Is your edge real, or just variance?

Every serious prediction-market trader eventually asks the same question: am I actually good, or just running hot? It has a real answer, and it comes down to four things, measuring your edge against the price, proving it isn’t variance, tracking your calibration, and sizing to survive. This is the complete toolkit for answering it honestly.

01
Edge means beating the price
The market price is already a sharp probability estimate. You only have an edge when your win rate clears the prices you paid, not just when you win. Winning at 90-cent favorites isn't an edge.
02
Variance hides the truth
A real edge still produces losing streaks, and a coin flip still produces hot ones. Over a small sample, results tell you almost nothing. Separating skill from luck takes a large enough record.
03
Calibration is the skill
The deepest edge is having probabilities you can trust: your 70% calls happening about 70% of the time. It's measurable, trainable, and it's what the market pays for.
04
Sizing lets you collect it
An edge only pays if you survive the downswings to collect it. Sizing to weather the bad stretches is the difference between an edge on paper and money in the account.

The toolkit, from quick check to full analysis

Edge Lab
The full picture on your own trades: a skill-vs-luck significance test, your edge by price band, tilt, and day-of-week leaks.
Sharp Score
One 0-100 rating of your edge, benchmarked against the field. Estimate yours free, no upload.
Edge significance
Enter your record and get a straight verdict: real edge, real leak, or variance.
Trades to prove it
How many trades it actually takes to prove the edge you think you have.
Variance simulator
Watch 1,200 simulated seasons to see how often a real edge still ends down.
Prediction Journal
Log your probability calls and track your calibration on real markets over time.
Calibration test
A free 2-minute test: are your 80% calls right 80% of the time?

Go deeper

Can you make money on Kalshi?How many trades to prove an edge?How to get better calibratedBankroll management and sizing

Common questions

How do I know if my trading edge is real?
Compare your win rate to the average price you paid, across a large enough sample. If you win significantly more often than the prices implied, quantified by a z-score around 2 or higher, it's unlikely to be luck. Below that, your results can't be distinguished from variance yet.
Can you be profitable and still not have an edge?
Yes, over a short sample. Variance produces winning runs even with no real edge, so a profitable month or even a profitable few hundred trades at heavy favorites may not reflect skill. A significance test on your record is what separates the two.
What's the difference between edge and calibration?
Edge is beating the price on your trades. Calibration is whether your stated probabilities are accurate across the board. Good calibration tends to produce edge, and it's the more fundamental, trainable skill, which is why tracking it is so valuable.
How many trades do I need before I trust my results?
It depends on your edge and prices, but far more than most think. Thirty is far too few; around a hundred the signal firms up, and several hundred trades make a real edge hard to hide. The trades-to-prove-edge calculator gives your specific number.
Put a number on it rather than a feeling:
Is your edge real, or just variance?
Enter your record: find out if your win rate is skill or luck.
How many trades until you know you're good?
The sample size it takes to prove a Kalshi edge is real.
What does a real edge actually look like?
Simulate a season of trading and watch how wildly a real edge still swings.
Free, no signup, and nothing you type leaves your browser. Every calculator →