If a method worked well in the past, will it work well in the future?
Backtesting, which involves applying methods to past prices, has pitfalls that are easy to overlook.
Table of contents
Remember this first
First, 3 things to remember
- 1A backtest is an experiment calculated using historical data.
- 2Using information that was unknown at the time inflates the results.
- 3You must also include commissions and failure cases.
Check if it is a test that uses future information.
Backtesting is the process of calculating, "What would have happened if I had used this method in the past?" If you use news released in the afternoon for a decision made that morning, the test becomes impossible in reality.
Do not use daily highs or revised data that were determined later as if you knew them at the time. You must only use data that was available at that moment.
Check if you only selected stocks that performed well.
Including only companies that have survived until now or stocks that have risen significantly can make results look good. Excluding delisted or failed stocks makes the results different from the actual situation.
Be careful about constantly changing numbers to select only the combination that fits the past best. Results may differ between the data used for testing and data you see for the first time.
Do not look only at how many times you won.
If you earned 1 dollar eight times out of ten and lost 5 dollars twice, you earned 8 dollars and lost 10 dollars, resulting in a total loss of 2 dollars. Even though the number of wins is high, your money has decreased.
Conditions such as commissions, slippage between desired and actual execution prices, and trading halts also change the results. This is why past calculation results do not guarantee actual profits.
| Result | Frequency | Profit/loss per trade | Total |
|---|---|---|---|
| Profit | 8 | +$1 | +$8 |
| Loss | 2 | −$5 | −$10 |
| All | 10 | Win rate 80% | −$2 |
You should not make such assumptions
“The win rate is 80%, so it is a profitable strategy.”
Even if the number of wins is high, the total amount can decrease if a single loss is large. Calculate costs and the amount lost as well.
Check what you have read.
- Did you use only the information available at that time?
- Did you include companies that failed or disappeared?
- Did you also calculate costs and the magnitude of losses?
Apply what you read to the actual screen
Try it on WAPE
Source and scope of verification
Reviewed · Version 2
- Types of OrdersNew window
SEC Investor.gov · Verify 2026-10-20
Difference between market price and last traded price · Price conditions for limit orders - WAPE mock trading user guideNew window
WAPE · Verify 2026-10-20
Differences between paper trading and live trading; personal record login · MA and session VWAP in lib/toss-chart.ts and lib/paper-chart-indicators.ts.
This is educational material for observation and information comprehension. Hypothetical examples do not represent actual stocks or performance and do not provide buy or sell recommendations.