Resolves YES if before the end of 2023, two or more bots make at least 10 trades against each other in quick succession.
The time between consecutive trades must be at most 60 seconds and there can't be any human bets between them.
Limit orders count.
The bots have to be general-purpose bots, such as BotLab, v, acc, William Hagraves, ArbBot, quadrilateral, etc. Bots which are meant to trade on a specific market, manipulate a specific market (including this one!), exploit a bug on manifold, or exploit another bot do not count. The bots have to be reacting to each other and not to some external stimuli.
Multiple consecutive trades by the same bot count as one. For example this counts as 4 trades:
Related questions
🏅 Top traders
| # | Name | Total profit |
|---|---|---|
| 1 | Ṁ603 | |
| 2 | Ṁ350 | |
| 3 | Ṁ59 | |
| 4 | Ṁ37 | |
| 5 | Ṁ21 |
Hi everyone, @Yev delegated this market's resolution to @Austin who re-delegated it to me. Although I have a large stake in this market, I'm not going to re-re-delegate ownership until we come to a truly messy edge case.
Where last we left off, @MarcusAbramovitch was complaining that this market should resolve due to strings of bot trades by arbitrage bots, and I was disagreeing on the grounds that the following line in the resolution criteria:
The bots have to be reacting to each other and not to some external stimuli.
seems to clearly exclude trades by arbitrage bots that are responding to movements on another market rather than to each others' trades. If there's a contingent of people other than marcus who dispute that ruling, say your piece now and I'll go ahead and re-re-delegate immediately. In the meantime, I've asked some follow-up questions for @EliLifland and @Eliza.
@MichaelWheatley I think it depends on the specifics of the arb bots, if they were only trading back and forth on markets a and b it seems like it should resolve yes because they are still reacting to each other. But if they are both only both trading on the market a in reaction to other traders on market b it should resolve no. I'm not sure about the example Marcus sent because I don't see any 10 consecutive trades in that market
@EliLifland was just combing over that example. Yeah, an arb bot reacting to human trades on a second market doesn't count, but IMO, an arbbot reacting to bot trades on a second market does, because it's still reacting to another bot.
@Eliza There's been several examples. The problem is just that the market creator refuses to come on and resolve the question
Just got a chain of 14 here: https://manifold.markets/ButtocksCocktoasten/will-there-be-a-wellrecognized-lett-2ad90b30e3f9
Pretty sure this has already happened many times, but one example is on https://manifold.markets/NFL/will-the-kansas-city-chiefs-tomahaw when @EliBot @acc @Botlab had 18 trades in a row, due to an unintentional bug in my bot where it was supposed to be arbing against (1 - https://manifold.markets/KevinBurke/nfl-week-4-oct-1-will-the-new-york) but actually was arbing normally against it (if trades due to bugs aren't allowed I'm sure another could be found, but it doesn't say it's disallowed in description). It also says "meant to trade on a specific market" isn't allowed but I assume this doesn't exclude bots which arb specific groups of markets (which mine does) since ArbBot is included in the description.
@EliLifland Just realized the group of 18 included me manually selling to undo the bug, but there’s a group of 14 starting at 1:09:31 ET at https://manifold.markets/KevinBurke/nfl-week-4-oct-1-will-the-new-york that
is all bot trades and spans 40 seconds total which should count
bug in my bot where it was supposed to be arbing against (1 - https://manifold.markets/KevinBurke/nfl-week-4-oct-1-will-the-new-york) but actually was arbing normally against it
Could you clarify this part?
The disagreement between myself and Marcus is that I read
The bots have to be reacting to each other and not to some external stimuli.
as excluding arbitrage bots in the sense that if you arb another market 5 times and acc reacts to you 5 times, thats not a ten-trade chain reaction: you weren't reacting to acc but to trades on a different market.
@MichaelWheatley
Yeah, sure. Looking at my logs / the trades on the market in question:
1. For my bots first two trades, for some reason that I'm still not sure it was trying to bet the markets to the same price but failing, must be something wrong with my bet size calculation. Then on the third trade it succeeded in bringing the market to the same price. I think these should count because my bot only reacts to other user's trades and not its own so acc's trades prompted it to trade again; technically if acc had only traded in the other market it still would have prompted my bot to trade again in the market in question, but the same is true of acc only trading in the linked market so it seems to me that it should count.
2. My bot's fourth and sixth trades were directly caused by @Botlab's trades
3. My bot's fifth trade was caused by a combination of acc's trades on both markets, so similar situation to (1).
4. Acc had the last trade which was presumably caused by me
So overall (2) / (4) seem clear-cut, and if only (1) doesn't count there are still 10 trades in a row, so it basically hinges on whether (3) counts. Looking at the trades I'm guessing my bot wouldn't have traded if acc had only traded on the market in question, and also wouldn't have traded if acc only traded on the other market (I could do the math if needed). Personally I feel since with without acc's trade on the market in question my bot wouldn't have traded that counts as reacting, but it is a bit ambiguous and I could see arguing it has to be solely reacting to the trade on that market and nothing else can be affecting it.
Anyway, if this doesn't count I could write a script to pull candidates and I'm guessing if people manually went through them they'd find one that counted, at least by the end of the year.
@MichaelWheatley It's on the last page (5), starting with these trades. The first one is at Sep 29 1:09:31 ET
@MarcusAbramovitch Is that the policy? I thought we just waited until EOY and do an admin resolve.
I can DM Yev on Discord asking for an ruling if there's an edge case.
also, I'm just going to leave this here...
@MarcusAbramovitch I got a response form Yev who said they'll look into delegating market ownership to someone else when they're less busy.