What is Conviction Score?
Reviewed by Karthikeyan Anbarasan (Founder, TradeLyser) · Updated 19 Oct 2026 · Editorial policy
How to score conviction
A conviction score only means something if each level has a written description. Otherwise it drifts with mood: a 4 after a winning streak and a 2 after three losses for the same chart. Tie the levels to checklist items you can see before the order goes in. The score then becomes a summary of the setup’s quality rather than of your mood, and two people looking at the same chart should arrive at roughly the same number. Record it in the trade entry at the moment you place the order, alongside the setup tag, and do not change it afterwards, even if the trade turns out very differently from what you expected.
| Score | Example description (opening range breakout) |
|---|---|
| 5 | Every checklist item met, higher-timeframe trend agrees, clean level, volume above average |
| 4 | Every item met, one supporting factor missing, such as average volume |
| 3 | Checklist passes, but one item is borderline, such as a level tested many times already |
| 2 | Minimum criteria only; trend neutral or against |
| 1 | Taken for a reason outside the plan; tag it so it can be reviewed separately |
Reading the record: a worked example
In Figure 1 the levels separate cleanly. The 35 high-conviction trades made +16.1R in total, the 51 medium trades +10.71R and the 34 low trades +1.36R. Together that is +28.17R over 120 trades, or +0.23R per trade. At ₹2,000 risked per trade, the low-conviction trades earned about ₹80 each on average, before charges, and the high-conviction trades about ₹920. Once charges are taken off, the low-conviction trades may well be losing money, which is a question worth answering before deciding whether to keep taking them at all.
Even this tidy result needs care. Around 35 trades per level is a small sample: with 30 trades, a measured 50% win rate still has a 95% range of about ±18 points. A gap that shows up in one quarter and disappears in the next was probably noise. Look for the same order across two separate reviews before acting on it. Check two more things before you trust the split. First, are the levels spread across setups, or are all the 5s one setup? If so, you are comparing setups, not conviction. Second, how are the scores spread over time? If high scores bunch up after winning weeks, the score may be tracking confidence from recent results rather than the quality of the chart.
Using it for sizing
Conviction score vs emotion grade
Conviction score
Emotion grade
Reference guide
| Context | Value | Reading |
|---|---|---|
| Scored before entry from written criteria | Usable data | The score reflects what you knew at the time, so it can be tested against results |
| Scored after the trade closed | Contaminated | Hindsight pushes winners up and losers down, so the score just repeats the result |
| Size raised on high conviction with no record behind it | Untested sizing | Wait until each level has 30 or more trades and the gap holds across two reviews |
Questions
What is a conviction score in trading?
A pre-trade rating, often 1 to 5, of how well a trade matches your best version of the setup. Its value comes from comparing results by level after enough trades.
Should I trade bigger on high-conviction trades?
Only if your own record shows high-conviction trades earn clearly more per trade, across at least two review periods with 30 or more trades per level. Until then, equal size is the safer default.
How many trades do I need before conviction scores mean anything?
Aim for 30 or more at each level, so 100 or more in total for a three-level split. Fewer than that and differences between levels are easily due to chance.
Is high conviction the same as overconfidence?
No. High conviction backed by written criteria and a record of results is calibrated judgement. Overconfidence is rating trades highly without that record, or after a winning streak.