🏅 Top traders

Looks like there was not a bigger whale!

Looks to me like this should resolve to NO, based on Jack's M$100,000 limit order in that direction. Please confirm before I do so.

I guess there isn't :)

@Yev Lol, you win because I misread the close date. 😛

Wait a second, I knew I forgot something - this doesn't close yet:

To prevent sniping, I'll leave this market open until 2 days have passed since the current largest bet. (I might forget to change the close date proactivly, in which case I'll just reopen it for 2 days after it closed.)

@jack Curse you, the Show More button! 😠

@jack I also misread the title. It said it closes at 2:00 am, so I assumed that was 2 hours after midnight. It was 3 hours after midnight.

@Yev Lol... curse DST

Reopened for another 2 days. :)

I'll just reopen it for 2 days after it closed.

Does that mean that after it closed you'll reopen it for 2 days? Or that you'll set the close date to 2 days after the original close date?

@Yev Ah you're right, that's unclear, thank you. I meant that I would set a new close date such that the market is open for another 2 days.

Basically, I want everyone to have the chance to place any bets they want to place.

@IsaacKing Wouldn't it be more fair to have it open for another 2 days - 13 hours, since it was already open for 13 hours between my bet and market closure?

@Yev That would be more in keeping with the "2 day antisnipe window" concept, but it's not what I had said in the description, so I don't want to change it now.

I didn't think anyone would care about the exact amount of time it's left open, so I didn't think about it very hard when writing the description. Is there a reason the length of the antisnipe window matters more than I think it does?

@IsaacKing The more time people have to make a larger bid, the more likely someone is to do it.

@Yev Sure, but a difference of a few hours seems like its effect on that probability should be negligible, no?

@IsaacKing Probably

Nothing's happened for a few days, let's shake things up. :)

For limit orders, do you count however much was filled?

@jack Good question, I forgot about limit orders.

I don't fully understand the way Manifold displays limit orders in the trades tab, and I want to make sure whatever answer I give makes it relatively easy for me to figure out who won. Let's say that Alice places a limit order of M$100 on YES. M$30 of it is filled right away. Another M$20 is filled later when Bob bets on NO. Then Alice cancels the remaining M$50. How will that display in the trades tab?

@IsaacKing I think it'll display as M$50 out of M$100 (cancelled)

@Yev So there's no way for me to see that it was filled in two parts? It'll display identically to the case where all M$50 of it was filled right away?

I believe that info is available in the API but it is very hard to see on the website.

I would suggest any of the following rules could potentially work:

• Don't count limit orders

• Count the entire filled amount of the limit order as a single bet, regardless of whether it was filled in multiple parts.

(Where by "very hard to see on the website" I mean the website doesn't actually show it anywhere, you'd have to piece it together by trawling through the bet history.)

The problem with your second option is that it's impossible to know when a limit order was placed or filled, so if there's a tie I won't know how to resolve it. (The spot it displays at in the chronological bet history is the time at which any portion of it was first filled.)

But a tie is unlikely enough that I can just say I'll count it as having occurred wherever it shows up in the bets tab, and I don't think that'll cause a huge problem.

Another issue is that someone could place a large limit order and then fill it themselves. I think this would let them guarantee negligible losses, which is not possible via placing a large bet and then selling it right afterwards. (Since a bot could bet in between.)

Yeah I don't think ties are an important issue. You could even randomly resolve to any of the tied bettors or something, as long as there is some rule I don't think it's going to actually be a problem.

The latter is a problem no matter whether you count limit orders or not - I already considered this idea myself.

@jack Oh right. Ok yeah, let's go with your second option. Limit orders count as however much of them was filled. Ties determined by position in the trades tab.

@IsaacKing Hmm, thinking about it some more, I suspect there might be ways to exploit that somewhat. So far I haven't thought of any that would actually break the market, but maybe someone will. Might be safer to just exclude them, unless someone else can think of a reason that's not good either.

@jack I mean, this market is for fun anyway, so if someone figures out an exploit, just go for it. :)

@IsaacKing :) works for me

Bob places a M$30 limit order on NO.

Carol bets on YES, filling M$20 of Bob's limit order.

This example would resolve to YES.

I think this could be either YES or NO depending on which bet was bigger. It's possible for it to be YES if Carols' bet was M$30 and filled M$20 of Bob's order. Or it could be Carol betting M$1 and filling M$20 of Bob's order. It depends on the price and whether they also traded against AMM or another limit order.

@jack Is your comment meant to point out a problem with the market, or is it just an example? I don't see an issue with what you've described.

@IsaacKing It's your example in the market description. I think "This example would resolve to YES." is wrong

@jack Oh, I see what you're saying. Whoops.