If any non-bot participant placs one or more bets on YES that total more than Ṁ40, I reserve the right to resolve the market to either N/A or to NO, in whatever way best punishes the user(s) who placed those invalid bets.
All my markets of this type with various maximum mana investments:
1,000
3.00Based on a quick glance through the trades tab, Tassilo Neubauer and Market Maker are the only two accounts to have placed individual bets on YES greater than M$40. Tassilo sold out at a loss and Market Maker switched to a NO position, so it seems I can safely resolve YES based on that.
However, it's possible that someone placed multiple smaller bets that totaled to an invalid position. (For example a M$40 bet on YES and then another M$40 bet on YES later on.) I won't have caught those, since I can't remember everyone's username as I'm looking through the trades tab.
Seems likely I'm safe to resolve this to YES regardless though, any concerns before I do so?
Basic scripted API check results – no guarantees, but it passes sanity checks. Bots and NO (net) positions were filtered out, and only YES accounted. How much mana did each group bet in total:
`valid YES bets?` `how much at stake`
<lgl> <dbl>
1 TRUE 1327.How many people are on each side:
`valid YES bets?` `how many people`
<lgl> <int>
1 TRUE 45Transgressors (none):
character(0)Another easy one IMO, can resolve YES right away, because no one broke the rules. @IsaacKing
Alright, here's what seems fairest to me:
Get a list of everyone who placed invalid bets on YES.
Check to see how much mana those users would collectively lose from a YES or NO resolution. If they would lose the most from a NO resolution, then:
Get a list of everyone who did not place any invalid bets on YES.
Check to see how much those users would collectively lose from a NO resolution.
If the invalid bettors would lose more than the valid bettors, resolve to NO.
Otherwise, resolve to YES.
However, that also seems like it would take a lot of work. I'd have to write a script to download the trading history and calculate who placed invalid bets and how much they stand to lose each way, which would take a while especially given that I'm not very familiar with Manifold's API. I don't want to lose several hours to this, so I'm also thinking about doing something simpler like just resolving to YES. (Which is allowed per the description.)
Thoughts?
@IsaacKing I think most of the invalid bets are people like me (currently 48M yes) who placed a buy for 40M cost- my position is 48M I guess but I was never putting more than 40 at risk (which is how I interpreted the market, incorrectly)
Maybe someone else is jeopardising everything for 15M but it doesn’t seem likely to me
@IsaacKing No non-bot participant 'placs one or more bets on YES that total more than Ṁ40'. The market can safely resolve to YES.
@Gen Holding more than 40 YES shares doesn't mean your bets were invalid. Bets don't convert mana to shares at a 1:1 rate, you always get more shares than the mana you spent. (That's why it's profitable to place bets in the first place.)
@IsaacKing I guess my confusion stems from the title saying “max position 40M” rather than 40M cost basis. Description makes it sound more like 40M cost basis is the expectation
I don’t see any real harm done here by “invalid bets” if the latter was your expectation. Who are you trying to punish by delaying resolution?
@IsaacKing Although some non-bot participants did place one or more bets on YES that totaled more than Ṁ40, they did not, however, plac one or more bets on YES that totaled more than Ṁ40.
@Gen Ah, good point, I guess "max position" was poor phrasing. I did mean cost basis, since that's how all the other markets like this had been run, and it would be much more awkward to police share amounts, since people make their betting decisions by cost basis. (e.g. the default M$10 cost basis for tapping the betting arrow.)
