![scribb lio scribb lio](https://data.pacogames.com/images/300x300/skribblio.jpg)
![scribb lio scribb lio](https://i.pinimg.com/originals/46/3f/f8/463ff89bdab0c8b76aca726c6a86080f.jpg)
In reality, there are several other considerations that are involved in the game: a good knowledge of the type of words that is used in the database of such games (or at least have a good vocabulary that includes the word that you guess). This means that you should draw something that is easy to understand for most people in general (for example, using simple shapes and colors).įor the guessing people, the above analysis shows that if you want to guess right, the only thing that you can depend on in this setting is Pr(word), i.e. While Pr(picture) does not affect the choice of word, increasing this quantity together with Pr(picture|word) is in general helpful, since this would provide larger margins between the Pr(word|picture) for different words, and help the guesser make decisions.Increase Pr(picture|word), which means that you should tailor your drawing so that it reflects the key characteristics of the particular word.Thus, we see that for the drawing person, the best strategy to have people guess the right word involves the following: This again d epends on the style of the drawing person – whether he draws something that is unique to himself or accessible to most people. Pr(picture), the probability that the picture is drawn.This somewhat dependent on the guesser’s knowledge of word frequency, and the particular context of the word (if it is given).
![scribb lio scribb lio](https://pics.me.me/o-skribbl-io-free-multiplayer-dra-o-following-twitch-67055409.png)
Pr(word), the prior that indicates the frequency of the word.This is solely dependent on the style of the person who draws the picture. Pr(picture|word), which is the probability of the having this picture when given the word.We assume that the guesser don’t have any knowledge about the style of the person who draws the picture (this may not hold if you’re very close friends, or if you have already played the game for many rounds).īayes’ Rule states that this maximum depends on the following factors:.there is no additional interaction between the two sides beyond the picture drawn. We assume everyone follows the rule, i.e.Then a simple application of Bayes’ Rule states thatĪt each point, our best guess is to pick the word that has the largest LHS in the above equation. Let “word” and “picture” represent the current word we want to guess and the picture that is already drawn. In this blog post, I will try to provide an alternative interpretation of the game in terms of the Bayes’ Rule learned in this class (as this might give some clue to how AI guessers may work for this game). The goal for both sides is to guess (or let other people guess) the word correctly based on the picture. The other players then try to guess the word, given the length of the word and the picture. The idea of this game is that one of the players is given a word, and the player draws a picture that illustrates the word. Recently, the online game skribbl.io has been used in various social events. Bayes’ Rule and the Strategy for Winning in skribbl.io