Trading & signals glossary
What is backtesting?
Backtesting means testing a strategy on historical data: applying its rules to what already happened to see what result it would have produced. With trading signals, it means taking every signal a channel posted and checking against real prices whether it hit the target or the stop first.
How it works
You lock in the rules (when to enter, where the stop goes, where the targets go) and run them over past price data without cheating.
To know what happened first inside a candle (the stop or the target) you need tick-by-tick data, not just 1-hour candles.
You have to include costs (spread, commission) and the signals that never triggered, or the result comes out better than it really was.
Signal
entry, stop and targets exactly as posted
Real price
tick by tick from the second of the message
What hit first
target or stop, with the spread
Result in R
plus a note if it couldn’t be measured cleanly
Example: measuring a signal that was already posted
- Signal: buy gold at 4,200-4,205, stop 4,190, target 4,220, posted at 9:14.
- Find the first price inside the zone after 9:14: 4,204 at 9:16.
- Walk through the ticks: price hits 4,220 at 10:02 without touching 4,190 first.
- Result: target, (4,220 − 4,204) ÷ (4,204 − 4,190) = +1.14 R.
Our data · Verpips
How we measure
- 3,866
- signals measured tick by tick
- 22%
- not cleanly measurable
- 303
- never triggered
3,866 signals measured
■ measured: 3,032■ never triggered: 303■ ambiguous or outside the zone: 531
We measured 3,866 signals from 18 channels against real prices, tick by tick and with the spread. 834 of them (22%) can’t be measured cleanly: 303 never triggered, and the rest were ambiguous or arrived with price already outside the zone. We count them separately, without making up a result.
Verpips data as of Oct 7, 2026, measured against the real price. See how we measure and the channel directory.
Types and variations
- Manual backtest
- Scrolling through the chart by hand. Slow and easy to bias.
- Automated backtest
- A program applies the rules to years of data.
- Forward test (paper trading)
- Testing the rules in real time, without real money.
- Signal audit
- A backtest of what a channel actually posted, using its messages as the source.
Why it matters when choosing a signal channel
It’s the only way to know how someone following a channel would have done. But a sloppy backtest (candles instead of ticks, no costs, or missing the deleted signals) produces results that look better than reality.
Common traps
- Using candles instead of ticks and assuming the target always got hit first.
- Forgetting the spread and commission.
- Ignoring signals the channel deleted afterward (survivorship bias).
- Tweaking the rules until the past looks good (overfitting).
Key takeaways
- ✓Backtesting = testing rules on past data.
- ✓To know what got hit first, you need ticks.
- ✓Without costs or deleted signals, the result gets inflated.
- ✓A good backtest also tells you what it couldn’t measure.
FAQ: backtesting
Is backtesting reliable?
Only as reliable as its data and rules: with ticks, costs and every signal, it gets very close to reality; with candles and no costs, it doesn’t.
What’s the difference between backtesting and forward testing?
Backtesting uses past data; forward testing applies the rules in real time, going forward.
Can you backtest a Telegram channel?
Yes: that’s what we do. Every posted signal is measured against real prices from the second it was posted.
Channel by channel
See these numbers for each channel, not as an average.
With a free account you see each channel’s result in R, its drawdown, and what it announced versus what actually happened.