How the skill check works
The sample account's record since 2024-12-27, against the S&P 500 (SPY), in plain words. Nothing here predicts what comes next.
The short answer
Out of everyone with a record like yours, about 40% would have a real edge over the S&P 500 (SPY): one that isn't just a lucky run. It counts how rare real skill is among investors, which a p-value on its own doesn't.
1. Your edge, and how sure we can be of it
After taking out what the S&P 500 (SPY)'s moves explain, your account made +19.2% a year on its own (alpha). Daily ups and downs make that uncertain by about ±17.1%.
The dot is your alpha. The darker band is one step of uncertainty either side, the lighter band two.
2. The t-stat: edge against wobble
Your edge is 1.12 times its own uncertainty. Near 0 means no sign of an edge; 2 or more is the usual line for “probably not luck”.
The t-stat grows with the square root of time: about 3.8 years more of the same would reach 2.
What the p-value means
Picture 100 investors with no real edge over the S&P 500 (SPY), taking the same market risk as you over the same days. About 26 of them would still end up at least this far ahead of it, or as far behind, from random swings alone.
The usual line for "not random" is 5 in 100, so random swings could easily explain a gap like yours.
3. Why the chance is lower than the p-value suggests
A p-value treats you as the first investor ever tested. But most investors trail the market, so a good record is more often a lucky one than a skilled one. We read your record against how skill is spread among investors, and the two together give the chance.
- Investors in general: the typical one trails by 3% a year
- Your record on its own
- Both together: the share right of 0% is your chance, 40%
4. Time is the real test
If your edge stays what it has been, the chance it is real grows as the record gets longer, because luck evens out over time and skill doesn't. This is that “if”, not a forecast.
Our starting assumption, and how it improves
Studies of retail investors find most trail the market after costs. Until we can measure it ourselves, we start from this: the typical investor trails the market by 3% a year, and real skill varies by about 5% a year either way. As more broker-verified records join, we'll replace these with what those records show, so the chance gets more accurate for everyone. The method is called empirical Bayes; Baks, Metrick and Wachter used it on mutual funds in 2001.
A description of your past record, not a prediction and not investment advice. Sample data.