PredictPnL's skill ranking scores each wallet on how well it priced its trades, band by band. In the ranking of 2026-11-09, 55 of the top 100 by skill got the biggest piece of their score from shares bought below 25. In the rest of the ranking, a third did. The top 100 did not buy more of those cheap shares than anyone else. They bought them better.
How the score is built
The True Skill Score (TSS) groups every closed trade into nine bands by entry price, from under 15 to 85 and up. In each band it compares what the shares turned out to be worth, 100 or 0 at resolution or the sale price, with the average price paid. A band with few trades is shrunk toward zero, so a handful of lucky trades cannot move the score much. The bands are then combined into one number from −1 to +1, where 0 is a fair price. The metrics guide has the formulas.
One detail matters for this article: the bands do not weigh the same. A share at 50 swings a lot, so beating that price by 3 points is ordinary. A share at 10 swings much less, so the same 3 points are a bigger feat, and they count for more. That rule is the same for every wallet. The comparison below is between wallets scored the same way.
The strongest band of each wallet
For each wallet, the "strongest band" is the price band that adds the most to its score. The ranking of 2026-11-09 covers the top 1,000 Polymarket wallets by volume, and 952 of them have 50 or more closed trades.
- Longshots (below 25): 55 of the top 100 by skill, against 33% of the other 852.
- Middle (25 to 75): 21 of the top 100, against 34% of the rest.
- Favorites (75 and up): 24 of the top 100, against 33% of the rest.
- In the top 100, 35 have their strongest band below 15, the cheapest band of all. In the rest, 24% do.
- In the top 20, it is 13 of 20.
- Of the 55 in the top 100 whose strongest band is below 25, 38 have enough trades there for the band to count with at least half its weight after shrinkage. For them it is not a handful of lucky trades.
What they buy is not what sets them apart
The same ranking also records where each wallet's shares were bought.
- Longshots (below 25): 29% of the shares the top 100 bought, against 34% for the other 852.
- Middle (25 to 75): 37%, against 43%.
- Favorites (75 and up): 34%, against 23%.
The top 100 bought slightly fewer longshots than the rest, and more favorites. Yet their score comes from the longshots far more often. On average, the money they put below 25 came back better than the price said.
They also did better on the bottom line. On closed trades, the median top-100 wallet returned 2.6% of what it put in, against 0.5% for the rest. 92 of the 100 were positive, against 62% of the others.
Why the cheap end
A cheap share is a statement that something is unlikely. If the market leans on hopeful buyers at the cheap end, some of those prices end up too high, and a buyer who passes on them and picks the ones that are too low builds an edge there. Studies of other prediction and betting markets have long described this pattern, often called the favorite–longshot bias. This data does not prove that it is the reason on Polymarket. It shows where the skilled wallets' edge sits, not why.
Two other things shape the result. The score weighs the cheap end more, as explained above, so an edge there shows more clearly. And a cheap share is where luck looks largest: one share at 5 that pays 100 is a 20× return. The shrinkage and the 50-trade minimum are there to keep that from passing for skill.
What this does not say
It does not say that buying cheap shares works. The rest of the ranking bought more of them, and most of it did not get its edge there. It describes what already happened to the top 100, not what will happen next. A price that was too low last month may be fair today. PredictPnL shows the data; it is not investment advice.
See the numbers
The skill ranking lists the top wallets by True Skill Score, and each wallet's page shows its score and where it trades. The Markets page shows which long-running events the top 100 hold positions in right now. For more on how to read the score, see How to measure skill on Polymarket. To see your own strongest band, add your wallet.
Method: PredictPnL's daily skill ranking of 2026-11-09, covering the top 1,000 Polymarket wallets by volume. Only the 952 wallets with 50 or more closed trades are counted, ranked by True Skill Score. The strongest band is the price band with the largest positive contribution to the score, after shrinkage. "Enough trades" means a shrinkage factor of 0.5 or more, about 10 effective trades. For the most active wallets, the ranking reads the 10,000 most recent closed positions.