Track record: Intraday Hunter
His actual trades land 57.3% at open to close; the 95% interval includes 50%. Direction is not a proven edge. The one slice that holds up: trend days.
Method: T08b (2026-08), pre-registered primaries at open to close, exact binomial vs 50%, Benjamini-Hochberg. Direction is the sign of the move with a 0.1% dead band; a long call is right when the move is up, a short call when it is down.
Pre-registered test on his recorded trades
| Subject | Horizon | Days | Right | Hit rate | 95% interval | p vs coin flip |
|---|---|---|---|---|---|---|
| His actual trades | open to close | 110 | 63 | 57.3% | 47.9 to 66.1% | 0.152 |
| His actual trades | open to +60m | 110 | 62 | 56.4% | 47.0 to 65.3% | 0.215 |
| His evening plan bias | open to close | 208 | 90 | 43.3% | 36.7 to 50.1% | 0.061 |
| Generic technical-analysis model | open to close | 178 | 63 | 35.4% | 28.7 to 42.7% | 0.0001 |
Slices worth watching (small samples)
| Slice | Horizon | Days | Hit rate | p |
|---|---|---|---|---|
| His actual trades on trend days (|open to close| >= 0.5%) | open to +60m | 37 | 75.7% | 0.0026 |
| His actual trades on big gap-up opens (small sample) | open to close | 13 | 92.3% | 0.0034 |
The engine's own 09:45 read agreed with his actual call on 72 paired days 66.7% of the time. That measures faithfulness to him, not edge.
Reads run on this site
Every read anyone runs here is logged and scored against the bars that followed; repeats of the same moment count once. "Grounded" reads are mornings on the markets he trades; "extrapolated" reads apply his logic where he has no recorded evidence. A day he actually traded is marked in-sample: his own trade on that day was part of what built the profile, even though the engine never sees that day's trade or anything after it. No-trade reads are counted separately: right when the index stayed flat.
| Evidence | Sample | Horizon | Directional reads | Right | Hit rate | 95% interval | No-trade reads | Flat rate |
|---|---|---|---|---|---|---|---|---|
| grounded | out-of-sample | h30 | 3 | 1 | 33.3% | 6 to 79% | 0 | - |
| grounded | out-of-sample | h60 | 3 | 3 | 100.0% | 44 to 100% | 0 | - |
| grounded | out-of-sample | close | 3 | 1 | 33.3% | 6 to 79% | 0 | - |
| grounded | in-sample (a day he traded) | h30 | 1 | 1 | 100.0% | 21 to 100% | 0 | - |
| grounded | in-sample (a day he traded) | h60 | 1 | 1 | 100.0% | 21 to 100% | 0 | - |
| grounded | in-sample (a day he traded) | close | 1 | 1 | 100.0% | 21 to 100% | 0 | - |
4 distinct moments read so far; model reads 2, the rest deterministic.
Recent reads
| Instrument | Day | Time | Bias | Evidence | +30m | +60m | Close | |
|---|---|---|---|---|---|---|---|---|
| bank nifty | 2026-09-23 | 09:50 | lean long | grounded | up (right) | up (right) | up (right) | open |
| bank nifty | 2026-09-22 | 10:05 | lean short | grounded (in-sample) | down (right) | down (right) | down (right) | open |
| bank nifty | 2026-08-04 | 09:45 | lean short | grounded | flat (wrong) | down (right) | up (wrong) | open |
| bank nifty | 2026-08-04 | 09:45 | lean short | grounded | flat (wrong) | down (right) | up (wrong) | open |
| bank nifty | 2026-08-04 | 09:45 | lean short | grounded | flat (wrong) | down (right) | up (wrong) | open |
| bank nifty | 2026-08-04 | 09:45 | lean short | grounded | flat (wrong) | down (right) | up (wrong) | open |
| bank nifty | 2026-08-04 | 09:45 | lean short | grounded | flat (wrong) | down (right) | up (wrong) | open |
| bank nifty | 2026-08-04 | 09:45 | lean short | grounded | flat (wrong) | down (right) | up (wrong) | open |
| bank nifty | 2026-08-04 | 09:45 | lean short | grounded | flat (wrong) | down (right) | up (wrong) | open |
| bank nifty | 2026-08-04 | 09:45 | lean short | grounded | flat (wrong) | down (right) | up (wrong) | open |
| bank nifty | 2026-08-04 | 09:45 | lean short | grounded | flat (wrong) | down (right) | up (wrong) | open |
| bank nifty | 2026-08-04 | 09:45 | lean short | grounded | flat (wrong) | down (right) | up (wrong) | open |