@mods - I don't use Discord - please delete this question/refund all traders accordingly, not concerned with the fee being refunded or not as long as betters are made whole.

seems like a clear yes, right? It just ranked higher

[potentially wrong]

@jacksonpolack @ian I'm not sure I follow. Claude 3.5 Sonnet only just a few minutes ago was added to the leaderboard, and it is not higher than GPT-4o. It may rise, but the issue is that the market was very ambiguous and the creator @vVv has said they would N/A.

If so this is why I always re-check markets before resolving them, lol!

I still think it's better to read it as 'will 3.5 ever rank higher than gpt-4o on the leaderboard' rather than nullifying though? Not sure

@jacksonpolack I think the issue is that since vVv said they would N/A, so people bet based on that. I would N/A. (Personally, my intuition is that the question means elo rank at launch.)

Oh I thought someone said it ranked higher already, whoops! Nullifying is still bad precedent and should be avoided when possible, which it seems very easy to avoid in this case, (it hasn't ranked higher and given the confidence interval, doesn't seem like it's going to).

@ian - yeah I agree, not a fan of N/A but just for clarification - what I had always intended was for this to read as "Will 3.5 Sonnet rank higher than GPT-4o..." in terms of when it happens, agreed this bit wasn't clear but I would have used the question closing time. If 3.5 Sonnet made it to the leaderboard and beat GPT-4 on Day 1 it would have resolved as YES and if 3.5 Sonnet made it to the leaderboard but not placed higher I would have waited until the end time and resolved NO if it had not placed higher by then.

My reasoning for N/A was two-fold - one, a question with much higher liquidity already existed (this didn't show up during my question creation suggestions) and as there was already ambiguity from users - hence the u-turn.

Anyway - I would much rather a mod resolve this on my behalf so there can be no accusations against me - for the record, as of this post Sonnet 3.5 has not/did not place higher than GPT-4o

The question is open until Dec 31 anyway - there's a chance it could go the other way as well but I am also conscious of those that traded based on it being N/A or any other conversations in this thread - I'd like to reduce any impact on traders due to this

Gotcha. I still think resolving NO makes sense even with the ambiguity, but I'll honor your wishes here.

When will this evaluate seems quite critical.

The other question here seems more robust.

@vVv Is this market asking whether claude 3.5 sonnet will outrank gpt4o at any point in 2024, or as soon as it's added to the leaderboard, or somethning else?

Okay - use this one then.

https://manifold.markets/Moscow25/where-will-anthropics-claude-35-son

@Bayesian it was for the former - if it outranks GPT4o at any point. Will be nullifying this question anyway so recommend not betting :)