Bot Stack Correlation

Running several trading bots? Upload each bot's TradingView List of Trades export below and see whether your “diversified” stack is really one bet in disguise. Pairwise correlations above 0.5 get flagged as the same bet twice — the check Kevin Davey uses before adding a bot to his portfolio. Everything runs in your browser; your trade files never leave this device. See also the expectancy calculator.

1. Upload trade lists

Export each bot’s trade list from TradingView as CSV and drop one file per bot.

Drop .csv files here

or click to browse

No bots loaded yet. Drop CSV files above or press “Load demo data”.

🔒 All computation happens in your browser. Your trade files never leave this device.

No uploads, no network calls — CSVs are read locally with File#text() and parsed in memory.

3. Correlation matrix

Daily P&L correlation. r > 0.5 means two bots are effectively placing the same bet.

Add at least two bots with trading history to see correlations. A pair shows a value only with ≥ 20 overlapping active days.

4. Overlap analysis

Real diversification loses in different weeks — shared red weeks are a warning sign.

0
weeks where ALL bots lost
Pairwise — % of active weeks both lost

Need at least two bots with weekly history.

5. Combined equity (risk-normalised)

Each bot is scaled to 1σ of daily risk, then summed. Thin lines = single bots, bold line = the combined book.

Not enough daily data to draw the curve yet.

Combined max drawdown
0.00σ

In σ units — with every bot sized to equal daily risk, the whole book gave back 0.00σ of one daily σ from its peak.

Worst single-bot max drawdown

No bots loaded.

6. Davey rules checklist

A quick sanity check before you add another bot to the book.

  • PASSOne bot per pair
    No two bots correlate above r = 0.5.
  • NEEDS TAGSMix trend + mean reversion
    Tag every bot's style — at least one is still “other”/untagged.
  • INFO1–3% risk per trade
    Cannot be verified from trade lists — confirm your per-trade risk is 1–3% of account.
  • INFOMax DD < 20% per bot
    No bots loaded.

How it's calculated

r = Σ(x−x̄)(y−ȳ) / √(Σ(x−x̄)²Σ(y−ȳ)²)

Each bot's closed trades become a daily P&L series (profits summed per exit date). Pairwise Pearson correlation is computed on the union of dates, ignoring days where neither bot traded, and requires at least 20 active days before a value is shown. For the combined equity curve each bot is scaled to 1σ of its own daily risk first, so a big-size bot can't drown out a small one — then the daily means are cumulated.

DD = max peak-to-trough decline

Max drawdown is measured on each bot's cumulative $ curve and expressed as a % of the account size you enter (Davey's gate: under 20%). The weekly overlap counts weeks where every bot lost money, plus per-pair “lose together” rates.

Davey's diversification rules

One bot per pair. If two bots correlate above 0.5, they're the same bet — keep one.

Mix styles. A stack of five trend-followers is one regime bet. Pair trend with mean-reversion so something works when the other doesn't. Tag each bot's style above to check this.

Size 1–3% risk per trade. Can't be read from a trade list — confirm it yourself.

No bot over 20% max drawdown. A bot that drew down more than a fifth of its account fails the gate.

Built for education — not financial advice.

connecting…