Two kinds of sources, two kinds of verification.
Not all trader data is created equal. We track two kinds of sources, and the verification method is different for each — which matters when you're deciding how much weight to give a result.
Copy-trader positions are executed and settled on the exchange. We fetch actual trade history directly — real fills, real profit and loss, stamped and signed by the exchange. We do not re-simulate these. The exchange is the source of truth. When you see exchange-verified in the product, the numbers are from the execution record, not a model.
Signal channels post trade calls — entry price, take-profit, stop-loss — but those calls are never executed on an exchange we can see. We simulate each call against real historical market data using the fixed rules described in the next section. The simulation is objective and applied identically to everyone, but it is still a model: it reconstructs what would have happened, not what did happen.
Every trader profile shows which type of verification applies, so you always know what kind of data you're looking at before you decide how much to trust it.
How we simulate a signal.
These rules are fixed. They don't change per trader, per asset, or per market condition. The same rulebook applies to every signal on the platform.
- 1.
Entry must be reached.
A trade only enters if the market price actually trades through the stated entry level within a set time window after the signal is posted. If price never reaches entry, the call is voided — not counted as a win, not counted as a loss. A signal that never triggered is not a trade.
- 2.
First target only.
If a signal lists multiple take-profit targets (TP1, TP2, TP3…), we credit only reaching the first target — not the best-case final one. This deliberately understates good traders rather than flattering them. A trader who posts TP3 and actually hits all three only gets credit for TP1 in our model.
- 3.
TP vs. stop: whichever comes first in 1-minute data.
After entry, we walk real 1-minute candle data forward and check which is touched first: the take-profit or the stop-loss. The first one hit determines the outcome.
- 4.
Tie in the same candle: stop-loss wins.
When a single 1-minute candle touches both the take-profit and the stop-loss, we assume the stop hit first. This is deliberately conservative — we could assume the best case, but we assume the worst. This means our model slightly underestimates performance in ambiguous candles.
- 5.
Time limit: closes at market if neither hits.
If neither the target nor the stop-loss is reached within the resolution window, the trade is closed at the prevailing market price at that time. It neither wins nor loses at the stated levels — it exits wherever the market is.
- 6.
Fees are deducted.
A standard trading fee is subtracted from each simulated trade. Profit isn't inflated by assuming zero-cost execution. We use a fixed fee rate applied consistently across all traders.
- 7.
Fixed stake per trade.
Every simulated trade uses the same fixed notional amount. This makes results comparable across traders regardless of the sizing they recommend. It also means our model doesn't compound wins — each trade is evaluated independently.
- 8.
Frozen on receipt.
A signal is recorded the moment it is posted. If a trader later edits the entry, changes targets, or deletes the post entirely, we've already captured the original. The historical record is what was posted live — this is the core mechanism that prevents channels from quietly erasing their losers.
Window and sample size.
All performance metrics are computed over a rolling 90-day window from the current date. A trader with a great year but a poor last three months will reflect the recent three months. This keeps rankings current rather than coasting on older results.
Before displaying full metrics, we require a minimum number of completed trades in the window. Traders below that threshold are shown as limited history or excluded entirely from the main rankings. A handful of trades cannot distinguish skill from luck — a 5-from-5 win rate on five trades is meaningless compared to 80 wins from 100.
Trade count is always shown alongside metrics, so you can judge how much confidence a number deserves. A Profit Score of 2.4 from 200 trades is a very different claim than a Profit Score of 2.4 from 22 trades.
What each metric means.
Every metric shown is derived from the verified trade record — not from the trader's own claims. Below: what it means, how it's computed, and where it can mislead you.
Profit Score
What it is: Total gross profit divided by total gross loss over the window. A score above 1.0 means the trader is net profitable. A score of 1.5 means they make $1.50 for every $1.00 lost.
How it's computed: Sum of all winning trade P&L ÷ absolute value of sum of all losing trade P&L. Fees are already deducted.
Limitation: A high Profit Score doesn't mean low risk. A trader could have a great score from a few big wins while carrying enormous drawdown risk. Always read it alongside the risk metrics.
Win Rate
What it is: The share of completed trades that closed profitably. Shown as a band (e.g. "50–60%") rather than an exact figure for free-tier users.
How it's computed: Winning trade count ÷ total resolved trades.
Limitation: A high win rate can still lose money if the losses are large relative to the wins. Profit Score is a more complete picture. Win rate alone is a common way traders make their record look better than it is.
Worst-Case Risk
What it is: The largest single-trade loss seen in the window (scaled to your budget), plus the worst consecutive losing streak. It's the most any single trade would have cost you at your capital level.
How it's computed: Worst single-trade loss % × your stated budget for the dollar figure. Worst streak = maximum consecutive losses without a winner.
Important limitation: Future losses can and do exceed the worst loss seen so far. This metric describes the historical floor — not a cap. Never treat it as "the most you could lose." Markets can move beyond any historical reference point.
Consistency
What it is: How steady vs. bumpy the returns are — our plain-language version of a risk-adjusted return. Shown as High / Medium / Low.
How it's computed: We measure the variance in per-trade outcomes relative to the average return. A trader who grinds out similar small wins scores higher than one whose results swing wildly.
Limitation: Measures past variance, not a guarantee of a smooth future. Consistency can change — especially when a trader's edge is regime-dependent.
Max Bet Size %
What it is: The largest share of account balance that a single trade would have used, based on the trader's stated sizing. A measure of aggression.
How it's computed: Maximum single-trade notional ÷ account balance at the time of that trade (approximated at fixed stake for simulated signals). Higher numbers mean more capital concentrated in one bet.
Stubbornness
What it is: How much longer a trader holds losing positions compared to winning ones. A high value suggests they let losses run rather than cutting them — which is a common reason traders underperform their own stated win rate.
How it's computed: Average time-to-exit for losing trades ÷ average time-to-exit for winning trades. A value above 1.0 means losers are held longer than winners.
Minimum Budget Match
What it is: Whether your stated capital is large enough to follow every trade this trader makes, including their smallest positions. An affordability calculator — not a recommendation.
How it's computed: The platform runs your stated budget against the minimum margin required for this trader's smallest leverage-adjusted position in the window. If your budget covers it, you see a match. This is purely arithmetic on your own number — we make no judgment about whether following this trader is wise.
Trend (improving / declining)
What it is: Whether performance has been getting better or worse within the current window. Shown as an up arrow, down arrow, or stable (no arrow).
How it's computed: We split the 90-day window in half and compare the Profit Score in the more recent half against the earlier half. Only a change above a minimum threshold shows an arrow — small fluctuations stay as stable. One arrow means a meaningful direction change within the window.
Limitation: Trend reflects recent direction within the window — not a prediction. An improving trend may reverse; a declining trend may recover. It tells you which way things have been moving, not which way they'll go.
What we deliberately don't do.
These aren't aspirational statements — they're operational constraints we've built into how the platform works.
-
We don't take affiliate money from anyone we rank.
No trader, signal channel, or exchange pays us for placement, visibility, or favorable treatment. Revenue comes from user subscriptions only.
-
We don't tell you who to follow.
TraderVerity shows a verified record. What you do with it is your decision. We deliberately avoid recommendations — the same data can lead different people to different conclusions depending on their risk tolerance and goals.
-
We don't accept self-reported performance.
Screenshots, PDFs, Telegram announcements about win rates — none of these are inputs to our rankings. Only exchange-confirmed executions or verifiable simulated signals count.
-
We don't hide losing traders to make the board look good.
Survivorship bias — only showing traders with good records — is one of the most common ways performance data misleads people. We track and display records regardless of whether they flatter the trader. A channel that's been losing for 90 days is still shown, not quietly dropped.
-
We don't claim to predict the future.
No metric on this platform is a forecast. All rankings are retrospective — they describe what happened, not what will happen.
Limitations.
We'd rather name these ourselves than have you find them. Owning a limitation is more useful than hiding it — you can account for it; you can't account for what you don't know about.
Vague signals can't be verified.
Our simulation requires a clear entry price, take-profit, and stop-loss. Some signal channels post vague calls ("buy BTC around 60k"), post-hoc entry adjustments, or no stop-loss. These calls can't be objectively simulated and are not counted. Channels that habitually post unverifiable calls will show fewer trades — not better ones.
Simulation isn't execution.
Even with real 1-minute market data, our simulation can't perfectly reconstruct what an actual live execution would have produced. Slippage (the difference between the stated price and the filled price), order book depth, and partial fills are not modeled. Real results — whether better or worse — will differ from our simulation. This is inherent to all backtested or simulated frameworks, not unique to us.
Historical vs. live-captured data.
Some traders were added to our tracking after they started posting. For those, we have live-captured data from when we began tracking them — and potentially reconstructed data from before that, which is labeled separately. We clearly distinguish "live-verified since [date]" from reconstructed history. Any claim about live-capture is only made where it's true.
Past performance doesn't predict the future.
No verified track record — no matter how rigorous the verification — is a guarantee of future results. Markets change. Edges erode. Traders change strategies. Everything here is historical data about a period that has already ended.
Disclaimer.
TraderVerity is a data platform, not a financial adviser. Nothing on this site is financial advice. Independently verified historical performance data is still historical data — the decision about whether to follow anyone, and how much capital to risk, belongs entirely to you.
Crypto markets are volatile. Following any trader, however strong their verified record, carries real risk of loss. Position sizing, timing, and personal risk management are your responsibility.
Questions about the method? We mean it when we say check our math. If you find a flaw in the approach, an inconsistency between this page and the product, or something that seems wrong — we want to know. The methodology must always describe what the product actually does. Write to us.