Eleven testable applications of Google’s TimesFM 3.0 (330M-param time-series foundation model) on real Bybit perpetual futures data — organized by edge type and expected value. Every test is measured as MAE vs the right naive baseline + p10–p90 coverage; if it doesn’t beat naive, it’s noise.
TimesFM-3 is a risk-input tool, not a price-direction predictor. The validated sandbox baseline shows well-calibrated uncertainty bands (p10–p90 coverage 81–84%), beats naive on funding level and volume, and has no directional skill at 3 days — which is exactly why the edge-test suite targets levels, volatility, cross-series structure, and mean reversion rather than price direction.
Status of the edge tests: early runs (Sep 9, 2026) were promising — realized-vol
forecasting for sizing (A1), quantile-width as a risk-regime meter (D2), real event
covariates (B2), band-breakout anomaly detection (D1), funding/carry timing (C1/A3), daily
range (C3), and session-volume timing (A2). But those evaluation windows were
misaligned (fixed in commit b986e09), so the numbers are
pending re-validation on the alignment-fixed notebook and must not be trusted as-is.
So far only C2 (vol-clustering persistence) and B3 (liquidation precursors) show no clear edge — both are shelved until more data.
Cheap, high-confidence tests that feed existing systems: vol-based sizing (A1), session-aware volume (A2), range-based stops/TP (C3), and the quantile-width risk meter (D2).
| Test | Direction | Status (repo HEAD) |
|---|---|---|
| A1 realized-vol → sizing | Predicted 24h vol; size inversely | Promising — pending re-validation |
| A2 volume → session timing | Rank hours for breakout/ORB entries | Promising — pending re-validation |
| C3 daily range → stops/TP | Predicted range as TP/stop multiplier | Promising — pending re-validation |
| D2 quantile-width → risk regime | Wider band → cut size / widen stops | Promising — pending re-validation |
Forecast 24h realized vol (rolling std of log returns) on BTC/ETH/SOL.
Vol is strongly autocorrelated; this is the model's best skill.
Size positions inversely to predicted vol instead of lagging ATR (per-trade risk input).
Beat naive vol-persistence MAE by >5%, coverage 75–85%.
Highest expected value. Early runs showed a strong signal; the fixed notebook must confirm the numbers.
Tier 1
Forecast hourly volume over the next 24–48h, session effects included (NY/Asia/London).
Volume is far more predictable than price; session structure (NY-open 8h edge) is real.
Rank which hours deserve breakout/ORB entries; avoid thin-book traps.
Beat naive per-hour volume; session-hour forecasts beat naive within those sessions.
Volume forecasting for session timing.
Tier 1
Forecast next-day high–low range from hourly context + volume/vol covariates.
Range is much more predictable than direction.
Stop and take-profit placement (TP at predicted-range multiples).
Beat naive range-persistence by >5%; actual range inside p10–p90 70–85%.
Predicted range as TP/stop distance multiplier.
Tier 1
Test whether the model's own p10–p90 spread predicts next-period realized vol.
If the model's uncertainty is forward-looking, its width is a vol-of-vol indicator.
As a risk-regime gauge for sizing (wider band = reduce size).
Correlation of quantile-width with subsequent realized vol > current-vol persistence.
Free — works on every forecast with no extra data.
Tier 1
Harder tests: BTC→alt lead-lag (B1), real known-future event covariates (B2), and funding-spike reversion timing (C1).
| Test | Direction | Status (repo HEAD) |
|---|---|---|
| B1 cross-asset BTC→alts | Hedge ratio / correlation input | Direction only — re-validation pending |
| B2 real event covariates | Unlock / FOMC / CPI calendar input | Promising — pending re-validation |
| C1 funding-spike reversion | Carry entry timing (post-spike) | Promising — pending re-validation |
Multivariate forecast SOL/ETH price conditioned on BTC as a past covariate; compare vs univariate.
If BTC history helps predict alts, it's a hedge-ratio / correlation input (not direction).
Portfolio construction, beta hedging.
Multivariate MAE < univariate MAE on alts by a meaningful margin.
Not in the repo's interim recommended list until the fixed notebook confirms it.
Tier 2
Use real known-future events with magnitudes — token unlocks, halving dates, FOMC/CPI calendar, listings/airdrops — as continuous past-future covariates, short horizon (24–72h).
Binary flags demonstrably didn't work; magnitudes may.
Event-aware sizing (skip/half-size into known events).
Forecast reacts to the event window AND that reaction beats naive.
The covariate mechanism showed a real signal with the funding-settlement schedule; next test is real known-future events.
Tier 2
After |funding| > 2σ, forecast time to revert and reversion magnitude.
Funding mean-reverts; level-forecasting is the model's strength.
Entry timing on carry trades; avoid paying high funding into a squeeze.
Reversion-timing MAE beats naive 'next-period revert' by >10%.
Post-spike reversion + funding-level forecasts as carry-entry timing inputs.
Tier 2
Cheap speculative tests: funding-level convergence (A3), liquidation precursors (B3), vol-clustering persistence (C2), and band-breakout anomaly detection (D1).
| Test | Direction | Status (repo HEAD) |
|---|---|---|
| A3 funding-level | Carry timing input | Promising — pending re-validation |
| B3 liquidation precursor | Pre-event warning flag | No clear edge so far |
| C2 vol-clustering persistence | ORB/breakout regime filter | No clear edge so far |
| D1 band-breakout anomaly | Regime-change alert for screener | Promising — pending re-validation |
Forecast funding 1–3 days out with funding history + spikes as covariates.
Predict when funding normalizes → carry entry timing.
Carry entry timing.
Beat naive on level + correct sign of convergence.
Grouped with C1 as funding/carry timing.
Tier 3
Feed funding + OI + realized vol as covariates; target near-future vol spike (4h max abs move).
Pre-event warning — reduce size / widen stops flag (not directional).
Pre-event warning — reduce size / widen stops flag (not directional).
Quantile width or vol forecast spikes significantly before actual vol spikes (lead time test).
Not recommended yet — revisit only with more data.
Tier 3
Forecast 6–24h vol given a recent vol spike (regime persistence).
Regime-state input for ORB/breakout filters.
Regime-state input for ORB/breakout filters.
Distinguishes spike-sustained vs spike-decay regimes better than AR(1) vol.
Not recommended yet — revisit only with more data.
Tier 3
When actual deviates outside the model's p10–p90 band, treat it as an anomaly / change-point event.
Band-breaks cluster before large moves (lead time > random).
Change-point detector rather than forecaster; regime-change alert for the screener.
Band-breaks cluster before large moves (lead time > random).
Actuals leaving the p10–p90 band as an anomaly/change-point flag.
Tier 3
All tests run in timesfm3_bybit_edge_tests.ipynb on the same data layer as the
sandbox (Bybit hourly klines + funding, with automatic Hyperliquid fallback). Walk-forward
windows are validated: context = forecast_start − CTX … forecast_start,
actual = forecast_start … forecast_end.
Persistence for vol/range, last-value for funding level, session-average for volume. A test passes only when it beats its own baseline.
≈80% is ideal. Funding and volume runs land 81–84% — the model's uncertainty intervals are trustworthy.
No look-ahead: context ends exactly where forecasting begins; 8–24 iterations per test depending on horizon.
log-price / log-volume scaled for the model, inverted for reporting. Rows below are in original units.
| Target | Series | MAE | naive MAE | Beat naive? | p10–90 cov% |
|---|---|---|---|---|---|
| Funding | BTC | 2.33e-6 | 3.59e-6 | ✅ | 84.1 |
| Volume | BTC | 905 | 1,001 | ✅ | 83.3 |
| Volume | ETH | 12,383 | 12,762 | ✅ | 82.6 |
| Volume | SOL | 125,645 | 158,730 | ✅ | 81.5 |
Consistent picture across runs: well-calibrated baselines, beats naive on funding level and volume, no price-direction skill at 3 days, and synthetic binary covariates ignored — which motivated the real-event test (B2).
b986e09. Per-test numbers from the earlier runs are not published here — they must be re-validated on the alignment-fixed notebook and recorded in HYPOTHESES.md before being trusted.Ranked by fit with the existing stack, as recorded in the repo.
Replace lagging-ATR input with TimesFM predicted 24h vol; size inversely. Vol is strongly autocorrelated — highest expected value.
The model's own p10–p90 width as a forward-looking uncertainty gauge — wider band → cut size / widen stops. Free, no extra data.
The covariate mechanism showed a real signal with the funding-settlement schedule; next test is real known-future events (unlock / FOMC / CPI) with magnitudes.
Actuals leaving the p10–p90 band as an anomaly/change-point flag for the screener.
Post-spike reversion and funding-level forecasts as carry-entry timing inputs.
Predicted range as TP/stop distance multiplier.
Forecast volume to rank which hours deserve breakout/ORB entries.
C2 (vol-clustering persistence) and B3 (liquidation precursors) — no clear edge; revisit only with more data.