A five thousand dollar account has a very specific problem, and it is not the one people think. It is not that the account is too small to matter. It is that the position sizes it implies, somewhere between one and five hundred dollars, are large enough to move a genuinely thin book and small enough that a few dollars of spread eats a real share of the return.
So the question is never which categories are interesting. It is which categories contain books that will take 250 dollars in and hand it back later without charging you for the privilege twice. That test disqualifies more of the platform than most people expect.
Start from the exit, then work backwards to the entry
Pick a maximum share of visible resting liquidity that you are willing to be. Mine is five percent, and I would call ten percent the outer edge of defensible. Above that you are not taking a price, you are setting one, and you will set it again in the other direction on the way out.
Run a 250 dollar ticket through some real depth figures from the Prediction Alpha markets feed. Against a book of 282.0K, 250 dollars is under one tenth of one percent and you are invisible. Against 497.3K or 648.8K, the same. Against a book of 686 dollars, a 250 dollar order is 36 percent of everything resting, and that number is not a warning, it is a disqualification.
That gives you a working floor. At a 250 dollar ticket and a five percent share limit, you need at least 5,000 dollars of visible resting liquidity before a market is tradeable for you at all. Write that number down, because it converts a vague worry about thinness into a filter you can actually apply in twenty seconds per row.
Where the categories actually live
Two honest notes before the category list, both visible in that screenshot. The index came back empty when I went to capture it, which is a reminder that a single reading of any research page is one reading. And the categories themselves are not on this page at all.
The category chips live on the Prediction Alpha markets tab, and they read All, Politics, Sports, Crypto, Finance, Geopolitics, Tech and Culture. The stats tab on the same module counts 20 categories in total. So the seven chips are a grouping over a longer underlying list, which means selecting a chip is applying a filter rather than choosing a market. Two questions inside the same chip can have nothing in common in terms of depth, horizon or who is on the other side.
Keep that in mind through everything below. The category is a starting point for where to look. It is never the unit you trade, and no category is uniformly deep.
The three depth tiers you can observe
Across the readings available on this module, resting depth falls into three fairly distinct bands.
The top band is a handful of marquee questions with millions of dollars of book. The most liquid leaderboard on the trending tab topped out at 4.7M. There are not many of these, and on the capture I have they were concentrated in politics on a single venue.
The middle band is the high turnover rows on the markets tab, sitting between roughly 280K and 650K of resting liquidity. Comfortable for a five thousand dollar account and comfortable for an account ten times that size.
The bottom band is everything else, and it is where the surprises live. A market showing 796.4K of 24 hour volume was resting on 686 dollars of book. High turnover is not depth. A five figure account can operate in the top two bands without being noticed and cannot operate in the third at any size worth taking.
Category by category, on the readings available
Politics and geopolitics contain the deepest single questions on the platform, and they carry the longest dates. The tradeoff is capital lockup rather than depth. A contract resolving 01/01/27 commits your money for well over a year, and on a five thousand dollar account two thousand dollars parked until then is forty percent of your capital doing nothing while you wait. Deep enough to trade, expensive in a way that has nothing to do with spread.
Sports is deep on the day and gone the next. The 24 hour volume board was led by a football fixture at 4.1M and filled out with esports and tennis matches resolving the same day. The risk there is not the book, it is the clock: there is no time to be early and wrong, and the research you do has a half life measured in hours.
Crypto and finance sit in the middle band and are the one place where knowledge you already have might be worth something. The Fed rate question in the capture showed 648.8K of resting liquidity against 15.1M of lifetime volume, which is a real market with a real history.
Tech and culture are the ones to be careful about, and I want to be honest about the basis for saying so. I have not sampled every question in them. What I can tell you is that the pattern across this module is a small number of deep questions surrounded by a long tail of thin ones, and that the further you get from the categories that draw institutional attention the thicker that tail gets. Do not take my word for it. Apply the 5,000 dollar liquidity floor to the chip and count how many rows clear it. If the count is in single digits, the category is not fundable for you, whatever the questions look like.
The pass to run this week, and the two rules to keep
Twenty minutes, once, and then repeat it on a different day of the week because sports and esports distort any single reading. Sort by 24 hour volume. Filter to one category chip. Write down the liquidity figure for every row, not the volume. Count how many clear your floor. Then look at the end dates on the ones that did and ask how long your money would be committed.
Two rules come out of that and they are the whole discipline. First, never take a position larger than your stated share of visible resting book, checked against the Liquidity column before the ticket rather than after. Second, never open a position you would not be content to hold to resolution, because on anything below the top band the exit may simply not be there when you want it, and a binary you cannot sell is a binary you are holding to the end whether or not you planned to.
If a category leaves you with three tradeable questions and all of them resolve in 2027, that is not a category you trade with five thousand dollars. It is a category you read about for free until the account is bigger, and the screen will still be there when it is.