This seems incredibly reasonable to me, maybe with an option for a question creator to disable the limit of they think the question needs that much precision.

Most markets don't, and you just get more problems from that last 1% than you get information.

Also avoids problems where someone dumps m10,000 pushing my question to 99.9% and then gets upset when I don't resolve it the same day.

An interesting idea tbh. Curious if anyone at Manifold considers any of these traits as relevant things to solve.

When even Mira doesn't have enough mana to get all the NO markets under 1% before they close, I still like this idea.

@Mira I expect this is the right prediction but it's going to feel brutal when this market gets bet below 1%.

This also slightly reduces the profit advantage of a large bankroll. Betting a market to 1% takes 10x less capital than betting it to 0.1%.

Someone brought up issues specific to multiple choice / free response markets this general idea might help alleviate somewhat

I'd rather they actually fix the reliability of extreme probabilities. There's a lot of value in being able to bet on the difference between 0.1% and 1% for important events.

@IsaacKing I think that value is 90% smaller than being able to get on the difference between 1% and 10%, and requires 90% more capital, and Manifold isn't well suited to it.

I think having the markets work well around 0.1-1% would require some adjustments to the market mechanisms, I don't think there's an easy solution. E.g. adding a new form of leveraged trading could be helpful.

@MartinRandall I disagree! Even for low-risk events it can be very useful to know how low the risk is.

@Boklam I think the issue is that whether the probability is 99.5 or 99.8 probably has more to do with other factors than with the probability of the underlying event.

@jack sure, I think a different market structure might work, I see that insurance markets for rare events exist in the world. But that could be a different market type, and perhaps a problem solved by a different start-up.

Would you consider anything around that ballpark within the spirit of the question? Suggested ~1% and ~99% as rough starting points, but I'd still consider my idea implemented even if it was implemented at slightly different exact percents

@PatMyron a 2-98% limit resolves this yes, because they prevent betting beyond 2% and thus beyond 1%

A 0.5%-99.5% limit does not, maybe worth a separate market. I think 1% is a good number though.

I think it's a good idea, but one concern is what actually happens when the price reaches 1% and someone wants to buy more NO? The seemingly obvious answer is that you just can't, which means there is a cap on how much you can buy at once, which is reasonable but may cause confusion.

@jack Yes, either nothing, or place a limit order.

@jack the more challenging question is what you would do about selling YES? selling YES is kind of the same as buying NO, but disallowing selling YES seems wrong to me.

@harfe The AMM won't buy your near worthless YES shares so you would be left with unsellable shares, similar to other prediction markets with the same feature.

I think this change would make a lot of sense!

How will this resolve if Manifold only disables quick bet arrows for extreme probabilities, but allows regular bets and limit orders?

@Yev that's a great feature but doesn't change market resolution. The market is not limited.

But I would accept a UI limit that can be bypassed via the API. Grudgingly.

• More resilient to tranche corrections.

This is important.

@IsaacKing What is a tranche correction? Does it have something to do with this user called spindle who likes to spell thingz with z's?

@Boklam I wouldn't know, I'm not an expert.