Guide
How to check a stock picker's track record
Short answer
Check five things: who set the timestamps, whether every pick is still on the record, whether you could have bought at the stated price, whether the picks beat the market after costs rather than riding it, and whether the record is long enough to rule out luck. A seller's own spreadsheet fails the first check, because the seller controls its dates.
- technical rules that looked like winners over 30 years
- 34 of 324
- of those that beat the market after risk
- 0
- months a real 1%-a-month edge needs to show, one try
- 99–223
The five checks at a glance
| Ask | Why it matters | A credible record shows |
|---|---|---|
| Who set the timestamps? | Whoever controls the dates can add picks after the fact. | Times from a clock the seller does not control. |
| Are the losers still there? | Dropping bad picks is the easiest way to look good. | Every pick ever sent, open ones included, none deleted. |
| Could you have bought at that price? | Subscribers get the pick after the seller decides. | Entry at the next market open after publication, with costs. |
| Did it beat the market, or ride it? | In a rising market almost any list of stocks goes up. | Returns compared with the market over the same days. |
| Is it long enough to rule out luck? | A year or two of results is mostly noise. | Years of monthly results, or a clear "not enough evidence yet". |
Who set the timestamps?
A track record is only as good as its dates. If the seller typed the date into a spreadsheet, they could have typed any date, and a winning pick can be added after it won. Screenshots have the same problem.
Look for times that come from somewhere the seller does not control: a third party's clock, or an email or post that a platform stamped when it went out. At Autrady the time of each pick is our receipt time. The seller cannot set it.
Are the losing picks still on the record?
Most flattering records are not fake. They are incomplete. Watch for these:
- Deleted or edited picks. Ask for every pick ever sent, not a highlights page.
- Closed picks only. An "average gain of closed picks" leaves out losers that are still open.
- The best of several portfolios. A seller with five model portfolios can advertise the one that did well.
- A start date chosen afterwards. Every record looks good from the right starting point.
A record where each entry is chained to the one before makes a removal visible: change or drop one pick and every later link breaks. Autrady publishes the newest link of every track each day at /audits/chain.txt, so a later edit would not match an archived copy.
Could you have bought at that price?
Picks are often quoted at the price when the seller decided, before any subscriber could act. A fair test buys at the next market open after the pick was published and charges spreads and costs on the way in and out.
This matters more than it sounds. A viral grid bot's own backtest showed +24.5% a month. Replayed on 1-minute prices, it lost 1.6% a month.
Did the picks beat the market, or ride it?
A list of stocks bought in a bull market makes money. The question is whether it made more than the market did over the same days, after costs. Much of what looks like stock-picking skill is exposure to the market, to small companies, to cheap stocks or to recent winners.
In our own research, 34 of 324 technical rules looked like winners over 30 years on raw returns. None of them beat the market after risk.
Is the record long enough to rule out luck?
Rarely. An edge of 1% a month over the market, with a typical monthly swing of 4% to 6%, needs about 99 to 223 months of results before a test can tell it from luck 80% of the time. If the record is the best of several tries, it needs much longer. How long a track record must be has the table and the math.
You can test any claimed record in a minute with the free luck check: enter the average monthly return, the monthly swing, the months and how many versions were tried.
Red flags in a stock picker's track record
- Screenshots or a spreadsheet that only the seller can edit
- Returns without the date and entry price of each pick
- Averages of closed picks, with open picks left out
- No benchmark, or one chosen after the results were known
- Several newsletters, portfolios or accounts, with only the best one shown
- A record of a few months presented as proof
- "Verified" or "audited" without saying what was checked
What should "verified" or "audited" mean?
Ask what exactly was checked: the dates, the prices, the returns, or skill. These are different claims. A record can have honest dates and still not beat the market.
That is why an Autrady track page shows two things and never merges them. The first is authenticity: "Tracked since 15 November 2026, 214 picks, none deleted." The second is the verdict: beats the market after costs, does not beat it, or not enough evidence yet. We never call a record "verified profitable".
How Autrady tracks a seller
- The seller proves their public profiles, such as X or Substack, with a one-time code.
- Each pick is stamped with Autrady's clock when it arrives, chained to the one before and never deleted. A typo can be voided within 15 minutes, and the void stays on the record.
- Open picks show only a hash until the holding period and an embargo of at least 30 days end, so paying subscribers see them first. Then the ticker is revealed and anyone can check the hash.
- Closed picks are tested against the market after costs. A verdict needs at least 12 months with closed picks.
See the tracked records, or read how tracking works for sellers.
Selling picks, or buying them?
Sellers get a public record nobody can backdate, free for founding sellers until 31 March 2027. Buyers read every track and audit free, and can check any claimed record for luck.
Autrady records and tests evidence. This guide is research, not investment advice, and past results do not predict future returns. Questions or corrections: [email protected].