Announcing it here but will also add it to community guidelines:

@AndrewG and @derikk found a bug with managrams that allowed them to make the balance of one account infinitely high and the other infinitely negative. They created a detailed PDF which they privately sent us alerting us of the bug.

We have awarded each of them 5,000 mana, and when I eventually get around to it I'll give them an "infinite mana bug-finder" badge as well which I also owe @8. We have awarded this to them as they reported it in a sensible way and didn't abuse it to seriously manipulate any markets (aside from this one, which was more of a by-product).

Going forward, we will continue to award bounties to people who responsible find and report bugs.

That being said, we don't want to incentivise people to abuse bugs and manipulate markets. Therefore, going forward, ALL markets which have its resolution affected by a Manifold-specific bug should ignore the bug if possible, or N/A if not possible. Eg. if this market + bug were to happen again then it would stay open and eventually resolve NO unless someone actually earned mana not using a bug.

Bugs/exploits on other websites that affect a market resolution criteria we will leave to the creator's discretion. Although, I am considering actions that go against other websites TOS eg. follower botting on Twitter, to not count toward Manifold resolutions. Still undecided on this.

Please reply to this with comments of any weird examples you can think of that you are not sure would count as a bug or not. That way we can make sure we account for all the weird edge cases now in wording updated in the guidelines and are all on the same page.

How does this policy apply to markets which were created before before now?

@Joshua yeah meant to N/A this one too

@DavidChee Would markets that accept bugs as a potential resolution criteria still be resolved as N/A to discourage bug abuse?

+ do i get sick perks

@8 Yeah we'd either N/A them or forcibly change the resolution criteria to not count bugs.

I think this is a very good policy change.

The one edge case I can think of is markets about who will or will not become trustworthy/untrustworthy/be banned. These markets can hinge on bugs and other exploitative behavior. Is it still fine to resolve a market about someone losing the trustworthy badge if that happens they abused an exploit?

After internal discussion we we will be N/A'ing this market.

In the future bug exploits on Manifold will not count towards fulfilling resolution criteria.

I'll be writing a longer post explaining the reasoning behind this decision and examples of bug manipulation in a bit (getting on a flight rn)

@DavidChee I support!

@DavidChee this is admittedly not entirely consistent with previous decisions (although I believe we did also NA this last instance of bug related market). But going forward will be

@DavidChee Nice, been waiting for this

@DavidChee I'm not terribly opposed to this market having been resolved N/A. Personally I would think this should count under the "manipulation" clause, but I understand that opinions differ.
As I see it I did not explicitly exploit the market, but found after pumping my mana balance that I had caused it to become true. I did not use my 600 trillion mana to directly influence markets, so I don't think that I should be penalized for small bets here (beyond an N/A).

A similar market that explicitly allows for exploits in its resolution:

Will another infinite mana exploit be found by the end of 2023?

50% chance. Resolves YES if proof is posted of an account not belonging to the Manifold team with a current balance of at least Ṁ1,000,000,000 (however obtained).

@jacksonpolack @Loppukilpailija @ItsMe @RobertCousineau @Joshua @ian

Now that the votes on resolving NO has barely surpassed votes on YES (11 NO vs 10 YES as of now) , and much discussion and arguments was raised, I am a little unsure about what I should do.

@ian Can I leave the resolution up to you/other staff/ a trustworthy-ish user who would volunteer?

If you think this should resolves YES, consider this market:

https://manifold.markets/jacksonpolack/will-the-acx-post-hypergamy-much-mo-9991c723b94b

Someone botted 80ish likes on this ACX post, pushing it above the threshold at the last minute. Should I have resolved this yes? Allowing these posts to resolve YES creates incentives for, a pattern of, and implicitly endorses taking commons-burning action that hurts our ability to predict in markets. If it's consistently allowed, people will specialize in it, and it'll increase in frequency. As another examle, in a twitter market about HMYS followers, someone gave him 100 fake followers to resolve it YES. Emboldened by this, the same person bought fake followers in the twitter market about mana-chan's followers. How is this good? Why encourage it? We should vote no.

Also, consider law IRL, where the technical wording of contracts famously takes precedent. Even then, in a contract depending on the 'price' of an item, without any direct reference to hacking - if someone hacked the price to be 100x what it usually was, a judge wouldn't require you to honor that fake price. That'd be stupid, and against whatever the purpose of the contract is initially.

Finally, if this trend continues, every manifold meta-market now has a secret hidden clause where if you manage to exploit manifold well enough, you get to change the resolution, unless the description explicitly mentions that case. How many market descriptions currently mention that? How many descriptions will be edited after this market? How many new markets, created by a hundred different causal users, will mention that exploits don't count?

A quote from Vo on the HMYS market:

@Tarl 200 MANA profit for a USD 0.077 investment 🤑

He then made #1 profit with 300 M$ on the sneakysly market.

Please scroll down and vote NO on the poll.

Disclosure: I hold NO shares and did before the poll. But this is my consistent opinion on this matter from months earlier, as my posts on the acx market and twitter market suggest.

Agree. The foremost target in designing resolution criteria is that people should know what makes the market resolve YES/NO. As a consequence. you should strongly dis-incentivize pushing the market/reality to edge cases where resolution is unclear - or at the very least not incentivize finding loopholes etc.

There's also the standard critique of prediction markets incentivizing affecting the predicted phenomenon, changing what the prediction is about. This is happening here, but worse: the predicted phenomenon is being affected in a way that makes it go wildly out-of-distribution.

(Compare to markets of form "Will someone achieve thing X?" and someone goes and does the thing and gets prediction points - the state of reality is well within the scope of the prediction market. This is often okay and even hoped for!)

https://manifold.markets/ItsMe/will-the-ninthcause-market-resolve

Will the NinthCause market resolve N/A?

34% chance. This market: https://manifold.markets/AmmonLam/will-someone-surpass-the-portfolio [link preview]

Edit: See the discussion below. Ian roughly says it is up to the market creator, he has no strong opinion on this specific market.

In the infamous "Will the Sum of Shares..." market, @Ian explicitly states that the proper course of action after someone used a glitch to cause it to resolve Yes was to keep the Yes resolution and refund the exploited with fined winnings from the exploiter.

He instead resolved it N/A because it was a /very/ large market and they did not have time to do so. Nevertheless, the proper way, per my understanding of @ian's comment is to resolve this Yes.

https://manifold.markets/JonnyBairstow/will-the-sum-of-shares-on-any-marke-99ccfd2db099#pZTYVvMwC5HdIkM6BsH

Will the sum of shares on any market exceed 100 000 000 in the next two weeks?

Resolved CANCEL. Resolves YES when there is a market with the sum of YES and NO shares bigger than 100 000 000 (100 million). I will not search for any of them, you will have to provide me a link for that market. Doesn't have to be at close, any moment will count towards YES resolution. Has to happen after this market creation. I don't care how and where, I just need to see the Big number. @/JonnyBairstow/will-the-sum-of-shares-on-any-marke @/JonnyBairstow/will-the-sum-of-shares-on-any-marke-fd62b674234d @/JonnyBairstow/will-the-sum-of-shares-on-any-marke-018666941c28 @/JonnyBairstow/will-the-sum-of-shares-on-any-marke-a1630708cfe4 @/JonnyBairstow/will-the-sum-of-shares-on-any-marke-a010815e211b @/JonnyBairstow/will-the-sum-of-shares-on-any-marke-40c0f3a6eb24 @/JonnyBairstow/will-the-sum-of-shares-on-any-marke-8f2901c9436a @/JonnyBairstow/will-the-sum-of-shares-on-any-marke-c75056d0dc8c @/JonnyBairstow/will-the-sum-of-shares-on-any-marke-d8ddf5081eaf @/JonnyBairstow/will-the-sum-of-shares-on-any-marke-9f970346ae77 @/JonnyBairstow/will-the-sum-of-shares-on-any-marke-63681e4f0ae8 @/JonnyBairstow/will-the-sum-of-shares-on-any-marke-fca2c69b97e1 @/JonnyBairstow/will-the-sum-of-shares-on-any-marke-e60d99961b3f @/JonnyBairstow/will-the-sum-of-shares-on-any-marke-99ccfd2db099 <-- you are here @/JonnyBairstow/will-the-sum-of-shares-on-any-marke-37e1d02a8a04

@RobertCousineau under no interpretation of @ian's comment should this resolve no.

I can see the argument for N/A, but resolving yes seems very silly to me. I think when someone finds an exploit they get rewarded by the bounty, but otherwise they shouldn't be able to profit from it.

@RobertCousineau To be clear, I wished it to be resolved yes because the creator had already resolved yes, in this case it’s still up to the creator

@Joshua it should resolve Yes because the resolution criteria were fulfilled. In the much hairier situation. (More vicious exploiting), @ian still said the market should resolve yes.

I see no reason why this should resolve no - that goes against the resolution criteria and caselaw/historical judgements made.

I mean at this point it's additionally complicated by Ian saying it's up to the creator, and the creator saying it's up to the poll in the comments, and then now people voting on all that information being public.

@ian Count me confused - you said "it should rightfully be resolved YES". Why would you put "rightfully" in there if you did not think that was the right resolution?

Aside from all the debate about which way this should go, I definitely think the market should be closed right now instead of continuing to run while the poll runs. But up to the creator, of course.

@RobertCousineau I’m saying I’ve no comment on how this market should be resolved. The other market was similarly hairy and I had no opinion on the “absolute and true” way it should’ve resolved, only that I wished to leave it resolved as the creator saw fit.

@ian fair enough - we are all guilty of confusing word choice at one time or another.

@RobertCousineau I'll just mutter "rightfully" and a couple expletives over and over to myself for a moment :P

@RobertCousineau I mean, there’s a chance you’re correct that I did have an opinion back then, I only vaguely remember that decision. Either way I’m not to be taken as any supreme judge.

The creator should resolve it as they see fit. It does make me uncomfortable one of the bug exploiters stands to gain profit from this market, but imo it’s not enough to justify any enforcement. We paid out a bounty that is smaller than it would’ve been had they not exploited the bug, i think.