Hi folks!
I apologize for not blogging much lately; the Canadian Media Fund deadline is coming up so I’ve been slaving away on an application for government support that will help pay for the giant advertising and PR campaigns that will be accompanying Prismata’s release. (Exciting, I know….)
In the interim, I have two small things for you:
(1) A quick survey on what we should do with the “Good Game” emote, which has recently been removed from Prismata. It should take less than two minutes to complete.
(2) I want to run a quick little “Guess who wins” contest, inspired by a few recent threads on our subreddit.
The “Guess Who Wins” Challenge
The challenge: How well can you estimate the amount of player advantage in a Prismata unit set?
When: After the grand prix stream on Saturday; probably starting between 7pm and 8pm EST.
Where: twitch.tv/lunarchstudios
Prize: This is just for fun, but to up the stakes a little so you’re all trying your hardest, I’ll provide the winner with a code for 650 shards.
Rules
- There will be 30 rounds.
- During each round, the units from a Prismata replay between two 1900+ Master Tier players will be shown (these replays are randomly chosen from all games played since the Savior patch).
- Players will then be given an opportunity to give an estimate of what level of player advantage they believe is present in the set. For example, a guess of “0.6” would represent “I think that player 1 has a 60% chance of winning”.
- Players will input their guess into a google docs spreadsheet. After a time limit of a few seconds, all player guesses will be copied into a separate read-only scoresheet.
- If a player makes a guess p, they will receive points according the following formula:
- If player 1 wins, you gain 1+log(p)/log(2) points
- If player 2 wins, you gain 1+log(1-p)/log(2) points
- Note: you can LOSE points by guessing wrong.
- Highest score after 30 rounds is the winner.
Notes on the scoring method
This scoring method gives one point per guess, minus one point per “bit of surprisal” induced by the outcome of the game (effectively, it’s the amount of “entropy loss” that the revelation of the winner adds to the system). Equivalently, it’s how unlikely the result should seem to you, measured in units of “coinflips”. For example, if you think that player 1 is 75% likely to win, a player 2 victory results in -2 points (for a total of -1 points) because that result is, to you, as unlikely as losing two consecutive coinflips.
This method carries the advantage of encouraging to vote what you truly believe. For example, if you were betting on whether a 6 would be rolled on a single 6-sided die, betting 1/6 = 16.666 leads to the lowest expected losses (exercise for the mathematically inclined: prove it!)
A table of some sample scores is given below:
| Bet | score if p1 wins | score if p2 wins |
| 0.005 | -6.64385619 | 0.9927684308 |
| 0.01 | -5.64385619 | 0.9855004303 |
| 0.05 | -3.321928095 | 0.9259994186 |
| 0.15 | -1.736965594 | 0.7655347464 |
| 0.25 | -1 | 0.5849625007 |
| 0.35 | -0.5145731728 | 0.3785116233 |
| 0.5 | 0 | 0 |
| 0.65 | 0.3785116233 | 0.5145731728 |
| 0.75 | 0.5849625007 | -1 |
| 0.85 | 0.7655347464 | -1.736965594 |
| 0.95 | 0.9259994186 | -3.321928095 |
| 0.99 | 0.9855004303 | -5.64385619 |
| 0.995 | 0.9927684308 | -6.64385619 |
Participating
This is only intended to be a fun little experiment. There is no need to register; just show up in the stream at twitch.tv/lunarchstudios and links will be provided in the Twitch chat. Participation is open to everyone.
The 4th Grand Prix round is upon us!
See you Saturday for round 4 of the PAWC Grand Prix! If you want to participate in it, please sign up if you haven’t already done so!
