1) Profits and losses from trading on your own markets no longer count towards your league's earnings. This was to reduce the number of people winning their leagues by creating a new market with significant liquidity and correcting the probability from the initial 50%.
2) Unique trader bonuses are back for leagues! Any unique trader bonuses gained from a new trader betting on your market since the season start date will count. However, to keep things fair, we will check if the market was created within the last 31 days of the bonus and only count those that are. This keeps market creation more in line with the rules of trades having to be within the given month; however, it doesn't create weird incentives of waiting to create your market.
Next league season we'll count trades on one's own markets 1hr after creation: https://manifold.markets/post/next-league-season-well-count-trade
Two proposals:
- Trickle liquidity in at market creation, so that the creator can bet it up or down to whatever probability they want at market creation without "making fake profit by stealing market liquidity". and if they wanna wait 20 minutes before betting, anyone else can see the market and bet before the creator does. Disadvantage: for markets that are meant to last 20 minutes, this sucks. maybe there are ways around this. Maybe: when exceptions are needed, liquidity added by moderators doesn't trickle bc they are trusted not to abuse this? maybe this idea needs workshopping, open to feedback
- simple creation toggle to set prob at market creation, sends the same request to the same market creation endpoint but with that extra arg. I am not a very competent dev but to my badly trained eyes this seems simple enough. if not i am curious to know why not. maybe complexity with mc and it's weird if a setting exists for binary but not mc?
Is it possible to set the initial probability to anything other than 50% ?
Not sure what the alternative is.
@Pete6Eucc yes for binary via the API, although it is very inconvenient so few ever do it.
No for MC. In theory it’s possible but it would be a lot of work to implement (it’s not just a quick UI change)
@Pete6Eucc Used to be easy, fucked with the liquidity so it was disabled (outside the API apparently) years ago.
TL;DR: The premise of these changes is good, but in practice, it’s very bad to entirely exclude profits from creators trading on their own markets. This penalizes users from making good markets on topics they want to trade on, and encourages them to instead wait & hope “someone else does it”. It's important that Manifold not penalize market creation.
Solution: Only exclude profits from trades placed within X days of market creation (I’d suggest 1d, but you could also do 30d to match the trader bonus cutoff [1]). The issue for leagues is when creators trade first against their own liquidity [2]—after X days, that's no longer relevant.
Explanation: I’ll use my past Manifold usage as an example (I’m less active now so it probably wouldn’t impact me much, not sure). I like to trade & predict the Oscars. Last year, I created many markets about the Oscars, so that I would have more opportunity to trade. Under this system, if I care about leagues, it’s actively bad for me to create Oscars markets—instead I should hope that someone else does, so that my trades can count for leagues. I am confident that last year (since I cared somewhat about leagues), that would have made me create fewer Oscars markets.
E.g., this week I invested a bunch of liquidity in this market for the best picture nominees. After creation, I immediately placed bets to initialize the market at a suitable price. The premise of this change is that those trades shouldn't "count" for leagues, and that's absolutely correct (my only goal was to preserve liquidity, it wasn't real trading). But now that the market is created, and others are trading on it on an even footing, that's no longer relevant. Under the new system, my reward for investing liquidity in a topic I care about is that for the next 4 months, none of my trades on that topic will count for leagues. Instead, I should have waited for someone else to create the market (and in this case I'd want to see if it's possible to get that market N/A-ed so an identical market created by someone else can take its place, even if it has less liquidity than what I was willing to pay for—very bad incentive!).
It is bad if Manifold punishes you for creating markets on topics you care about. This issue would be solved (while still achieving the original objective) if only trades from the first day after market creation were excluded.
[1] Even better would be something that just excludes the trades that initialize a market, but I’m starting with the simple version
[2] There's also the separate issue of creator bias in resolution. I'm setting that entirely aside, because that's not the stated reason for the change (and if that's the issue you'd want to block creators from trading in their own markets entirely).
Well said. I think the creator bias problem (as much as it even exists, I think most users operate in good faith) is best handled by making ratings more visible.
I saw an API field for initial probability when making markets. Perhaps expose that in the UI for honest market creators? We could also limit trading by the creator for the first hour or so to limit the effect and give traders a chance at mispriced markets before the creator.
@benmanns initial probability was formerly available in the UI, but my understanding is that it had to be disabled, due to exploits surrounding initializing markets far away from the true value and then "correcting" them. Initialization at 50% is the lesser evil.
but my understanding is that it had to be disabled, due to exploits surrounding initializing markets far away from the true value and then "correcting" them.
i'm fairly confident this isn't true—the ability to trade into the initial liquidity is equivalent whether the market starts at 5% or 50%.
the primary issue is that the site isn't particularly close (AFAIK) to this feature being available for MC markets. so while it exists (& works fine) for binary, i think the site is less likely to complicate market creation if the same feature isn't available for MC markets.
@Ziddletwix to clarify, the issue isn't just that the UI is more complicated for MC (due to all the different options). what's good about initializing binary markets at a different target probability is that it's more efficient, due to maniswap. this math doesn't exist atm for MC (dependent MC because it's more complicated, independent MC because it's not actually configured properly as a set of binary markets). it would be actively bad atm to allow independent MC to be initialized far from 50%, because much of the liquidity would be wasted (the mana would just get burned, same as if you inject liquidity into independent MC that are currently far from 50%!).
this is a fixable issue—independent MC should in theory function just like binary—but practically speaking, it would require more than just creating a new UI.
@Ziddletwix I can see your point from a creator's perspective. The opposing view: Creators shouldn't even be allowed to trade in their own markets. Add to that Manifold's guideline saying that creators have (almost) full control over their markets. Merely not counting their own markets against leagues is still very far on the pro-creators side in my opinion.
@Primer this is a not uncommon view, although my argument against it is largely the same as the above—as someone who has made many popular markets, I would have been substantially less likely to create those markets if I were barred from trading in them. To me, any concerns about conflicts of interest (which will arise regardless of whether creators trade) are a rounding error compared to concerns that people will have less reason to engage with the site. But some people feel differently about that trade off and it’s not an issue that people are likely to change their mind on.
I can only state with confidence that if creators could not trade in their own markets I would not have created most of my markets (and I do not think I am the only person on the creator leaderboard who feels that way). For this change to be worth it, you’d need to offset that loss in activity. In practice, it’s not obvious that most traders actually care that much (or else they would refrain from trading in markets where the creator is trading). But I do take seriously that a small minority feel strongly about this.
@Primer This might be beside the point, but I am very confident that if creators were never allowed to trade on their markets from the start, the site could not have gotten as popular as it is today. A huge portion of markets, especially the niche ones that keep various subgroups active on the site, feature creators making markets that they want to trade on.
Whether that has to be the case into the future is certainly up for discussion. I'm sure there is some version of repositioning the market creator role where that idea could succeed.
@Ziddletwix Makes sense to me, I support this! Would it be better to just exclude trades after the first hour or first 5 minutes? Better to let the creator start trading for real as fast as possible, and it doesn't take that long to correct the prices, right?
i'm fairly confident this isn't true—the ability to trade into the initial liquidity is equivalent whether the market starts at 5% or 50%.
Manifold instituted a workaround (a long time ago I believe) that makes the particular exploit I am referring to infeasible, so they probably could bring the initial probability option back for binary and unlinked MC.
I believe unique trader bonuses being back could be a good incentive for new users to create markets, since they need little mana to promote and they could get it that way. It's not great for having an array of diverse market topics since they'll easily figure out AI/Trump/Crypto is what gets the most traders. I think more diverse topics could attract more users. About a month ago I somehow (probably someone linking the market outside manifold to some community) got a bunch of new users that were made just to trade on a manga market, it was through a topic not very popular in manifold (other than One Piece) that the site got new users.
Profits and losses from trading on your own markets not counting toward leagues makes sense from the "correcting the new market odds for easy mana" angle, but it doesn't make sense once the market is established, that's to say, once the first non-bot non-creator trade has been placed. At that point, assuming it's not a personal market or otherwise subjective, it should be fine for the profits and losses to apply to leagues.
That said, profits and losses from trading on your own markets still count towards your overall profit, so it doesn't matter that much (opinions may vary), but I think my point about market diversity still stands.
TL;DR: I agree with @Ziddletwix that these measures penalize market creation, but the trading bonuses thing seems more worrying to me.
Can the admins do anything about the long-term markets where the odds say that I'm wrong, but I know that I'm right?
Praised be the announcement post!
Great to see you've found a way to inform Manifolders about what's changing on Manifold!
#2 is extremely biased in favor of well-established creators. My starship-related market (an extremely popular topic here) took almost a month to get just 10 traders, plus creating markets is very expensive if you don't have a massive net worth. This problem is exacerbated by #1, since making markets is less worth it if you can't capture the initial correction from 50%.
Before this change, I was actually beating @bens in our Diamond league. Now that his trader bonuses count, he's beating me.
As-is, I wish I was able to downvote this post, even though I appreciate the value of its goals.
Agreed that it's a problem, but it's mostly unrelated to leagues. We should find other ways to help new users create more enticing markets (and then highlight them when they are good).
100 mana to create a binary market means you just need 10 traders to break even. Less if you trade on it initially, or it closes before converging to the answer. I don't think this is a particularly unfair advantage. Many older users are willing to gift new users mana to create markets (Manifold is too!). If you feel like your balance is blocking you from making more markets, let me know and I'll fund some.
100 mana to create a binary market means you just need 10 traders to break even.
trader bonuses for 100m markets should be 3m, so 34 traders to break even (10m for 1k liquidity markets, but those require 100 traders).
that's not a problem—the goal of making markets should not be to necessarily earn mana, you are "paying" for liquidity"—and not that relevant IMO to leagues but just flagging the math.
It was unrelated to leagues up until this change. Big creators get more traders on their markets, and so get bigger trader bonuses. Worse, higher-liquidity markets get bigger bonuses and attract more traders, so the advantage compunds. Trader bonuses are currently way too biased towards established creators with lots of mana to be fair for inclusion in leagues right now.
@SirSalty I appreciate the offer for market creation financing, I'll try to keep that in mind.
100 mana to create a binary market means you just need 10 traders to break even.
This is a suboptimal statement of the incentive to create markets. Markets should not be created purely for engagement farming, they're for finding the answers to questions about the future. Framing it like this only reinforces the idea that you 'lose' the mana you used to create the market.
The reality is you fully control your liquidity in a market and can close trading or remove liquidity at any time. You only pay for the information you gained when traders disagreed with your initial guess at the answer.
When I make markets with 10,000 liquidity I don't need any number of traders to break even -- I control the liquidity completely and manage it just like I would manage any other investment on the platform. Overall, I've found that traders often do a poor job of predicting and in many cases I have gotten back more mana from the pool than I put in to such markets.
@Eliza The ability to remove liquidity doesn't apply to all market types, but I appreciate your perspective on how best to look at the "cost" of market-making.
@SimonWestlake Don't worry, I've also done my fair share of kicking and screaming about that specific point 😂
Ooh yay!!! I got a big chunk of my profit in last month's league from exactly this and it felt weird 😅, will be glad to see it gone
It does create a weird, niche disincentive. Newly trading on a league competitor's market gives them a trickle of mana that might be enough to dethrone you in some close races.
Would you consider modifying rule 1 to not apply to markets created before the start of September? This is a massive bummer for one of my markets, and it does not give me any unfair advantage.
Better yet, make this change always not apply to markets older than the current month, except for unlinked multi-choice markets (which would provide a massive loophole). A buffer might be a good idea (markets created a day before the current month wouldn't count either)
@SimonWestlake I’d +1 to this suggestion. It’s a good idea to remove the early bets that initialize the probability from leagues. But it’s unfortunate to penalize creators for making popular markets they want to trade on. Ideally you’d just exclude only initializing bets, but if that’s difficult to do, excluding only markets created within the month of the league is a pretty good approximate fix.
(And it’s especially important for this first month to not penalize creators under the old rules).
Unlinked MC is a “loophole” but it’ll be a loophole regardless of how you handle the above case bc people can still add options to another creator’s markets. So I’m not sure if that’s planned to be fixed anyways.
@Ziddletwix Good point about that loophole already being there. You could fix the loophole entirely, by tracking profit on each independent option separately and filtering out disqualified profit from any independent option you added. That would make it safe to count already-created unlinked MC markets as well.
(side note, it would be wayyy more helpful if profit was actually displayed by answer for independent MC markets. It's not very helpful to know that I lost profit somewhere on a market with 200 independent options.)
I have been self imposing a rule on my own markets to say "Market Creator may only trade on this market after 1 week of operation or after 10 unique traders have traded, whichever happens first," to not game the leagues.
Perhaps that could be used as a jumping off point for the conversation to see what the community thinks is fair?
@Quroe The market that prompted me to start this conversation didn't get 10 traders until literally today, when it's been open for 2 months (combination of me not having followers and it being a niche topic). Even my Starship-related market took almost a month to get 10 traders.
I think a minimum-traders requirement might need to be as low as 2 or 3 (not including yourself) in order to be fair to people who haven't built a reputation for creating great markets.
(to be clear, when I say "minimum-traders requirement", I mean a requirement before trades against a market you created start counting for leagues)
Another thing to consider is that for someone who doesn't have a huge net worth, the only way making markets is worth it is if you get to do the initial probability correction. Under the current system, I would never consider letting someone else get that profit, given the number of traders I get and my limited liquidity.
Perhaps markets from creators with less than 100 ratings could get boosted in the rankings algorithm? That could go a long way towards helping the trader bonuses be worth it.
Also consider that I'm saying this as someone who joined Manifold when daily bonuses were much higher and mana could be bought much more cheaply (but after the first round of making mana more expensive). It's no wonder the site isn't growing when new users today are so extremely disadvantaged. Now that mana can't be converted to real money for charities, I think it's high time at least some of the changes to reduce the amount of mana being added get reverted.
I think unlinked MC market questions actually treat each question submitter as the market creator for the purposes of trader bonuses so that implies to me they need not be a loop hole. If you submit a MC option in someone else's unlinked MC question you would be the market creator for that option.
Rather deadlines or buffers based on the month or season, they could use the same criteria, but inverted, as they use for rule 2. Profits from market creator trades made within 31 days of market creation are excluded.
the only way making markets is worth it is if you get to do the initial probability correction.
The platform already lets you set the probability for binary markets but for some reason it forces you to use the API, they removed it from the web UI some time ago and no amount of pleading has convinced them to add it back.
All these forms of inefficient capital use for creators really add up and drag down those users who would be making the best markets. There are so many cases where the platform's current system "requires" you to use like 2x the capital on a question due to the reasons you outlined here. It's really important to me that they do something to address it but so far no luck.
@Ziddletwix if you took the path of only ignoring profit from self-trades in first month after creation, one consideration I'd add would be to maybe also ignore profit in the ~24hrs before resolution to avoid situations where the creator could self trade then resolve instead of just resolving and getting the liquidity back
Eg there's no reason I couldn't have reopened /Nat/starship-flight-10-soft-touchdown , traded it to 99% then resolved YES - instead of just resolving YES
Admittedly this could theoretically provide an incentive to still self-trade then just intentionally wait 24hrs before resolving but hopefully that behaviour would be obvious enough that social pressures and market ratings would sufficiently negate that incentive?
Would also admittedly make it awkward to count profit in leagues as you'd either have to wait 24hrs before the profit from a self-trade can be counted, or sometimes remove self-trade profits that you'd already counted
But still feels somewhat worthwhile given that if a creator already knows how they're planning to resolve a market, it takes no forecasting ability and adds nothing for them to simply trade it to 99% or 1% beforehand
@Nat I think that’s largely an issue with a creator opening and closing a market with the purpose of taking the liquidity. And I think that’s best handled via the rules (simply say “don’t close and reopen a market for the purposes of being the first to trade on it”).
The initialization piece is different because it’s vital that creators be able to trade immediately to be efficient with liquidity. The case you’re describing has no actual benefit and doesn’t need to be protected.