Overview & Experiments 55 Synthesis Roadmap Lookahead Audit

Stablecoin-Drift — 7d Supply Δ as BTC Directional Signal

Inconclusive

2026-05-19 stablecoinmacrodirectionalon-chain-adjacent

Hypothesis

Δ in total stablecoin USD market cap (USDT+USDC+others) over the past 7 days predicts BTC forward 1d return at IC ≥ 0.05 walk-forward; top-quintile supply expansion days are followed by ≥40 bps higher fwd-1d return than baseline.

Verdict

**WEAK** — IC +0.0376, spread +25 bps. Daily-resolution stablecoin supply doesn't carry strong forward-return information at the 1d horizon. The lit's '1-30d delay' may be too slow to capture as a tradable edge.

pooled_n_1d

2,327

q1_mean_bps

-0.0693

q5_mean_bps

+25.2963

n_windows_1d

pooled_CI_1d

[-0.0028, +0.0739]

pooled_IC_1d

+0.0376

pooled_IC_4h

-0.0250

stability_1d

+0.5714

windows_pos_1d

q5_minus_q1_bps

+25.3657

Stablecoin-Drift — 7d Supply Δ as BTC Directional Signal

2026-05-19 · status: inconclusive · 3.9s

Hypothesis: Δ in total stablecoin USD market cap (USDT+USDC+others) over the past 7 days predicts BTC forward 1d return at IC ≥ 0.05 walk-forward; top-quintile supply expansion days are followed by ≥40 bps higher fwd-1d return than baseline.

Verdict: WEAK — IC +0.0376, spread +25 bps. Daily-resolution stablecoin supply doesn't carry strong forward-return information at the 1d horizon. The lit's '1-30d delay' may be too slow to capture as a tradable edge.

Key metrics

metric	value
pooled_IC_1d	`+0.0376`
pooled_CI_1d	`[-0.0028, +0.0739]`
pooled_IC_4h	`-0.0250`
pooled_n_1d	`2,327`
windows_pos_1d	`12`
n_windows_1d	`21`
stability_1d	`+0.5714`
q5_mean_bps	`+25.2963`
q1_mean_bps	`-0.0693`
q5_minus_q1_bps	`+25.3657`

Approach

Daily total stablecoin USD market cap (DefiLlama aggregate) → 7d log-Δ as causal feature. Tested against BTC fwd-1d log return (primary) and fwd-4h log return (secondary). Walk-forward 12mo train / 3mo test, 1-day embargo. Quintile analysis on daily, sampled-every-24h for 4h panel.

Pooled OOS IC

1d horizon: IC = +0.0376 (n=2,327, 95% CI [-0.0028, +0.0739], p=0.0696)
4h horizon: IC = -0.0250 (n=2,320, 95% CI [-0.0699, +0.0190], p=0.229)

Quintile table — fwd-1d BTC return by stablecoin Δ7d quintile

quantile	count	mean_bps	std
0	466	-0.1	0.0300444
1	465	-12	0.0247435
2	465	11.3	0.0356338
3	465	26.7	0.0333834
4	466	25.3	0.0371588

Q5 − Q1 spread: +25.4 bps
Q5 mean: +25.3 bps
Q1 mean: -0.1 bps

Per-window IC (1d horizon, 21 windows)

window	n	IC	ci_low	ci_high	p_value
2021-01-02 → 2021-04-02	90	0.0409	-0.1468	0.2411	0.7017
2021-04-02 → 2021-07-02	91	-0.1174	-0.3077	0.1204	0.2676
2021-07-02 → 2021-10-02	92	0.0363	-0.1655	0.2432	0.731
2021-10-02 → 2022-01-02	92	-0.1089	-0.3061	0.0917	0.3014
2022-01-02 → 2022-04-02	90	-0.0582	-0.2428	0.1408	0.586
2022-04-02 → 2022-07-02	91	0.0217	-0.1763	0.2365	0.8383
2022-07-02 → 2022-10-02	92	0.0324	-0.1661	0.2382	0.7591
2022-10-02 → 2023-01-02	92	-0.1918	-0.3828	0.0237	0.0671
2023-01-02 → 2023-04-02	90	-0.0241	-0.2269	0.1991	0.8217
2023-04-02 → 2023-07-02	91	0.0727	-0.1399	0.2595	0.4936
2023-07-02 → 2023-10-02	92	-0.0507	-0.2609	0.1655	0.6314
2023-10-02 → 2024-01-02	92	0.0262	-0.1753	0.2311	0.804
2024-01-02 → 2024-04-02	91	0.0764	-0.1511	0.2815	0.4716
2024-04-02 → 2024-07-02	91	-0.065	-0.2435	0.1445	0.5407
2024-07-02 → 2024-10-02	92	0.1975	-0.0233	0.4116	0.0591
2024-10-02 → 2025-01-02	92	0.0701	-0.1765	0.2771	0.5067
2025-01-02 → 2025-04-02	90	-0.1396	-0.3542	0.0724	0.1894
2025-04-02 → 2025-07-02	91	0.0021	-0.2234	0.2439	0.9841
2025-07-02 → 2025-10-02	92	-0.1784	-0.4239	0.0542	0.0888
2025-10-02 → 2026-01-02	92	0.0184	-0.207	0.2131	0.862
2026-01-02 → 2026-04-02	90	0.0103	-0.2245	0.2278	0.9234

Per-window IC (4h horizon, 21 windows)

window	n	IC	ci_low	ci_high
2021-01-09 → 2021-04-09	90	-0.0021	-0.233	0.216
2021-04-09 → 2021-07-09	91	0.0451	-0.1792	0.2541
2021-07-09 → 2021-10-09	92	0.2137	0.0203	0.4114
2021-10-09 → 2022-01-09	92	-0.1659	-0.3593	0.0465
2022-01-09 → 2022-04-09	90	-0.1413	-0.3425	0.0736
2022-04-09 → 2022-07-09	91	0.1344	-0.0626	0.3622
2022-07-09 → 2022-10-09	92	-0.0055	-0.2114	0.222
2022-10-09 → 2023-01-09	92	-0.1996	-0.4167	0.0334
2023-01-09 → 2023-04-09	90	0.0593	-0.1407	0.278
2023-04-09 → 2023-07-09	91	0.0138	-0.1922	0.224
2023-07-09 → 2023-10-09	92	0.0117	-0.1888	0.2304
2023-10-09 → 2024-01-09	92	0.0746	-0.1542	0.2534
2024-01-09 → 2024-04-09	91	0.0324	-0.1738	0.2432
2024-04-09 → 2024-07-09	91	-0.0519	-0.2578	0.1894
2024-07-09 → 2024-10-09	92	-0.0138	-0.2273	0.2095
2024-10-09 → 2025-01-09	92	0.0967	-0.0887	0.298
2025-01-09 → 2025-04-09	90	0.0524	-0.1756	0.2618
2025-04-09 → 2025-07-09	91	0.034	-0.1788	0.2356
2025-07-09 → 2025-10-09	92	0.0103	-0.2134	0.2409
2025-10-09 → 2026-01-09	92	-0.1055	-0.3154	0.1366
2026-01-09 → 2026-04-09	90	-0.0535	-0.2904	0.1739

per-window IC 1d

quintile chart