/ Check 7 · Text versus chart

Does the screenshot match the claim?

A backtest usually arrives as words and a picture. They should describe the same run. When they disagree, at least one of them is from a different run, and you cannot tell which one the result belongs to.

The short answer

Read the claim, then look for each part of it in the picture. The trade count should match the markers you can count. The date range should match the time axis. The symbol and timeframe should match the chart header. The drawdown should match the deepest dip on the equity curve. None of this needs the code or the trade list, and it is the cheapest check on the list: if the words and the picture cannot both be true, stop there.

Four things to compare

  • Trade count against markers. Count the entry markers visible on the chart. If the panel or the text says 300 and you can count forty, the numbers and the picture are not the same run, or the picture shows a small slice of it.
  • Date range against the time axis. Read the first and last dates on the axis. A claim about two years shown on a chart that spans a few months is a claim the picture does not support.
  • Symbol and timeframe against the header. The chart header names the instrument, the venue and the bar size. They should be the ones the text names.
  • Drawdown against the equity curve. Find the highest point on the curve before its deepest dip, and the bottom of that dip. The fall between them, as a share of the high, is the drawdown the picture shows.

A worked example: one write-up, one screenshot

A write-up says: 312 trades over two years, maximum drawdown 8%. The screenshot under it shows this:

  • The time axis runs for five months.
  • You can count 41 entry markers.
  • The equity curve peaks at 12,400 and then falls to 10,540 before it recovers.

Now compare. The fall is 12,400 − 10,540 = 1,860, and 1,860 ÷ 12,400 = 15%, not 8%. Forty-one trades in five months is 8.2 a month; at that pace two years would hold about 197 trades, not 312. And five months is not two years. Three claims, three disagreements, and you have not opened the code.

That does not prove the write-up is false. It may describe a different run, a later version, or different settings. It does mean the picture is not evidence for the words, and you do not know which run the 8% belongs to. Ask for the run that produced both, with its settings stated. If the trade count survives, the win-rate interval calculator tells you how much that many trades can say.

How to run the check on any report

1. Write the claim down as numbers. Trade count, start and end date, symbol, venue, timeframe, drawdown, net result. Whatever the text states, list it.

2. Find each one in the picture. Count markers, read the axis, read the header, measure the deepest dip. On TradingView the strategy report sits under the chart; its numbers and the chart above it should come from the same settings.

3. Mark every disagreement. One small one can be a cropped screenshot. Two or more, or one large one, means the words and the picture are from different runs.

4. Move on to the settings. Check 8, window, venue and costs, asks whether the run can be reproduced at all.

Why this is the check people skip

Words and a picture side by side feel like two pieces of evidence. Usually nobody compares them, because each one looks fine on its own. The chart looks like a good chart and the numbers look like good numbers. The check is to read them together, and it costs a minute.

This is check 7 of 10

The one-page Backtest Verification Checklist, free. Confirm the email and the PDF opens straight away; one worked check each Tuesday after that.

How these numbers were worked out

Every figure is a made-up write-up and screenshot, not a result of any system. Drawdown is the fall from the peak to the bottom of the dip as a share of the peak: 12,400 − 10,540 = 1,860, and 1,860 ÷ 12,400 = 0.15, or 15%. Trade pace is markers divided by months: 41 ÷ 5 = 8.2 a month; two years is 24 months, and 8.2 × 24 = 196.8, about 197. Nothing here is investment advice.

Historical simulations are not live results. Nothing on this site is investment advice.