Week Ahead: Funding Rate Predictions
The Ornstein-Uhlenbeck fits are in. Here's what the data says about the week starting March 23.
The Signal: 206 Pairs Modeled, 15 Worth Watching
Our OU engine ran overnight across all active perpetual markets with sufficient settlement history. Of 206 pairs that cleared the R² > 0.5 threshold, the top 15 by fit quality break into three distinct regimes — and the split matters.
| Exchange | Symbol | Persistence (β) | Equilibrium (%) | Positive % | R² | Half-Life (h) |
|---|---|---|---|---|---|---|
| Hyperliquid | BABYUSDT | 0.7049 | -0.003 | 50.0 | 0.962 | 15.8 |
| Hyperliquid | DOTUSDT | 0.9564 | -0.0035 | 33.9 | 0.911 | 124.4 |
| Hyperliquid | VIRTUALUSDT | 0.9259 | -0.0002 | 65.6 | 0.898 | 72.1 |
| Binance | LRCUSDT | 0.8954 | -0.0132 | 42.2 | 0.897 | 50.2 |
| Hyperliquid | KPEPEUSDT | 0.9203 | -0.0013 | 32.8 | 0.884 | 66.8 |
| Binance | SAHARAUSDT | 0.9392 | -0.1015 | 0.0 | 0.881 | 88.5 |
| Hyperliquid | WIFUSDT | 0.9401 | -0.0092 | 2.8 | 0.875 | 89.7 |
| Binance | YALAUSDT | 0.9966 | 2.0419 | 93.4 | 0.891 | 1,617.4 |
Rows filtered to R² > 0.87 for readability. Full table: 206 pairs.
Regime 1: Degenerate Fits (R² = 1.000)
Five pairs — Bybit STRKPERP, NOTPERP, EDUUSDT, KuCoin SOLUSD, and Binance DMCUSDT — posted perfect R² scores with half-lives of 99,999 hours. That's not signal. That's structural ceiling rates. These instruments are pinned to their exchange minimum (0.01%), likely due to low open interest or exchange-side capping. They look predictable because they never move. Ignore them for carry strategies this week.
Regime 2: Fast Mean-Reversion (Half-Life < 72h)
Hyperliquid BABYUSDT is the standout. Half-life of 15.8 hours, R² of 0.962, equilibrium sitting at -0.003%. It reverts fast and the fit is tight. Fifty percent positive rate suggests it oscillates symmetrically around zero — classic noise-driven mean reversion, not trend. Good candidate for short-duration delta-neutral carry if spreads are manageable.
VIRTUALUSDT (72.1h half-life, 65.6% positive rate, equilibrium -0.0002%) is the other fast mover in this group. Slightly positive bias historically, nearly zero equilibrium. Rates here are likely to oscillate around flat this week — neither a strong long nor short carry case, but low regime-risk.
Regime 3: Persistent Negative Drift
Seven of the top-fit pairs have equilibria firmly negative. Some by a little; some by a lot.
SAHARAUSDT on Binance is extreme: equilibrium at -0.1015%, 0% positive settlements, 88.5-hour half-life. Every single historical settlement was negative. This is a structurally bearish perp — shorts consistently get paid. Whether that persists into next week depends on whether the underlying keeps underperforming spot, but the OU model is confident in the direction.
WIFUSDT on Hyperliquid shows similar structure: 2.8% positive rate, -0.0092% equilibrium, half-life of 90 hours. Memecoin funding drift in one direction.
LRCUSDT on Binance (equilibrium -0.0132%, R² 0.897) sits between the two — meaningful negative drift but not extreme. Half-life of 50 hours means it reverts but doesn't snap back fast.
The Outlier: YALAUSDT
This one stands alone. Equilibrium of 2.04% per settlement, 93.4% positive rate, R² of 0.891, half-life of 1,617 hours. That's not a typo. The OU model fitted a rate that is structurally, persistently elevated — and mean-reverts over 67 days. It's not "mean-reverting" in any practical trading sense. It's a carry opportunity with a very slow decay.
The 2.04% equilibrium rate annualizes to roughly 890% APR at 8-hour intervals, or 178% at 24-hour. Those numbers are almost certainly inflated by a short data history — YALA is a recent listing. Treat with caution. But if the structure holds, this is worth monitoring for a long-spot, short-perp setup.
What the Model Doesn't Know
OU parameters are backward-looking. They describe what happened. They say nothing about macro catalysts, liquidity events, or whether a token is about to get delisted. SAHARAUSDT's persistent negativity could reverse overnight if sentiment shifts. BABYUSDT's clean reversion could break down if volume collapses.
The model is a filter, not a forecast. Use it to rank which pairs deserve attention, not as a trade signal on its own.
Week Ahead: Three Things to Watch
- BABYUSDT half-life compression. If it drops below 10 hours, the noise is amplifying — good for scalp, bad for multi-day carry.
- SAHARAUSDT floor test. Any settlement above -0.05% would be a regime break worth flagging.
- YALAUSDT data accumulation. The R² is already 0.891 with limited history. A week of additional data will either validate the structure or collapse it.
Funding rate patterns are regime-dependent. The regimes are shifting.
Explore the data:
- All Funding Rate Data — compare current rates across every pair
- Predict Markets — trade these setups directly
- ETH/USDT Funding Rate — track ETH's rate dynamics
The research arm of Zirodelta. Data-driven analysis of crypto sentiment markets, model development, and market microstructure research. Data-driven. Real-time. Across 6 exchanges and 3,700+ perpetual futures.
Learn more about Zirodelta →