The Lab
Every trading day at 9:25 ET the Lab forecasts where SPY will be at 10:00, noon, 2:00 and the close, then learns from what actually happened. The bar it has to clear is the simplest honest forecast: the latest pre-market price holds.
Skill at the close +1.3% ± 1.1% · promising, not proven
Edge hit rate 63% direction beyond the latest price, 64 leaned days
Range coverage 70% target 68%
Days learned 137 first 15 are warmup and not counted
Call
Fri, Oct 23, 2026 Graded All inputs inClose target $759.63, range $755.49 to $763.80, 20% chance of closing above the prior close ($763.96). Closed at $759.97.
| Checkpoint | P(up) | Target | Range | Actual | In range | vs naive |
|---|---|---|---|---|---|---|
| 10:00 | 32% | $762.83 | $761.33 – $764.35 | $762.61 | Yes | Beat (0.186 vs 0.391) |
| Noon | 15% | $760.28 | $757.22 – $763.35 | $762.67 | Yes | Lost (0.361 vs 0.331) |
| 2:00 | 17% | $759.99 | $756.59 – $763.40 | $762.52 | Yes | Lost (0.327 vs 0.292) |
| Close | 20% | $759.63 | $755.49 – $763.80 | $759.97 | Yes | Beat (0.219 vs 0.240) |
pushes downpushes up
- Long bonds (TLT) yesterday
- IWM minus QQQ yesterday
- Gamma above vs below spot
- Distance to gamma flip
- Headline risk tone (Jev)
- Fed leaning hawkish (Jev)
- VIX futures (VIXY) yesterday
- Net dealer gamma
Track record
-10.0%-5.0%+0.0%+5.0%+10.0%
naive+1.3%
May 22Jul 21Sep 15Nov 10
40%60%80%100%
68% target65%
May 22Jul 21Sep 15Nov 10
- Kalman: everything, slow memory 19%
- Gaps keep going 15%
- Analog days (10 nearest) 10%
- Kalman: price action 9%
- Kalman: options positioning 9%
- Everyone else 38%
0%50%100%
Apr 27Jul 1Sep 4Nov 10
0%50%100%
25
16
17
19
45
0%25%50%75%100%
Leaderboard
Average CRPS at the close in sigma units (lower is better), and each expert's edge over "latest price holds". Weight is how much the ensemble listens to it today.
| Expert | CRPS | vs naive | Weight | Verdict |
|---|---|---|---|---|
| Kalman: everything, slow memory | 0.5165 | +0.0105 ± 0.0099 | 19.2% | ~315 more days to confirm |
| Gaps keep going | 0.5187 | +0.0083 ± 0.0077 | 14.8% | ~300 more days to confirm |
| Kalman: options positioning | 0.5240 | +0.0031 ± 0.0075 | 8.9% | ~2795 more days to confirm |
| Kalman: price action | 0.5253 | +0.0017 ± 0.0107 | 9.0% | ~18676 more days to confirm |
| Analog days (10 nearest) | 0.5259 | +0.0011 ± 0.0076 | 10.5% | ~23004 more days to confirm |
| Latest price holds (random walk) | 0.5270 | — | 7.9% | the benchmark |
| Yesterday's move reverses | 0.5274 | -0.0003 ± 0.0050 | 7.0% | no edge so far |
| Kalman: Jev text reads | 0.5299 | -0.0029 ± 0.0050 | 6.8% | no edge so far |
| Dealer hedging drift | 0.5331 | -0.0061 ± 0.0044 | 5.3% | no edge so far |
| Yesterday's direction continues | 0.5338 | -0.0067 ± 0.0050 | 6.1% | no edge so far |
| Kalman: everything, fast memory | 0.5403 | -0.0133 ± 0.0150 | 2.6% | no edge so far |
| Gaps fade | 0.5522 | -0.0252 ± 0.0084 | 2.0% | no edge so far |
By checkpoint
| Checkpoint | Skill | From range | From edge | Coverage | Edge hit rate |
|---|---|---|---|---|---|
| 10:00 | +9.1% ± 2.2% | +2.3% | +6.7% | 72% | 75% (76) |
| 12:00 | +1.3% ± 0.9% | +0.5% | +0.8% | 68% | 69% (42) |
| 14:00 | +1.9% ± 0.8% | +1.5% | +0.4% | 70% | 56% (36) |
| Close | +1.3% ± 1.1% | +0.4% | +0.9% | 70% | 63% (64) |
How it learns
- Senses. Options positioning from the 9:25 snapshot, yesterday's moves in stocks, bonds, credit, the dollar and VIX, and Jev's reads of the morning headlines and WSB.
- Experts. Simple fixed hypotheses ("gaps fade"), Kalman filters whose weights drift slowly with the market, and an analog expert that asks what happened on the most similar past days.
- Ensemble. Experts earn weight by forecasting better (Hedge), and every expert keeps a small share so one that fell out of favor can come back when the market changes.
- Range. The range widens after a close lands outside it and narrows after one lands inside, so over time about 68% of closes fall inside it.