Why is my trading strategy not working?
It looked great in the backtest. Live, it loses or goes nowhere. That is the normal story, not bad luck - and the reasons are usually boring and checkable.
1. The costs in the backtest were too small
Fees, spreads, slippage and funding add up. Many strategies have a real but thin edge, and the real costs eat it. In Indian F&O, SEBI found that charges were about 27% of traders' losses (FY22-24). A backtest with zero or tiny costs is a story, not a test.
Check: add 0.2% per trade (fees + slippage) and look at the profit factor again. If it falls below about 1.1, the edge lives in the costs you left out.
2. A few lucky days made all the money
Many strategies earn most of their profit on a handful of unusual days - a crash, a squeeze, a news spike. Those days may not come back while you trade it.
Check: remove the 5 best days (not trades) and recount. If the strategy stops making money, it was carried by events, not by a repeatable edge.
3. You tried many versions and kept the best one
Test 20 settings and keep the winner, and luck alone will usually give you something that looks good. In our own study of classic TradingView strategies, picking the best setting from a grid raised the backtest profit factor from about 1.15 to 1.54 - and in the next two unseen years it was about 1.0 either way. The tuning gain was an illusion.
Check: count every version you tested, and judge the final one on a period you did not use to design it.
4. The market changed under the strategy
A rule can work in one kind of market and fail in another. Classic trend rules looked profitable over long histories partly because markets rose; in our tests the Golden Cross was profitable in 9 of 10 markets over ten years but made money in 0 of 10 over the next two unseen years.
Check: split your history into two halves. If one half has all the profit, the edge is fading or was never steady.
5. The backtest saw the future
The most common backtest bug is look-ahead: using a bar's close to decide, and filling at that same close. It produces impossible results - very high win rates with tiny drawdowns.
Check: win rate above 85% with profit factor above 5 on 30+ trades is a red flag. Make sure orders fill on the next bar.
6. Live trading isn't the backtest
Live, you skip signals (asleep, nervous, busy), size up after wins, and stop after losses. The backtest never does. Small differences in behaviour turn a thin edge into a loss.
Check: paper-trade the next 30 signals exactly as tested, write down every skipped signal, and compare the win rate, average trade and worst fall with the backtest.
What to expect even when it is real: strategies that passed our checks kept making money in unseen data about 7 times in 10, against about 3.5 in 10 for an average backtest - and usually with about half the backtest edge. Plan on half.
Frequently asked questions
- How many trades do I need before I can judge a strategy?
- Under 30 trades almost any result can be luck. 100 or more starts to mean something. Paper-trade until you reach at least 30.
- My strategy wins 80% of the time but still loses money. How?
- A high win rate with rare large losses can lose overall. Look at the profit factor (money won divided by money lost) and the worst trade, not just how often you win.
- Should I keep tweaking the settings until it works?
- Usually no. Each tweak makes luck more likely. Change the idea, not the numbers, and test the final version on fresh data.
- Is it the market or my strategy?
- Split the history in two and compare. If the strategy only worked in one kind of market, expect long flat or losing spells when conditions change.
Want your own strategy checked?
Send your trade list. You get a plain-English PDF: verdict, 10 checks, a position size for your chosen risk, a daily stop and a 30-trade paper-test plan. Free in the beta: run it yourself in the Workspace.