I have tracked my wins and losses in a video game I've been playing called openfront.io. It looks like more than statistically chance would predict, I win games in streaks; a large % of my wins appear to be part of a streak where I win multiple games in a row (almost always just 2, but once i won 3 in a row).
Will o3, when I ask it based on the win loss data I have, judge that for the 62 games on which I gave it data, there's this statistically significant pattern where i win more often as part of winstreaks than random chance would predict? (ie if games were independent). I have seemed to improve across those 62 games so I am worried that there might be some kind of effect where a statistical test doesn't take this into account and that makes the results false, but for simplicity this market will resolved based on this simple prompt and I encourage people to let me know if they think the conclusion o3 comes to is wrong or wtv.
Update 2025-07-08 (PST) (AI summary of creator comment): The creator has shared the results of the analysis that will determine the market's resolution. The analysis yielded a p-value of 0.048.
1,000🏅 Top traders
| # | Trader | Total profit |
|---|---|---|
| 1 | Ṁ706 | |
| 2 | Ṁ23 |
People are also trading
@4fa it seems hard to not miscount at places if i tried doing this, but here is how i stored the data:
#games, #wins
1,1
2,2
15,3
18,4
19,5
22,6
39,7
40,8
41,9
45,10
50,11
55,12
59,13
60,14
62,15
63,16
I would note a few things:
significant at 0.048. lol, lmao even
the question we had verbally agreed to bet on, which you said we would bet on manifold about and which you were 80% about, was that your pattern of win streaks was unusual
o3 agrees, upon being given your pattern of win streaks alone, that that pattern alone is not statistically significant (p=0.32). as in, if you tell it: "over x games i had 1-win streaks y times, 2-win streaks z times, etc".
it manages to extract statistical significance from your full win-loss streaks. but we did not agree to bet about loss streaks, because it was clear to both of us that they were not a fluke; you indeed play like a brick when you're tired and pissed. but i was skeptical, and you were overconfident, that it was so unlikely to hit so many 2 and a 3 streak with a 24% winrate over this many games
even on the lie that is this question, i win in expectation
@BionicD0LPH1N this is incredibly misleading. i agree i bet wrong. i don't remotely agree with the framing