Results: how the rules really did
495 trading days (05 Nov 2024 to 05 Nov 2026), 2,391 NSE stocks, the same fixed rules for every trade. Rules version 0.23.0; the test is rerun every month with the rules in use then.
About the names. The famous names are credit for publicly described ideas. Every test here uses our own simplified, fully coded version of the idea, so the numbers show how our code behaved on past prices, not how the named person trades or traded. No affiliation with, or endorsement by, any named person.
Backtests are simulations on past prices; real trading has costs, slippage and taxes, and past results, real or simulated, do not predict future results. Not investment advice.
By market day: 🟢 GOOD, 🟡 CAUTION, 🔴 BAD
Before any stock is looked at, the market is scored on 7 checks (Nifty trend, VIX, breadth, distribution days). How the tags work. Across all rule picks the average winner was +1.19R and the average loser -0.95R.
| Market day | Trades | Win rate | Avg result (R) | vs Nifty | Beat the Nifty |
|---|
| All days | 1,460 | 40% | -0.10 | -0.70% | 39% |
| 🟢 GOOD market | 309 | 40% | -0.10 | -0.43% | 41% |
| 🟡 CAUTION market | 941 | 41% | -0.08 | -0.44% | 40% |
| 🔴 BAD market | 205 | 34% | -0.22 | -2.31% | 31% |
A win rate near 40% is normal for trend-following. What decides the result is the size of the winners against the losers. Is a 40% win rate bad?
Every setup
Do more setups agreeing help?
| Setups on the same day | Trades | Win rate | Avg (R) | vs Nifty |
|---|
| 1+ setups agreeing | 75,363 | 38% | -0.12 | -0.52% |
| 2+ setups agreeing | 26,291 | 39% | -0.11 | -0.50% |
| 3+ setups agreeing | 7,335 | 39% | -0.11 | -0.45% |
Would a different exit help?
| Exit | Win rate | Avg (R) | Avg win | Avg loss |
|---|
| Current exit (target 2.5 ATR, stop 1.5 ATR, 10 days) | 39% | -0.11 | +1.19R | -0.95R |
| Target at 3R, up to 20 days | 31% | -0.10 | +1.99R | -1.02R |
| No target, sell on the first close below the 10-day average | 31% | -0.14 | +1.37R | -0.83R |
| Half at day 3, stop to break-even, rest trails the 10-day average | 44% | -0.14 | +0.68R | -0.77R |
| Half at +2R, stop to break-even, rest trails the 20-day average | 32% | -0.12 | +1.63R | -0.94R |
Skipping stocks that already ran
| Filter | Trades | Win rate | Avg (R) | Avg per trade |
|---|
| Current rules | 1,460 | 40% | -0.10 | -0.77% |
| Skip overheated stocks (RSI above 70) | 1,338 | 40% | -0.08 | -0.55% |
| Skip jumpy stocks (daily range above 5%) | 1,445 | 41% | -0.08 | -0.51% |
| Skip stocks up 40%+ in 3 months | 1,399 | 42% | -0.03 | -0.23% |
| All three together | 1,308 | 40% | -0.09 | -0.46% |
Buying the dip instead of the breakout
| Entry | Filled | Trades | Win rate | Avg (R) |
|---|
| Current entry (buy above the last high) | 72% | 1,460 | 40% | -0.10 |
| Limit buy at the 10-day average | 9% | 189 | 22% | -0.22 |
| Limit buy at the 20-day average | 7% | 137 | 25% | -0.17 |
How the test works
- Entry above the last day's high; stop 1.5 ATR below; target 2.5 ATR above; out after 10 days; 0.1% costs.
- Technical rules only: past news and company results are not used.
- Delisted stocks are missing, so live results would likely be a little worse (why).
The live paper-trade record (90 days delayed) →