Predict with Tweets.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "0dab174e",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-02T03:22:42.835932Z",
     "start_time": "2022-02-02T03:22:42.610954Z"
    }
   },
   "source": [
    "## Wordle From Tweeted Scores"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4470d46c",
   "metadata": {},
   "source": [
    "Inspired by [Ben Hamner's Kaggle](https://www.kaggle.com/benhamner/wordle-1-6/) and borrowing some methodology, one can predict the wordle of the day from publicly shared scores/guesses.\n",
    "\n",
    "I tweet the results from yesterday's Wordle (so not to spoil anything) pretty much every morning at [thewordlebot](https://twitter.com/thewordlebot).\n",
    "\n",
    "Major differences from above Kaggle notebook:\n",
    "\n",
    "* **No simulations of hypothetical games.** I do use a similar word commonality lookup dictionary, and the same [word frequency data](https://www.kaggle.com/rtatman/english-word-frequency).\n",
    "* **No cosine similarity or comparison of specific (e.g. penultimate) guess.** Only the list, and to some extent count, of all tweeted wordle score lines is needed.\n",
    "* **Slightly different filtering of bad tweets** Upon further reflection, the code now rejects obviously bad tweets. The kaggle data set does some [light filtering](https://www.kaggle.com/benhamner/pull-wordle-tweets). In general, using a minimum count threshold mostly eliminates a few fake or spurious scores that were posted, and the penalty term is small enough that even if the data set has a bunch of retweets of a [tetris pattern](https://twitter.com/TomWritesBlog/status/1489676441562361858), the algorithm still converged accurately. However, non-English wordles or people tweeting out multiple variants at once are both easy to detect so I'm [now removing these](https://github.com/astrowonk/TwitterWordle/commit/98ee732bf50d896e007c01520a95e90dc4edd4a7).\n",
    "* **100% Accuracy** This algorithm has 100% accuracy from Wordles 210-233 (the original project initially failed on 223, it was successful on later [reruns after some fixes](https://twitter.com/benhamner/status/1489364155370926080). It also failed on 231,236,and 249.) With the restricted target list, `TwitterWordle` is 100% accurate (so far).\n",
    "  * _Note: Starting on Feb 15, the NY Times removed a few target words from the official wordle list, such as 'papal' and 'agora.' Some people continue to tweet results from the unaltered list, presumably cached/saved versions. So, on Wordle 247, using the full 12,000+ dictionary, `TwitterWordle` did fail to solve correctly. However, the default mode is using the smaller target dictionary._\n",
    "* **By default, the code only considers the known 2315 possible wordles.** The Kaggle project doesn't give the wordle list special treatment, and runs simulations considering all 12K words as possible answers. While my [wordlebot](https://github.com/astrowonk/wordle) has rolled its own dictionary, I used the actual wordle list here. \n",
    "  * Use the keyword argument `use_limited_targets = False` to load precomputed dictionaries across the full 12972 word list, and the `Create Lookup dictionary` notebook can generate this larger set of dictionaries.\n",
    "  * The code still solves with (almost, see above) 100% accuracy using all 12K+ words, I solve the dataframe using the full list in this notebook at the end.\n",
    "  * As of Feb 15, 2021 I now use a revised dictionary after [changes made by the New York Times](https://arstechnica.com/gaming/2022/02/heres-how-the-new-york-times-changed-wordle/).\n",
    "\n",
    "\n",
    "This notebook uses the precomputed dictionaries inside the `TwitterWordle` class to solve a game from tweets. The other notebook generates the dictionaries.\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2801fb3d",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-02T03:44:11.042723Z",
     "start_time": "2022-02-02T03:44:10.970339Z"
    }
   },
   "source": [
    "### Basic Solving Methodology\n",
    "\n",
    "Each target word has a dictionary that maps a wordle line score `10102` (I map 🟩 to 2, 🟨 to 1, and ⬛ and ⬜ to 0.) to how likely that guess is for the target word. To solve:\n",
    "\n",
    "1. Count and collect the line scores in public tweets. The order doesn't matter.\n",
    "2. Using some minimum threshold for using the line score, go through every tweeted line score, and sum the average likelihood score for each unique tweeted score line.\n",
    "3. Subtract a penalty term if the result is impossible for all 12,000 guessable words. For example, only one word ends in 'uery' : query. It is impossible to guess and get `02222` as a result. There for if `02222` occurred in a tweet, the code would subtract a penalty term for `query` as a possible solution.\n",
    "4. The target word with the maximum summed lookup frequency is the wordle.\n",
    "\n",
    "\n",
    "Not every penalty term and minimum count is optimal for every wordle. By default the algorithm uses every tweet available (~2000+), requires at least 5 occurences of the line score, and uses -5E7 (around 1 standard deviation of all frequency scores) as penalty. (That makes it sound like I derived the penalty term in some rigorous way. In fact I just tried a bunch of them, -5E7 worked fairly well, then realized for different wordles, the code should iterate on different values until it finds a strong signal.)\n",
    "\n",
    "If the top answer is less than 10% higher than the runner up, the code iterates through min counts and penalty terms until it finds a prediction that is, or returns the best one."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "2ea31f8f",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-03T12:07:41.161851Z",
     "start_time": "2022-02-03T12:07:41.061334Z"
    }
   },
   "outputs": [],
   "source": [
    "\n",
    "import plotly.io as pio;# plots should work with matplotlib backend too but ... plotly is better!\n",
    "pio.renderers.default = \"png\" #here only for github, for plotly\n",
    "from TwitterWordle import TwitterWordle\n",
    "import pandas as pd\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c401dc13",
   "metadata": {},
   "source": [
    "### Load in data\n",
    "\n",
    "`tweets.csv` from the [kaggle project](https://www.kaggle.com/benhamner/wordle-tweets). Storing tweets is [complicated](https://developer.twitter.com/en/developer-terms/agreement-and-policy), so I'm not hosting any tweets in this repository, you should download from there. The dataset filters out some spurious tweets, but I have tested this code on unfiltered tweets collected with [searchtweets](https://pypi.org/project/searchtweets-v2/) which you can see in the subsequent section.\n",
    "\n",
    "You will need to download from there, as there are various rule on storing tweets (i.e. one must delete them if the account goes private), which requires some automation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "20a92572",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-03T12:07:44.234396Z",
     "start_time": "2022-02-03T12:07:43.985276Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loaded 2309 pre-computed lookup dictionaries.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/marcosh/Development/active_projects/TwitterWordle/TwitterWordle.py:50: SettingWithCopyWarning:\n",
      "\n",
      "\n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "\n"
     ]
    }
   ],
   "source": [
    "\n",
    "#load in the tweet data from kaggle.\n",
    "df = pd.read_csv(\"tweets.csv\")\n",
    "t = TwitterWordle(df)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "34e0bc0a",
   "metadata": {},
   "source": [
    "### Solve a Wordle, see a bar graph.\n",
    "\n",
    "The list of dictionaries is all that is needed to solve from a large number of tweeted results. Below is solving wordle 223, grabbing the tweets from the dataframe `df`. The default parameters are sufficient for most wordles with large N for tweets. The database includes about 2000 or so per tweet.\n",
    "\n",
    "For wordle 223 below, the default parameters find a very strong signal, so the code does not iterate through other penalty terms and minimum counts."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "d7b99682",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-03T12:07:47.749867Z",
     "start_time": "2022-02-03T12:07:46.621859Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TwitterWordle analyzed 15094 tweets for Wordle 223.\n",
      "\n",
      "50756 score patterns. 143 unique.\n",
      "\n",
      "95.30%, (142/149) valid final guess patterns found. Impossible pattern count: 1.\n",
      "\n",
      "Wordle 223 prediction: PERKY.\n",
      "2.5 STD above mean. 1.4 above runner up.\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "'perky'"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "t.solve(223,mask_result=False,plot=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "de72e129",
   "metadata": {},
   "source": [
    "The penalty term is averaged in when a wordle score line is impossible for a given target word. However, too high and spurious/fake tweets will penalize real possible target words.\n",
    "\n",
    "With fewer tweets, the likelihood of bogus tweets is lower, and minimum count can be reduced. Wordle 210 again with a downsampled size and a lower min_count. \n",
    "\n",
    "**100 tweets is enough (for some words)!** I lower the min count when downsampling, and the alg iterates through penalty terms until the signal means the 10% threshold."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "48aa8f9b",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-03T12:08:01.192283Z",
     "start_time": "2022-02-03T12:08:00.868388Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "354 score patterns. 75 unique.\n",
      "\n",
      "Wordle 223 initial signal low 1.01. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -3.10E+08\n",
      "50.34%, (75/149) valid final guess patterns found. Impossible pattern count: 0.\n",
      "\n",
      "Wordle 223 prediction: PERKY.\n",
      "2.7 STD above mean. 1.11 above runner up.\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "'perky'"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Solve with only 100 tweets\n",
    "t.solve(223,downsample=100,min_count=1,plot=True,mask_result=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "13313d46",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-01T12:13:42.743548Z",
     "start_time": "2022-02-01T12:13:42.700165Z"
    }
   },
   "source": [
    "### Solve all test wordles in the test dataframe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "5b35c340",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-03T12:08:17.112970Z",
     "start_time": "2022-02-03T12:08:15.408269Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TwitterWordle analyzed 2724 tweets for Wordle 210.\n",
      "\n",
      "7677 score patterns. 132 unique.\n",
      "\n",
      "89.58%, (129/144) valid final guess patterns found. Impossible pattern count: 3.\n",
      "\n",
      "Wordle 210 prediction: PANIC.\n",
      "1.8 STD above mean. 1.12 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 2764 tweets for Wordle 211.\n",
      "\n",
      "8014 score patterns. 164 unique.\n",
      "\n",
      "92.49%, (160/173) valid final guess patterns found. Impossible pattern count: 4.\n",
      "\n",
      "Wordle 211 prediction: SOLAR.\n",
      "3.2 STD above mean. 1.27 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 2287 tweets for Wordle 212.\n",
      "\n",
      "6521 score patterns. 159 unique.\n",
      "\n",
      "89.66%, (156/174) valid final guess patterns found. Impossible pattern count: 3.\n",
      "\n",
      "Wordle 212 prediction: SHIRE.\n",
      "3.3 STD above mean. 1.34 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 2018 tweets for Wordle 213.\n",
      "\n",
      "7260 score patterns. 94 unique.\n",
      "\n",
      "Wordle 213 initial signal low 1.06. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -1.30E+08\n",
      "80.00%, (88/110) valid final guess patterns found. Impossible pattern count: 6.\n",
      "\n",
      "Wordle 213 prediction: PROXY.\n",
      "1.9 STD above mean. 1.11 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 2585 tweets for Wordle 214.\n",
      "\n",
      "6648 score patterns. 142 unique.\n",
      "\n",
      "86.79%, (138/159) valid final guess patterns found. Impossible pattern count: 4.\n",
      "\n",
      "Wordle 214 prediction: POINT.\n",
      "1.9 STD above mean. 1.24 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 2820 tweets for Wordle 215.\n",
      "\n",
      "8140 score patterns. 126 unique.\n",
      "\n",
      "87.77%, (122/139) valid final guess patterns found. Impossible pattern count: 4.\n",
      "\n",
      "Wordle 215 prediction: ROBOT.\n",
      "2.0 STD above mean. 1.26 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 5878 tweets for Wordle 216.\n",
      "\n",
      "16778 score patterns. 121 unique.\n",
      "\n",
      "Wordle 216 initial signal low 1.1. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -7.00E+07\n",
      "91.27%, (115/126) valid final guess patterns found. Impossible pattern count: 6.\n",
      "\n",
      "Wordle 216 prediction: PRICK.\n",
      "2.4 STD above mean. 1.17 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 2834 tweets for Wordle 217.\n",
      "\n",
      "9353 score patterns. 124 unique.\n",
      "\n",
      "89.21%, (124/139) valid final guess patterns found. Impossible pattern count: 0.\n",
      "\n",
      "Wordle 217 prediction: WINCE.\n",
      "1.8 STD above mean. 1.12 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3357 tweets for Wordle 218.\n",
      "\n",
      "9843 score patterns. 108 unique.\n",
      "\n",
      "85.48%, (106/124) valid final guess patterns found. Impossible pattern count: 2.\n",
      "\n",
      "Wordle 218 prediction: CRIMP.\n",
      "2.0 STD above mean. 1.1 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 2772 tweets for Wordle 219.\n",
      "\n",
      "9967 score patterns. 102 unique.\n",
      "\n",
      "88.50%, (100/113) valid final guess patterns found. Impossible pattern count: 2.\n",
      "\n",
      "Wordle 219 prediction: KNOLL.\n",
      "2.0 STD above mean. 1.16 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 6388 tweets for Wordle 220.\n",
      "\n",
      "19594 score patterns. 163 unique.\n",
      "\n",
      "96.39%, (160/166) valid final guess patterns found. Impossible pattern count: 3.\n",
      "\n",
      "Wordle 220 prediction: SUGAR.\n",
      "3.6 STD above mean. 1.47 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 6361 tweets for Wordle 221.\n",
      "\n",
      "19754 score patterns. 103 unique.\n",
      "\n",
      "85.00%, (102/120) valid final guess patterns found. Impossible pattern count: 1.\n",
      "\n",
      "Wordle 221 prediction: WHACK.\n",
      "2.1 STD above mean. 1.2 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 6459 tweets for Wordle 222.\n",
      "\n",
      "18530 score patterns. 133 unique.\n",
      "\n",
      "87.76%, (129/147) valid final guess patterns found. Impossible pattern count: 4.\n",
      "\n",
      "Wordle 222 prediction: MOUNT.\n",
      "2.1 STD above mean. 1.15 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 15094 tweets for Wordle 223.\n",
      "\n",
      "50756 score patterns. 143 unique.\n",
      "\n",
      "95.30%, (142/149) valid final guess patterns found. Impossible pattern count: 1.\n",
      "\n",
      "Wordle 223 prediction: PERKY.\n",
      "2.5 STD above mean. 1.4 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 6652 tweets for Wordle 224.\n",
      "\n",
      "19326 score patterns. 134 unique.\n",
      "\n",
      "91.67%, (132/144) valid final guess patterns found. Impossible pattern count: 2.\n",
      "\n",
      "Wordle 224 prediction: COULD.\n",
      "1.9 STD above mean. 1.16 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 6629 tweets for Wordle 225.\n",
      "\n",
      "21485 score patterns. 96 unique.\n",
      "\n",
      "Wordle 225 initial signal low 1.04. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -7.00E+07\n",
      "89.72%, (96/107) valid final guess patterns found. Impossible pattern count: 0.\n",
      "\n",
      "Wordle 225 prediction: WRUNG.\n",
      "2.0 STD above mean. 1.13 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 5819 tweets for Wordle 226.\n",
      "\n",
      "16750 score patterns. 116 unique.\n",
      "\n",
      "89.52%, (111/124) valid final guess patterns found. Impossible pattern count: 5.\n",
      "\n",
      "Wordle 226 prediction: LIGHT.\n",
      "2.0 STD above mean. 1.15 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 6549 tweets for Wordle 227.\n",
      "\n",
      "17218 score patterns. 155 unique.\n",
      "\n",
      "93.87%, (153/163) valid final guess patterns found. Impossible pattern count: 2.\n",
      "\n",
      "Wordle 227 prediction: THOSE.\n",
      "3.3 STD above mean. 1.22 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 5728 tweets for Wordle 228.\n",
      "\n",
      "15530 score patterns. 153 unique.\n",
      "\n",
      "91.57%, (152/166) valid final guess patterns found. Impossible pattern count: 1.\n",
      "\n",
      "Wordle 228 prediction: MOIST.\n",
      "3.1 STD above mean. 1.46 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 6236 tweets for Wordle 229.\n",
      "\n",
      "19361 score patterns. 140 unique.\n",
      "\n",
      "89.17%, (140/157) valid final guess patterns found. Impossible pattern count: 0.\n",
      "\n",
      "Wordle 229 prediction: SHARD.\n",
      "2.7 STD above mean. 1.17 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 6396 tweets for Wordle 230.\n",
      "\n",
      "18346 score patterns. 173 unique.\n",
      "\n",
      "93.51%, (173/185) valid final guess patterns found. Impossible pattern count: 0.\n",
      "\n",
      "Wordle 230 prediction: PLEAT.\n",
      "3.9 STD above mean. 1.34 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 7521 tweets for Wordle 231.\n",
      "\n",
      "23563 score patterns. 135 unique.\n",
      "\n",
      "93.01%, (133/143) valid final guess patterns found. Impossible pattern count: 2.\n",
      "\n",
      "Wordle 231 prediction: ALOFT.\n",
      "2.2 STD above mean. 1.39 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3176 tweets for Wordle 232.\n",
      "\n",
      "10529 score patterns. 100 unique.\n",
      "\n",
      "Wordle 232 initial signal low 1.04. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -6.30E+08\n",
      "76.15%, (99/130) valid final guess patterns found. Impossible pattern count: 1.\n",
      "\n",
      "Wordle 232 prediction: SKILL.\n",
      "3.0 STD above mean. 1.1 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 6179 tweets for Wordle 233.\n",
      "\n",
      "22090 score patterns. 136 unique.\n",
      "\n",
      "96.40%, (134/139) valid final guess patterns found. Impossible pattern count: 2.\n",
      "\n",
      "Wordle 233 prediction: ELDER.\n",
      "2.8 STD above mean. 1.23 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 5945 tweets for Wordle 234.\n",
      "\n",
      "18443 score patterns. 141 unique.\n",
      "\n",
      "Wordle 234 initial signal low 1.05. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -7.00E+07\n",
      "86.96%, (140/161) valid final guess patterns found. Impossible pattern count: 1.\n",
      "\n",
      "Wordle 234 prediction: FRAME.\n",
      "3.5 STD above mean. 1.27 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 6240 tweets for Wordle 235.\n",
      "\n",
      "19809 score patterns. 125 unique.\n",
      "\n",
      "88.97%, (121/136) valid final guess patterns found. Impossible pattern count: 4.\n",
      "\n",
      "Wordle 235 prediction: HUMOR.\n",
      "2.3 STD above mean. 1.17 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 9626 tweets for Wordle 236.\n",
      "\n",
      "29157 score patterns. 157 unique.\n",
      "\n",
      "95.71%, (156/163) valid final guess patterns found. Impossible pattern count: 1.\n",
      "\n",
      "Wordle 236 prediction: PAUSE.\n",
      "3.0 STD above mean. 1.24 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 9735 tweets for Wordle 237.\n",
      "\n",
      "32603 score patterns. 146 unique.\n",
      "\n",
      "94.48%, (137/145) valid final guess patterns found. Impossible pattern count: 9.\n",
      "\n",
      "Wordle 237 prediction: ULCER.\n",
      "2.6 STD above mean. 1.43 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 6771 tweets for Wordle 238.\n",
      "\n",
      "21151 score patterns. 146 unique.\n",
      "\n",
      "97.18%, (138/142) valid final guess patterns found. Impossible pattern count: 8.\n",
      "\n",
      "Wordle 238 prediction: ULTRA.\n",
      "2.8 STD above mean. 1.3 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 4862 tweets for Wordle 239.\n",
      "\n",
      "14591 score patterns. 143 unique.\n",
      "\n",
      "92.16%, (141/153) valid final guess patterns found. Impossible pattern count: 2.\n",
      "\n",
      "Wordle 239 prediction: ROBIN.\n",
      "2.8 STD above mean. 1.36 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 6138 tweets for Wordle 240.\n",
      "\n",
      "21607 score patterns. 94 unique.\n",
      "\n",
      "87.50%, (91/104) valid final guess patterns found. Impossible pattern count: 3.\n",
      "\n",
      "Wordle 240 prediction: CYNIC.\n",
      "1.8 STD above mean. 1.13 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 6563 tweets for Wordle 241.\n",
      "\n",
      "20158 score patterns. 145 unique.\n",
      "\n",
      "94.29%, (132/140) valid final guess patterns found. Impossible pattern count: 13.\n",
      "\n",
      "Wordle 241 prediction: AROMA.\n",
      "2.3 STD above mean. 1.22 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3068 tweets for Wordle 242.\n",
      "\n",
      "9841 score patterns. 149 unique.\n",
      "\n",
      "96.45%, (136/141) valid final guess patterns found. Impossible pattern count: 13.\n",
      "\n",
      "Wordle 242 prediction: CAULK.\n",
      "2.1 STD above mean. 1.29 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 2625 tweets for Wordle 243.\n",
      "\n",
      "8674 score patterns. 145 unique.\n",
      "\n",
      "86.54%, (135/156) valid final guess patterns found. Impossible pattern count: 10.\n",
      "\n",
      "Wordle 243 prediction: SHAKE.\n",
      "2.3 STD above mean. 1.18 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3231 tweets for Wordle 244.\n",
      "\n",
      "11105 score patterns. 137 unique.\n",
      "\n",
      "92.48%, (123/133) valid final guess patterns found. Impossible pattern count: 14.\n",
      "\n",
      "Wordle 244 prediction: DODGE.\n",
      "2.0 STD above mean. 1.17 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 2778 tweets for Wordle 245.\n",
      "\n",
      "10579 score patterns. 134 unique.\n",
      "\n",
      "91.27%, (115/126) valid final guess patterns found. Impossible pattern count: 19.\n",
      "\n",
      "Wordle 245 prediction: SWILL.\n",
      "1.9 STD above mean. 1.11 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3140 tweets for Wordle 246.\n",
      "\n",
      "10029 score patterns. 140 unique.\n",
      "\n",
      "88.74%, (134/151) valid final guess patterns found. Impossible pattern count: 6.\n",
      "\n",
      "Wordle 246 prediction: TACIT.\n",
      "2.0 STD above mean. 1.23 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3380 tweets for Wordle 247.\n",
      "\n",
      "9968 score patterns. 127 unique.\n",
      "\n",
      "Wordle 247 initial signal low 1.03. Iterating for better parameters\n",
      "..............................................................................................\n",
      "Iterated to a better signal with min_count 3 and penalty -1.50E+08\n",
      "91.73%, (122/133) valid final guess patterns found. Impossible pattern count: 5.\n",
      "\n",
      "Wordle 247 prediction: OTHER.\n",
      "2.9 STD above mean. 1.13 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3488 tweets for Wordle 248.\n",
      "\n",
      "8834 score patterns. 148 unique.\n",
      "\n",
      "97.84%, (136/139) valid final guess patterns found. Impossible pattern count: 12.\n",
      "\n",
      "Wordle 248 prediction: THORN.\n",
      "2.7 STD above mean. 1.19 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 5933 tweets for Wordle 249.\n",
      "\n",
      "19991 score patterns. 146 unique.\n",
      "\n",
      "92.09%, (128/139) valid final guess patterns found. Impossible pattern count: 18.\n",
      "\n",
      "Wordle 249 prediction: TROVE.\n",
      "2.0 STD above mean. 1.12 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 6829 tweets for Wordle 250.\n",
      "\n",
      "21315 score patterns. 148 unique.\n",
      "\n",
      "Wordle 250 initial signal low 1.03. Iterating for better parameters\n",
      "............................................................................................................................................................................................\n",
      "Iterated to a better signal with min_count 7 and penalty -7.00E+07\n",
      "94.93%, (131/138) valid final guess patterns found. Impossible pattern count: 17.\n",
      "\n",
      "Wordle 250 prediction: BLOKE.\n",
      "2.1 STD above mean. 1.22 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 8601 tweets for Wordle 251.\n",
      "\n",
      "31065 score patterns. 120 unique.\n",
      "\n",
      "Wordle 251 initial signal low 1.04. Iterating for better parameters\n",
      "............................................................................................................................................................................................\n",
      "Iterated to a better signal with min_count 7 and penalty -9.00E+07\n",
      "92.86%, (78/84) valid final guess patterns found. Impossible pattern count: 42.\n",
      "\n",
      "Wordle 251 prediction: VIVID.\n",
      "2.0 STD above mean. 1.11 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 5915 tweets for Wordle 252.\n",
      "\n",
      "18127 score patterns. 141 unique.\n",
      "\n",
      "92.86%, (130/140) valid final guess patterns found. Impossible pattern count: 11.\n",
      "\n",
      "Wordle 252 prediction: SPILL.\n",
      "2.4 STD above mean. 1.11 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 7153 tweets for Wordle 253.\n",
      "\n",
      "19700 score patterns. 144 unique.\n",
      "\n",
      "93.29%, (139/149) valid final guess patterns found. Impossible pattern count: 5.\n",
      "\n",
      "Wordle 253 prediction: CHANT.\n",
      "2.3 STD above mean. 1.47 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 10321 tweets for Wordle 254.\n",
      "\n",
      "29009 score patterns. 135 unique.\n",
      "\n",
      "92.59%, (125/135) valid final guess patterns found. Impossible pattern count: 10.\n",
      "\n",
      "Wordle 254 prediction: CHOKE.\n",
      "2.5 STD above mean. 1.34 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 4352 tweets for Wordle 255.\n",
      "\n",
      "14713 score patterns. 150 unique.\n",
      "\n",
      "96.43%, (135/140) valid final guess patterns found. Impossible pattern count: 15.\n",
      "\n",
      "Wordle 255 prediction: RUPEE.\n",
      "2.5 STD above mean. 1.4 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 6792 tweets for Wordle 256.\n",
      "\n",
      "20400 score patterns. 151 unique.\n",
      "\n",
      "92.86%, (143/154) valid final guess patterns found. Impossible pattern count: 8.\n",
      "\n",
      "Wordle 256 prediction: NASTY.\n",
      "2.6 STD above mean. 1.13 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 7176 tweets for Wordle 257.\n",
      "\n",
      "20931 score patterns. 146 unique.\n",
      "\n",
      "90.48%, (133/147) valid final guess patterns found. Impossible pattern count: 13.\n",
      "\n",
      "Wordle 257 prediction: MOURN.\n",
      "2.4 STD above mean. 1.34 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 7014 tweets for Wordle 258.\n",
      "\n",
      "22687 score patterns. 149 unique.\n",
      "\n",
      "94.24%, (131/139) valid final guess patterns found. Impossible pattern count: 18.\n",
      "\n",
      "Wordle 258 prediction: AHEAD.\n",
      "2.6 STD above mean. 1.28 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3516 tweets for Wordle 259.\n",
      "\n",
      "10417 score patterns. 145 unique.\n",
      "\n",
      "90.45%, (142/157) valid final guess patterns found. Impossible pattern count: 3.\n",
      "\n",
      "Wordle 259 prediction: BRINE.\n",
      "2.2 STD above mean. 1.28 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 7405 tweets for Wordle 260.\n",
      "\n",
      "20831 score patterns. 136 unique.\n",
      "\n",
      "92.86%, (130/140) valid final guess patterns found. Impossible pattern count: 6.\n",
      "\n",
      "Wordle 260 prediction: CLOTH.\n",
      "2.7 STD above mean. 1.33 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 7253 tweets for Wordle 261.\n",
      "\n",
      "20517 score patterns. 163 unique.\n",
      "\n",
      "96.25%, (154/160) valid final guess patterns found. Impossible pattern count: 9.\n",
      "\n",
      "Wordle 261 prediction: HOARD.\n",
      "2.8 STD above mean. 1.47 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 5707 tweets for Wordle 262.\n",
      "\n",
      "18595 score patterns. 146 unique.\n",
      "\n",
      "90.32%, (140/155) valid final guess patterns found. Impossible pattern count: 6.\n",
      "\n",
      "Wordle 262 prediction: SWEET.\n",
      "2.3 STD above mean. 1.16 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 7343 tweets for Wordle 263.\n",
      "\n",
      "22132 score patterns. 135 unique.\n",
      "\n",
      "Wordle 263 initial signal low 1.06. Iterating for better parameters\n",
      "..............................................................................................\n",
      "Iterated to a better signal with min_count 3 and penalty -1.90E+08\n",
      "96.72%, (118/122) valid final guess patterns found. Impossible pattern count: 17.\n",
      "\n",
      "Wordle 263 prediction: MONTH.\n",
      "2.9 STD above mean. 1.1 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3658 tweets for Wordle 264.\n",
      "\n",
      "10489 score patterns. 170 unique.\n",
      "\n",
      "91.91%, (159/173) valid final guess patterns found. Impossible pattern count: 11.\n",
      "\n",
      "Wordle 264 prediction: LAPSE.\n",
      "2.4 STD above mean. 1.21 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 4923 tweets for Wordle 265.\n",
      "\n",
      "16476 score patterns. 132 unique.\n",
      "\n",
      "Wordle 265 initial signal low 1.08. Iterating for better parameters\n",
      "..............................................................................................\n",
      "Iterated to a better signal with min_count 3 and penalty -1.10E+08\n",
      "88.97%, (121/136) valid final guess patterns found. Impossible pattern count: 11.\n",
      "\n",
      "Wordle 265 prediction: WATCH.\n",
      "2.4 STD above mean. 1.1 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3596 tweets for Wordle 266.\n",
      "\n",
      "10524 score patterns. 150 unique.\n",
      "\n",
      "95.68%, (133/139) valid final guess patterns found. Impossible pattern count: 17.\n",
      "\n",
      "Wordle 266 prediction: TODAY.\n",
      "2.2 STD above mean. 1.13 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3643 tweets for Wordle 267.\n",
      "\n",
      "11632 score patterns. 124 unique.\n",
      "\n",
      "88.49%, (123/139) valid final guess patterns found. Impossible pattern count: 1.\n",
      "\n",
      "Wordle 267 prediction: FOCUS.\n",
      "1.8 STD above mean. 1.15 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3500 tweets for Wordle 268.\n",
      "\n",
      "11253 score patterns. 151 unique.\n",
      "\n",
      "Wordle 268 initial signal low 1.06. Iterating for better parameters\n",
      "..........................................................................................................................................................................................................................................................................................\n",
      "Iterated to a better signal with min_count 11 and penalty -9.90E+08\n",
      "96.64%, (144/149) valid final guess patterns found. Impossible pattern count: 7.\n",
      "\n",
      "Wordle 268 prediction: SMELT.\n",
      "2.4 STD above mean. 1.06 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3590 tweets for Wordle 269.\n",
      "\n",
      "9308 score patterns. 150 unique.\n",
      "\n",
      "86.13%, (149/173) valid final guess patterns found. Impossible pattern count: 1.\n",
      "\n",
      "Wordle 269 prediction: TEASE.\n",
      "3.2 STD above mean. 1.56 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 2865 tweets for Wordle 270.\n",
      "\n",
      "9081 score patterns. 161 unique.\n",
      "\n",
      "87.15%, (156/179) valid final guess patterns found. Impossible pattern count: 5.\n",
      "\n",
      "Wordle 270 prediction: CATER.\n",
      "2.5 STD above mean. 1.42 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3438 tweets for Wordle 271.\n",
      "\n",
      "11336 score patterns. 136 unique.\n",
      "\n",
      "Wordle 271 initial signal low 1.03. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -3.70E+08\n",
      "76.79%, (129/168) valid final guess patterns found. Impossible pattern count: 7.\n",
      "\n",
      "Wordle 271 prediction: BARON.\n",
      "2.1 STD above mean. 1.1 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3665 tweets for Wordle 272.\n",
      "\n",
      "10641 score patterns. 178 unique.\n",
      "\n",
      "93.44%, (171/183) valid final guess patterns found. Impossible pattern count: 7.\n",
      "\n",
      "Wordle 272 prediction: SAUTE.\n",
      "3.2 STD above mean. 1.23 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3509 tweets for Wordle 273.\n",
      "\n",
      "11558 score patterns. 115 unique.\n",
      "\n",
      "92.79%, (103/111) valid final guess patterns found. Impossible pattern count: 12.\n",
      "\n",
      "Wordle 273 prediction: ALLOW.\n",
      "1.9 STD above mean. 1.15 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3635 tweets for Wordle 274.\n",
      "\n",
      "11801 score patterns. 129 unique.\n",
      "\n",
      "96.88%, (124/128) valid final guess patterns found. Impossible pattern count: 5.\n",
      "\n",
      "Wordle 274 prediction: RENEW.\n",
      "2.3 STD above mean. 1.24 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3564 tweets for Wordle 275.\n",
      "\n",
      "9122 score patterns. 165 unique.\n",
      "\n",
      "Wordle 275 initial signal low 1.04. Iterating for better parameters\n",
      "..............................................................................................\n",
      "Iterated to a better signal with min_count 3 and penalty -7.00E+07\n",
      "96.82%, (152/157) valid final guess patterns found. Impossible pattern count: 13.\n",
      "\n",
      "Wordle 275 prediction: THEIR.\n",
      "2.7 STD above mean. 1.12 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3645 tweets for Wordle 276.\n",
      "\n",
      "11787 score patterns. 119 unique.\n",
      "\n",
      "87.69%, (114/130) valid final guess patterns found. Impossible pattern count: 5.\n",
      "\n",
      "Wordle 276 prediction: SLOSH.\n",
      "2.0 STD above mean. 1.13 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3682 tweets for Wordle 277.\n",
      "\n",
      "11686 score patterns. 146 unique.\n",
      "\n",
      "93.20%, (137/147) valid final guess patterns found. Impossible pattern count: 9.\n",
      "\n",
      "Wordle 277 prediction: PURGE.\n",
      "2.0 STD above mean. 1.17 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3574 tweets for Wordle 279.\n",
      "\n",
      "10494 score patterns. 153 unique.\n",
      "\n",
      "96.05%, (146/152) valid final guess patterns found. Impossible pattern count: 7.\n",
      "\n",
      "Wordle 279 prediction: DEPOT.\n",
      "2.8 STD above mean. 1.44 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 2942 tweets for Wordle 280.\n",
      "\n",
      "10397 score patterns. 102 unique.\n",
      "\n",
      "Wordle 280 initial signal low 1.09. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -7.00E+07\n",
      "87.50%, (91/104) valid final guess patterns found. Impossible pattern count: 11.\n",
      "\n",
      "Wordle 280 prediction: EPOXY.\n",
      "1.7 STD above mean. 1.11 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3041 tweets for Wordle 281.\n",
      "\n",
      "9665 score patterns. 88 unique.\n",
      "\n",
      "94.32%, (83/88) valid final guess patterns found. Impossible pattern count: 5.\n",
      "\n",
      "Wordle 281 prediction: NYMPH.\n",
      "2.3 STD above mean. 1.1 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 2693 tweets for Wordle 282.\n",
      "\n",
      "8961 score patterns. 122 unique.\n",
      "\n",
      "87.88%, (116/132) valid final guess patterns found. Impossible pattern count: 6.\n",
      "\n",
      "Wordle 282 prediction: FOUND.\n",
      "1.8 STD above mean. 1.12 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3333 tweets for Wordle 283.\n",
      "\n",
      "11148 score patterns. 135 unique.\n",
      "\n",
      "87.07%, (128/147) valid final guess patterns found. Impossible pattern count: 7.\n",
      "\n",
      "Wordle 283 prediction: SHALL.\n",
      "2.0 STD above mean. 1.13 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3088 tweets for Wordle 284.\n",
      "\n",
      "10466 score patterns. 127 unique.\n",
      "\n",
      "Wordle 284 initial signal low 1.07. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -1.10E+08\n",
      "69.89%, (123/176) valid final guess patterns found. Impossible pattern count: 4.\n",
      "\n",
      "Wordle 284 prediction: SLOPE.\n",
      "2.9 STD above mean. 1.11 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3162 tweets for Wordle 285.\n",
      "\n",
      "11486 score patterns. 121 unique.\n",
      "\n",
      "Wordle 285 initial signal low 1.06. Iterating for better parameters\n",
      "..............................................................................................\n",
      "Iterated to a better signal with min_count 3 and penalty -2.70E+08\n",
      "80.15%, (105/131) valid final guess patterns found. Impossible pattern count: 16.\n",
      "\n",
      "Wordle 285 prediction: DANDY.\n",
      "2.7 STD above mean. 1.1 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3124 tweets for Wordle 286.\n",
      "\n",
      "10361 score patterns. 150 unique.\n",
      "\n",
      "87.50%, (140/160) valid final guess patterns found. Impossible pattern count: 10.\n",
      "\n",
      "Wordle 286 prediction: SNOUT.\n",
      "2.7 STD above mean. 1.34 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3356 tweets for Wordle 287.\n",
      "\n",
      "8733 score patterns. 158 unique.\n",
      "\n",
      "91.98%, (149/162) valid final guess patterns found. Impossible pattern count: 9.\n",
      "\n",
      "Wordle 287 prediction: TROPE.\n",
      "2.1 STD above mean. 1.22 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3420 tweets for Wordle 289.\n",
      "\n",
      "11574 score patterns. 149 unique.\n",
      "\n",
      "90.97%, (131/144) valid final guess patterns found. Impossible pattern count: 18.\n",
      "\n",
      "Wordle 289 prediction: SHAWL.\n",
      "2.4 STD above mean. 1.22 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3402 tweets for Wordle 290.\n",
      "\n",
      "11773 score patterns. 157 unique.\n",
      "\n",
      "Wordle 290 initial signal low 1.09. Iterating for better parameters\n",
      ".............................................................................................................................................\n",
      "Iterated to a better signal with min_count 5 and penalty -1.70E+08\n",
      "95.00%, (133/140) valid final guess patterns found. Impossible pattern count: 24.\n",
      "\n",
      "Wordle 290 prediction: NATAL.\n",
      "2.9 STD above mean. 1.11 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 6834 tweets for Wordle 291.\n",
      "\n",
      "23868 score patterns. 151 unique.\n",
      "\n",
      "Wordle 291 initial signal low 1.1. Iterating for better parameters\n",
      ".............................................................................................................................................\n",
      "Iterated to a better signal with min_count 5 and penalty -7.00E+07\n",
      "96.95%, (127/131) valid final guess patterns found. Impossible pattern count: 24.\n",
      "\n",
      "Wordle 291 prediction: COMMA.\n",
      "2.4 STD above mean. 1.28 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3434 tweets for Wordle 292.\n",
      "\n",
      "12053 score patterns. 151 unique.\n",
      "\n",
      "93.46%, (143/153) valid final guess patterns found. Impossible pattern count: 8.\n",
      "\n",
      "Wordle 292 prediction: FORAY.\n",
      "2.5 STD above mean. 1.34 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3256 tweets for Wordle 293.\n",
      "\n",
      "9589 score patterns. 168 unique.\n",
      "\n",
      "Wordle 293 initial signal low 1.04. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -7.00E+07\n",
      "88.40%, (160/181) valid final guess patterns found. Impossible pattern count: 8.\n",
      "\n",
      "Wordle 293 prediction: SCARE.\n",
      "3.5 STD above mean. 1.16 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3461 tweets for Wordle 294.\n",
      "\n",
      "8367 score patterns. 166 unique.\n",
      "\n",
      "90.96%, (151/166) valid final guess patterns found. Impossible pattern count: 15.\n",
      "\n",
      "Wordle 294 prediction: STAIR.\n",
      "2.2 STD above mean. 1.12 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3645 tweets for Wordle 295.\n",
      "\n",
      "10146 score patterns. 143 unique.\n",
      "\n",
      "Wordle 295 initial signal low 1.0. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -1.30E+08\n",
      "75.72%, (131/173) valid final guess patterns found. Impossible pattern count: 12.\n",
      "\n",
      "Wordle 295 prediction: GRADE.\n",
      "2.3 STD above mean. 1.11 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3498 tweets for Wordle 296.\n",
      "\n",
      "11760 score patterns. 124 unique.\n",
      "\n",
      "Wordle 296 initial signal low 1.09. Iterating for better parameters\n",
      "..............................................................................................\n",
      "Iterated to a better signal with min_count 3 and penalty -1.10E+08\n",
      "93.04%, (107/115) valid final guess patterns found. Impossible pattern count: 17.\n",
      "\n",
      "Wordle 296 prediction: SQUAD.\n",
      "2.5 STD above mean. 1.11 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3631 tweets for Wordle 297.\n",
      "\n",
      "11233 score patterns. 154 unique.\n",
      "\n",
      "92.99%, (146/157) valid final guess patterns found. Impossible pattern count: 8.\n",
      "\n",
      "Wordle 297 prediction: ROYAL.\n",
      "2.5 STD above mean. 1.26 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3646 tweets for Wordle 298.\n",
      "\n",
      "10546 score patterns. 121 unique.\n",
      "\n",
      "84.35%, (97/115) valid final guess patterns found. Impossible pattern count: 24.\n",
      "\n",
      "Wordle 298 prediction: CHUNK.\n",
      "2.0 STD above mean. 1.11 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3705 tweets for Wordle 299.\n",
      "\n",
      "11552 score patterns. 143 unique.\n",
      "\n",
      "91.22%, (135/148) valid final guess patterns found. Impossible pattern count: 8.\n",
      "\n",
      "Wordle 299 prediction: MINCE.\n",
      "2.0 STD above mean. 1.23 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3241 tweets for Wordle 300.\n",
      "\n",
      "9678 score patterns. 150 unique.\n",
      "\n",
      "Wordle 300 initial signal low 1.05. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -7.00E+07\n",
      "85.89%, (140/163) valid final guess patterns found. Impossible pattern count: 10.\n",
      "\n",
      "Wordle 300 prediction: SHAME.\n",
      "2.9 STD above mean. 1.21 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3709 tweets for Wordle 301.\n",
      "\n",
      "11629 score patterns. 119 unique.\n",
      "\n",
      "88.03%, (103/117) valid final guess patterns found. Impossible pattern count: 16.\n",
      "\n",
      "Wordle 301 prediction: CHEEK.\n",
      "2.2 STD above mean. 1.19 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3579 tweets for Wordle 302.\n",
      "\n",
      "11388 score patterns. 145 unique.\n",
      "\n",
      "Wordle 302 initial signal low 1.08. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -7.00E+07\n",
      "89.54%, (137/153) valid final guess patterns found. Impossible pattern count: 8.\n",
      "\n",
      "Wordle 302 prediction: AMPLE.\n",
      "3.1 STD above mean. 1.34 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3621 tweets for Wordle 303.\n",
      "\n",
      "10203 score patterns. 149 unique.\n",
      "\n",
      "88.54%, (139/157) valid final guess patterns found. Impossible pattern count: 10.\n",
      "\n",
      "Wordle 303 prediction: FLAIR.\n",
      "2.4 STD above mean. 1.18 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 2472 tweets for Wordle 304.\n",
      "\n",
      "8954 score patterns. 144 unique.\n",
      "\n",
      "88.89%, (128/144) valid final guess patterns found. Impossible pattern count: 16.\n",
      "\n",
      "Wordle 304 prediction: FOYER.\n",
      "2.3 STD above mean. 1.2 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3589 tweets for Wordle 305.\n",
      "\n",
      "11705 score patterns. 151 unique.\n",
      "\n",
      "92.26%, (143/155) valid final guess patterns found. Impossible pattern count: 8.\n",
      "\n",
      "Wordle 305 prediction: CARGO.\n",
      "2.4 STD above mean. 1.25 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3190 tweets for Wordle 306.\n",
      "\n",
      "11611 score patterns. 122 unique.\n",
      "\n",
      "86.51%, (109/126) valid final guess patterns found. Impossible pattern count: 13.\n",
      "\n",
      "Wordle 306 prediction: OXIDE.\n",
      "2.3 STD above mean. 1.21 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3450 tweets for Wordle 307.\n",
      "\n",
      "8169 score patterns. 131 unique.\n",
      "\n",
      "87.94%, (124/141) valid final guess patterns found. Impossible pattern count: 7.\n",
      "\n",
      "Wordle 307 prediction: PLANT.\n",
      "2.3 STD above mean. 1.22 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3532 tweets for Wordle 308.\n",
      "\n",
      "10945 score patterns. 129 unique.\n",
      "\n",
      "82.98%, (117/141) valid final guess patterns found. Impossible pattern count: 12.\n",
      "\n",
      "Wordle 308 prediction: OLIVE.\n",
      "1.9 STD above mean. 1.17 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3555 tweets for Wordle 309.\n",
      "\n",
      "10872 score patterns. 155 unique.\n",
      "\n",
      "94.94%, (150/158) valid final guess patterns found. Impossible pattern count: 5.\n",
      "\n",
      "Wordle 309 prediction: INERT.\n",
      "3.2 STD above mean. 1.52 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3439 tweets for Wordle 310.\n",
      "\n",
      "12066 score patterns. 124 unique.\n",
      "\n",
      "83.87%, (104/124) valid final guess patterns found. Impossible pattern count: 20.\n",
      "\n",
      "Wordle 310 prediction: ASKEW.\n",
      "2.9 STD above mean. 1.14 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3499 tweets for Wordle 311.\n",
      "\n",
      "9453 score patterns. 167 unique.\n",
      "\n",
      "94.05%, (158/168) valid final guess patterns found. Impossible pattern count: 9.\n",
      "\n",
      "Wordle 311 prediction: HEIST.\n",
      "3.5 STD above mean. 1.39 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3618 tweets for Wordle 312.\n",
      "\n",
      "10728 score patterns. 138 unique.\n",
      "\n",
      "Wordle 312 initial signal low 1.07. Iterating for better parameters\n",
      "..............................................................................................\n",
      "Iterated to a better signal with min_count 3 and penalty -9.00E+07\n",
      "87.31%, (117/134) valid final guess patterns found. Impossible pattern count: 21.\n",
      "\n",
      "Wordle 312 prediction: SHOWN.\n",
      "2.0 STD above mean. 1.1 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3452 tweets for Wordle 313.\n",
      "\n",
      "12193 score patterns. 120 unique.\n",
      "\n",
      "84.43%, (103/122) valid final guess patterns found. Impossible pattern count: 17.\n",
      "\n",
      "Wordle 313 prediction: ZESTY.\n",
      "2.4 STD above mean. 1.12 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3394 tweets for Wordle 314.\n",
      "\n",
      "8380 score patterns. 156 unique.\n",
      "\n",
      "90.42%, (151/167) valid final guess patterns found. Impossible pattern count: 5.\n",
      "\n",
      "Wordle 314 prediction: TRASH.\n",
      "3.4 STD above mean. 1.31 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3389 tweets for Wordle 315.\n",
      "\n",
      "12258 score patterns. 133 unique.\n",
      "\n",
      "95.90%, (117/122) valid final guess patterns found. Impossible pattern count: 16.\n",
      "\n",
      "Wordle 315 prediction: LARVA.\n",
      "2.2 STD above mean. 1.12 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3489 tweets for Wordle 316.\n",
      "\n",
      "13017 score patterns. 128 unique.\n",
      "\n",
      "Wordle 316 initial signal low 1.1. Iterating for better parameters\n",
      "..............................................................................................\n",
      "Iterated to a better signal with min_count 3 and penalty -9.00E+07\n",
      "88.71%, (110/124) valid final guess patterns found. Impossible pattern count: 18.\n",
      "\n",
      "Wordle 316 prediction: FORGO.\n",
      "1.8 STD above mean. 1.1 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3428 tweets for Wordle 317.\n",
      "\n",
      "10235 score patterns. 158 unique.\n",
      "\n",
      "Wordle 317 initial signal low 1.04. Iterating for better parameters\n",
      "..............................................................................................\n",
      "Iterated to a better signal with min_count 3 and penalty -1.50E+08\n",
      "94.67%, (142/150) valid final guess patterns found. Impossible pattern count: 16.\n",
      "\n",
      "Wordle 317 prediction: STORY.\n",
      "2.4 STD above mean. 1.1 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 7115 tweets for Wordle 318.\n",
      "\n",
      "21130 score patterns. 142 unique.\n",
      "\n",
      "91.84%, (135/147) valid final guess patterns found. Impossible pattern count: 7.\n",
      "\n",
      "Wordle 318 prediction: HAIRY.\n",
      "2.4 STD above mean. 1.4 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3160 tweets for Wordle 319.\n",
      "\n",
      "7192 score patterns. 150 unique.\n",
      "\n",
      "84.57%, (148/175) valid final guess patterns found. Impossible pattern count: 2.\n",
      "\n",
      "Wordle 319 prediction: TRAIN.\n",
      "2.3 STD above mean. 1.17 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 2883 tweets for Wordle 320.\n",
      "\n",
      "9563 score patterns. 144 unique.\n",
      "\n",
      "88.89%, (136/153) valid final guess patterns found. Impossible pattern count: 8.\n",
      "\n",
      "Wordle 320 prediction: HOMER.\n",
      "2.2 STD above mean. 1.23 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3493 tweets for Wordle 321.\n",
      "\n",
      "11342 score patterns. 143 unique.\n",
      "\n",
      "91.10%, (133/146) valid final guess patterns found. Impossible pattern count: 10.\n",
      "\n",
      "Wordle 321 prediction: BADGE.\n",
      "2.0 STD above mean. 1.19 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3533 tweets for Wordle 322.\n",
      "\n",
      "11088 score patterns. 140 unique.\n",
      "\n",
      "91.73%, (122/133) valid final guess patterns found. Impossible pattern count: 18.\n",
      "\n",
      "Wordle 322 prediction: MIDST.\n",
      "2.0 STD above mean. 1.17 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3402 tweets for Wordle 323.\n",
      "\n",
      "12277 score patterns. 121 unique.\n",
      "\n",
      "Wordle 323 initial signal low 1.06. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -7.00E+07\n",
      "87.79%, (115/131) valid final guess patterns found. Impossible pattern count: 6.\n",
      "\n",
      "Wordle 323 prediction: CANNY.\n",
      "2.4 STD above mean. 1.16 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3444 tweets for Wordle 324.\n",
      "\n",
      "9476 score patterns. 169 unique.\n",
      "\n",
      "92.26%, (155/168) valid final guess patterns found. Impossible pattern count: 14.\n",
      "\n",
      "Wordle 324 prediction: SHINE.\n",
      "2.5 STD above mean. 1.16 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3472 tweets for Wordle 325.\n",
      "\n",
      "11587 score patterns. 130 unique.\n",
      "\n",
      "Wordle 325 initial signal low 1.02. Iterating for better parameters\n",
      "..............................................................................................\n",
      "Iterated to a better signal with min_count 3 and penalty -1.70E+08\n",
      "77.48%, (117/151) valid final guess patterns found. Impossible pattern count: 13.\n",
      "\n",
      "Wordle 325 prediction: GIRLY.\n",
      "2.2 STD above mean. 1.1 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3528 tweets for Wordle 326.\n",
      "\n",
      "10556 score patterns. 158 unique.\n",
      "\n",
      "91.18%, (155/170) valid final guess patterns found. Impossible pattern count: 3.\n",
      "\n",
      "Wordle 326 prediction: FARCE.\n",
      "2.3 STD above mean. 1.2 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3510 tweets for Wordle 327.\n",
      "\n",
      "11882 score patterns. 132 unique.\n",
      "\n",
      "87.68%, (121/138) valid final guess patterns found. Impossible pattern count: 11.\n",
      "\n",
      "Wordle 327 prediction: SLUNG.\n",
      "2.0 STD above mean. 1.16 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3495 tweets for Wordle 328.\n",
      "\n",
      "9742 score patterns. 139 unique.\n",
      "\n",
      "91.16%, (134/147) valid final guess patterns found. Impossible pattern count: 5.\n",
      "\n",
      "Wordle 328 prediction: TIPSY.\n",
      "2.3 STD above mean. 1.16 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3545 tweets for Wordle 329.\n",
      "\n",
      "10102 score patterns. 158 unique.\n",
      "\n",
      "Wordle 329 initial signal low 1.04. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -7.00E+07\n",
      "92.12%, (152/165) valid final guess patterns found. Impossible pattern count: 6.\n",
      "\n",
      "Wordle 329 prediction: METAL.\n",
      "3.0 STD above mean. 1.15 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3424 tweets for Wordle 330.\n",
      "\n",
      "11711 score patterns. 145 unique.\n",
      "\n",
      "93.84%, (137/146) valid final guess patterns found. Impossible pattern count: 8.\n",
      "\n",
      "Wordle 330 prediction: YIELD.\n",
      "2.8 STD above mean. 1.31 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3360 tweets for Wordle 331.\n",
      "\n",
      "11595 score patterns. 137 unique.\n",
      "\n",
      "Wordle 331 initial signal low 1.08. Iterating for better parameters\n",
      "..............................................................................................\n",
      "Iterated to a better signal with min_count 3 and penalty -1.30E+08\n",
      "93.98%, (125/133) valid final guess patterns found. Impossible pattern count: 12.\n",
      "\n",
      "Wordle 331 prediction: DELVE.\n",
      "3.0 STD above mean. 1.1 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3558 tweets for Wordle 332.\n",
      "\n",
      "11303 score patterns. 136 unique.\n",
      "\n",
      "85.42%, (123/144) valid final guess patterns found. Impossible pattern count: 13.\n",
      "\n",
      "Wordle 332 prediction: BEING.\n",
      "1.7 STD above mean. 1.11 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3605 tweets for Wordle 333.\n",
      "\n",
      "10104 score patterns. 142 unique.\n",
      "\n",
      "91.28%, (136/149) valid final guess patterns found. Impossible pattern count: 6.\n",
      "\n",
      "Wordle 333 prediction: SCOUR.\n",
      "2.3 STD above mean. 1.19 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3452 tweets for Wordle 334.\n",
      "\n",
      "11150 score patterns. 131 unique.\n",
      "\n",
      "Wordle 334 initial signal low 1.08. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -7.00E+07\n",
      "89.29%, (125/140) valid final guess patterns found. Impossible pattern count: 6.\n",
      "\n",
      "Wordle 334 prediction: GLASS.\n",
      "2.4 STD above mean. 1.2 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3181 tweets for Wordle 335.\n",
      "\n",
      "10664 score patterns. 150 unique.\n",
      "\n",
      "86.67%, (143/165) valid final guess patterns found. Impossible pattern count: 7.\n",
      "\n",
      "Wordle 335 prediction: GAMER.\n",
      "2.1 STD above mean. 1.12 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3526 tweets for Wordle 336.\n",
      "\n",
      "10304 score patterns. 148 unique.\n",
      "\n",
      "91.39%, (138/151) valid final guess patterns found. Impossible pattern count: 10.\n",
      "\n",
      "Wordle 336 prediction: SCRAP.\n",
      "2.6 STD above mean. 1.18 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3555 tweets for Wordle 337.\n",
      "\n",
      "10568 score patterns. 145 unique.\n",
      "\n",
      "88.46%, (138/156) valid final guess patterns found. Impossible pattern count: 7.\n",
      "\n",
      "Wordle 337 prediction: MONEY.\n",
      "2.7 STD above mean. 1.19 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3513 tweets for Wordle 338.\n",
      "\n",
      "10729 score patterns. 133 unique.\n",
      "\n",
      "89.51%, (128/143) valid final guess patterns found. Impossible pattern count: 5.\n",
      "\n",
      "Wordle 338 prediction: HINGE.\n",
      "2.0 STD above mean. 1.2 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3525 tweets for Wordle 339.\n",
      "\n",
      "10987 score patterns. 137 unique.\n",
      "\n",
      "91.97%, (126/137) valid final guess patterns found. Impossible pattern count: 11.\n",
      "\n",
      "Wordle 339 prediction: ALBUM.\n",
      "2.6 STD above mean. 1.36 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3321 tweets for Wordle 340.\n",
      "\n",
      "12268 score patterns. 100 unique.\n",
      "\n",
      "85.29%, (87/102) valid final guess patterns found. Impossible pattern count: 13.\n",
      "\n",
      "Wordle 340 prediction: VOUCH.\n",
      "1.8 STD above mean. 1.15 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3547 tweets for Wordle 341.\n",
      "\n",
      "10584 score patterns. 140 unique.\n",
      "\n",
      "Wordle 341 initial signal low 1.04. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -1.30E+08\n",
      "87.33%, (131/150) valid final guess patterns found. Impossible pattern count: 9.\n",
      "\n",
      "Wordle 341 prediction: ASSET.\n",
      "2.7 STD above mean. 1.11 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3516 tweets for Wordle 342.\n",
      "\n",
      "9690 score patterns. 142 unique.\n",
      "\n",
      "83.83%, (140/167) valid final guess patterns found. Impossible pattern count: 2.\n",
      "\n",
      "Wordle 342 prediction: TIARA.\n",
      "2.0 STD above mean. 1.13 above runner up.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "t.solve_all(mask_result=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d64f9ee0",
   "metadata": {},
   "source": [
    "### Solve with acquired tweets\n",
    "\n",
    "If you have a Twitter developer account, I have a helper function for [searchtweets](https://pypi.org/project/searchtweets-v2/)  that will pull recent tweets for a given wordle number. It assumes a configuration in `~/.twitter_keys.yaml`.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "ff289313",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loaded 2309 pre-computed lookup dictionaries.\n",
      "TwitterWordle analyzed 2775 tweets for Wordle 341.\n",
      "\n",
      "8326 score patterns. 136 unique.\n",
      "\n",
      "Wordle 341 initial signal low 1.07. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -7.00E+07\n",
      "84.67%, (127/150) valid final guess patterns found. Impossible pattern count: 9.\n",
      "\n",
      "Wordle 341 prediction: ASSET.\n",
      "2.8 STD above mean. 1.15 above runner up.\n",
      "\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/marcosh/Development/active_projects/TwitterWordle/TwitterWordle.py:50: SettingWithCopyWarning:\n",
      "\n",
      "\n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "'asset'"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from get_tweets import get_tweets\n",
    "tweets = get_tweets(341,return_df=True)\n",
    "t = TwitterWordle(tweets)\n",
    "t.solve(341,mask_result=False,plot=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "709ba6ad",
   "metadata": {},
   "source": [
    "### Solve using the full 12972 length dictionary.\n",
    "\n",
    "Limiting to the 2315 possible wordle answers likely improves performance, however the code works well considering all 12972 guessable words as possible answers.\n",
    "\n",
    "In mid-February, 2022, the NY Times removed a few wordle answers and allowable guesses. Specifically, some players continue to play cached/archive versions of the original word list, and 'impossible scores' have become much more common on Twitter, with the majority playing the official Wordle, but a few people playing against a different target word.\n",
    "\n",
    "The code has been generally robust to this, but raising the min_count may be necessary after 2022-02-15."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "80c258db",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loaded 12945 pre-computed lookup dictionaries.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/marcosh/Development/active_projects/TwitterWordle/TwitterWordle.py:50: SettingWithCopyWarning:\n",
      "\n",
      "\n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "\n"
     ]
    }
   ],
   "source": [
    "t_full_dictionary = TwitterWordle(df.query('wordle_id > 330'),use_limited_targets=False)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "bf15f050",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TwitterWordle analyzed 3360 tweets for Wordle 331.\n",
      "\n",
      "11595 score patterns. 137 unique.\n",
      "\n",
      "Wordle 331 initial signal low 1.09. Iterating for better parameters\n",
      "..............................................................................................\n",
      "Iterated to a better signal with min_count 3 and penalty -9.00E+07\n",
      "93.98%, (125/133) valid final guess patterns found. Impossible pattern count: 12.\n",
      "\n",
      "Wordle 331 prediction: DELVE.\n",
      "2.6 STD above mean. 1.1 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3558 tweets for Wordle 332.\n",
      "\n",
      "11303 score patterns. 136 unique.\n",
      "\n",
      "Wordle 332 initial signal low 1.04. Iterating for better parameters\n",
      "..............................................................................................\n",
      "Iterated to a better signal with min_count 3 and penalty -3.70E+08\n",
      "85.42%, (123/144) valid final guess patterns found. Impossible pattern count: 13.\n",
      "\n",
      "Wordle 332 prediction: BEING.\n",
      "2.5 STD above mean. 1.1 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3605 tweets for Wordle 333.\n",
      "\n",
      "10104 score patterns. 142 unique.\n",
      "\n",
      "91.28%, (136/149) valid final guess patterns found. Impossible pattern count: 6.\n",
      "\n",
      "Wordle 333 prediction: SCOUR.\n",
      "2.4 STD above mean. 1.12 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3452 tweets for Wordle 334.\n",
      "\n",
      "11150 score patterns. 131 unique.\n",
      "\n",
      "Wordle 334 initial signal low 1.08. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -9.00E+07\n",
      "89.29%, (125/140) valid final guess patterns found. Impossible pattern count: 6.\n",
      "\n",
      "Wordle 334 prediction: GLASS.\n",
      "2.4 STD above mean. 1.11 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3181 tweets for Wordle 335.\n",
      "\n",
      "10664 score patterns. 150 unique.\n",
      "\n",
      "Wordle 335 initial signal low 1.06. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -7.00E+07\n",
      "86.67%, (143/165) valid final guess patterns found. Impossible pattern count: 7.\n",
      "\n",
      "Wordle 335 prediction: GAMER.\n",
      "3.4 STD above mean. 1.14 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3526 tweets for Wordle 336.\n",
      "\n",
      "10304 score patterns. 148 unique.\n",
      "\n",
      "Wordle 336 initial signal low 1.1. Iterating for better parameters\n",
      "..............................................................................................\n",
      "Iterated to a better signal with min_count 3 and penalty -7.00E+07\n",
      "91.39%, (138/151) valid final guess patterns found. Impossible pattern count: 10.\n",
      "\n",
      "Wordle 336 prediction: SCRAP.\n",
      "2.9 STD above mean. 1.11 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3555 tweets for Wordle 337.\n",
      "\n",
      "10568 score patterns. 145 unique.\n",
      "\n",
      "88.46%, (138/156) valid final guess patterns found. Impossible pattern count: 7.\n",
      "\n",
      "Wordle 337 prediction: MONEY.\n",
      "2.6 STD above mean. 1.16 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3513 tweets for Wordle 338.\n",
      "\n",
      "10729 score patterns. 133 unique.\n",
      "\n",
      "89.51%, (128/143) valid final guess patterns found. Impossible pattern count: 5.\n",
      "\n",
      "Wordle 338 prediction: HINGE.\n",
      "2.0 STD above mean. 1.14 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3525 tweets for Wordle 339.\n",
      "\n",
      "10987 score patterns. 137 unique.\n",
      "\n",
      "91.97%, (126/137) valid final guess patterns found. Impossible pattern count: 11.\n",
      "\n",
      "Wordle 339 prediction: ALBUM.\n",
      "2.5 STD above mean. 1.12 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3321 tweets for Wordle 340.\n",
      "\n",
      "12268 score patterns. 100 unique.\n",
      "\n",
      "Wordle 340 initial signal low 1.07. Iterating for better parameters\n",
      "..............................................................................................\n",
      "Iterated to a better signal with min_count 3 and penalty -1.30E+08\n",
      "85.29%, (87/102) valid final guess patterns found. Impossible pattern count: 13.\n",
      "\n",
      "Wordle 340 prediction: VOUCH.\n",
      "2.1 STD above mean. 1.11 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3547 tweets for Wordle 341.\n",
      "\n",
      "10584 score patterns. 140 unique.\n",
      "\n",
      "Wordle 341 initial signal low 1.0. Iterating for better parameters\n",
      "...............................................\n",
      "Iterated to a better signal with min_count 1 and penalty -1.30E+08\n",
      "87.33%, (131/150) valid final guess patterns found. Impossible pattern count: 9.\n",
      "\n",
      "Wordle 341 prediction: ASSET.\n",
      "2.7 STD above mean. 1.11 above runner up.\n",
      "\n",
      "TwitterWordle analyzed 3516 tweets for Wordle 342.\n",
      "\n",
      "9690 score patterns. 142 unique.\n",
      "\n",
      "83.83%, (140/167) valid final guess patterns found. Impossible pattern count: 2.\n",
      "\n",
      "Wordle 342 prediction: TIARA.\n",
      "2.0 STD above mean. 1.12 above runner up.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "t_full_dictionary.solve_all(mask_result=False)"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "d71e3fad4d7ef819ec5614b52b98e2aaeb2a4012f1049388f932022bc4f151d3"
  },
  "kernelspec": {
   "display_name": "Python 3.9.6 ('base')",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.6"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}