FuturesAI Scoreboard: Where Crypto’s 52.2% Sits After Fees
Short answer: FuturesAI’s scoreboard shows 48 correct out of 92 crypto calls, or 52.2%. At the overall payoff ratio of 1.19, break-even before costs is about 45.7%. After round-trip fees and slippage on a $10,000 position with a 1% stop, it climbs to 48.9% (Bitget rates) or 51.6% (Binance rates). The crypto record clears both lines, but its rough 95% band runs from 42.0% to 62.4%. Not investment advice. Leveraged futures can lose your entire margin.
Educational arithmetic, not investment advice and not a forecast. Figures were read on FuturesAI on November 14, 2026 and can change.
Crypto is the best-looking line on FuturesAI’s scoreboard. It is also the smallest sample except one.
Both facts matter. This post puts the 52.2% next to what a real trade has to beat, then asks how much the sample size lets you trust the gap.
What This Works Out
- What does the scoreboard say about crypto calls?
- What win rate does a trade need after fees?
- Does 52.2% clear that line?
- How wide is the uncertainty on 92 calls?
- What should you check before acting on it?
What does the scoreboard say about crypto calls?
FuturesAI’s public scoreboard, last auto-checked on November 11, 2026, lists crypto accuracy at 52.2% (48 of 92 calls), against 46.9% for US stocks (53 of 113) and 41.2% for Korean stocks (54 of 131) (futuresai.io/en/quant/scoreboard).
Overall, the page shows 46.1% (148 of 321 graded), a payoff ratio of 1.19, and an average return of +0.10% per prediction against +2.21% for the market over the same windows.
When I checked the page, I found that it says no fees are included in any of these numbers. Trades are priced at a frozen price and scored at the resolved price. That one line decides everything below.
The page also gives no single date range. The ledger’s call dates run from about August 28 to November 7, 2026, so this is a ten-week record.
What win rate does a trade need after fees?
A trade that risks R dollars to win 1.19 times R breaks even at a win rate of (R + C) / (R x 2.19), where C is the round-trip fee plus round-trip slippage in dollars.
With C at zero, that is 1 / 2.19, or 45.7%. Every dollar of cost lifts the line.
The cost inputs here come from FuturesAI’s payback page: an effective taker rate of 0.015% on Bitget and 0.045% on Binance (futuresai.io/payback, read November 7, 2026). On a $10,000 position a round trip is $3.00 on Bitget and $9.00 on Binance. I added slippage of 0.02% per side, $4.00 a round trip, which is my assumption for a liquid pair and not a measured figure.
| Stop distance | No costs | With Bitget effective fees + slippage ($7) | With Binance effective fees + slippage ($13) |
|---|---|---|---|
| 1% ($100 risk) | 45.7% | 48.9% | 51.6% |
| 2% ($200 risk) | 45.7% | 47.3% | 48.6% |
| 5% ($500 risk) | 45.7% | 46.3% | 46.8% |
Check one cell by hand. A 1% stop on $10,000 risks $100. Binance cost is $13. (100 + 13) / (100 x 2.19) = 51.6%.
Notice what moves the line. The fee barely changes with a 5% stop, because $13 is small against $500 of risk. With a 1% stop it is the difference between 45.7% and 51.6%.

Does 52.2% clear that line?
Crypto’s measured 52.2% sits above every break-even figure in the table, including the strictest cell of 51.6%, but by only 0.6 points in that cell.
| Comparison | Break-even | Crypto 52.2% minus break-even |
|---|---|---|
| No costs | 45.7% | +6.5 pts |
| 1% stop, Bitget fees + slippage | 48.9% | +3.3 pts |
| 1% stop, Binance fees + slippage | 51.6% | +0.6 pts |
| 5% stop, Binance fees + slippage | 46.8% | +5.4 pts |
US stocks at 46.9% clear the zero-cost line of 45.7% by only 1.2 points, and any fee-based line above 46.9% takes that away. Korean stocks at 41.2% sit below even the zero-cost line, if the same payoff ratio applied to them.
If, though, you read the table the other way, the story changes. The longer your holding window and the wider your stop, the less fees matter. The scoreboard’s windows run days to a few weeks, so the 5% row is probably the closer match.
How wide is the uncertainty on 92 calls?
A 52.2% hit rate on 92 calls has a rough 95% band of about 42.0% to 62.4%, which is wide enough to contain every break-even line above, including 51.6%.
| Market | Correct / graded | Hit rate | Rough 95% band | Range |
|---|---|---|---|---|
| Crypto | 48 / 92 | 52.2% | ±10.2 pts | 42.0% to 62.4% |
| US stocks | 53 / 113 | 46.9% | ±9.2 pts | 37.7% to 56.1% |
| KR stocks | 54 / 131 | 41.2% | ±8.4 pts | 32.8% to 49.7% |
The formula is the square root of p x (1 – p) / n, times 1.96. For crypto: the square root of 0.522 x 0.478 / 92 is 0.0521, and times 1.96 is 10.2 points.
It is a normal-approximation shortcut. It assumes the calls are independent, and crypto moves cluster, so the true band is probably wider.
So: the record is consistent with a real edge, and also with no edge at all. That is the honest reading. I have a longer treatment of the same point in FuturesAI Scoreboard: How Wide Is the Error Band?.
What should you check before acting on it?
Before acting on a 52.2% crypto record, check three things: that the payoff ratio holds for crypto alone, that your own fills match the cost assumptions, and that the sample has grown.
- The 1.19 payoff ratio is an overall figure. The page does not split it by market, so this post assumes crypto resembles it. If crypto’s real payoff is lower, every break-even above is higher.
- Replace my $4 slippage with your own. Look at the gap between your intended and filled price on your last 20 trades.
- Wait for more calls. A band of 10 points shrinks only with the square root of the sample, so doubling to 184 calls only trims it to about 7.2 points.
Related: for the market-relative view, where the page shows +0.10% a call against +2.21% for the market, see FuturesAI Scoreboard: Do Its Calls Beat Holding the Market?
Tips
- Compare any hit rate with break-even after your own fees, not before.
- Use the lower end of the band for decisions, not the headline number.
- Re-read the scoreboard before relying on a figure. It updates automatically.
- Keep position sizes small enough that a full loss is survivable.
Warnings
- Not investment advice. FuturesAI makes no profit promise, and neither does this post.
- Leveraged futures can lose your entire margin.
- The scoreboard excludes fees, so it overstates what a trader could have earned.
- The 1.19 payoff ratio is overall, and applying it to crypto alone is an assumption.
FAQ
Is a 52.2% hit rate on crypto calls good enough to make money?
Not by itself. Whether 52.2% pays depends on the size of wins against losses and on fees. At the scoreboard’s overall payoff ratio of 1.19 the break-even rate before costs is about 45.7%, but a tight stop and Binance-level fees can lift it to about 51.6%.
Why does the scoreboard accuracy ignore fees?
The scoreboard says its trades are priced at a frozen price and scored at the resolved price, with no fees included. That keeps the record simple but means real trading results would sit lower by the round-trip fee and slippage on every call.
Can I trust the 1.19 payoff ratio for crypto alone?
Treat it with care. The scoreboard publishes 1.19 as an overall figure and does not split the payoff ratio by market, so applying it to the crypto subset is an assumption. The conclusion should hold only as long as crypto’s payoff is close to the overall one.
Sources
- FuturesAI: quant scoreboard (read November 14, 2026; last auto-check November 11, 2026)
- FuturesAI: exchange fee payback page
Next step: write down your typical stop distance and your real round-trip cost, find that row in the break-even table, and ask whether you would still take the trade if the hit rate were the low end of the band.
Fact-checked against the FuturesAI scoreboard and payback pages as of November 14, 2026. Calculations are ours. Not investment advice.
๐ Related guides
