Bitfinex lending strategy backtest results — 3 years of data
Every funding strategy Stratum ships is replayed against 1,095 days of real historical Bitfinex funding-market data. This page publishes the resulting dataset — average net APR, order fill rate and capital utilization per strategy and currency — so you can compare approaches with numbers instead of marketing claims. You are free to cite it with attribution.
Read the numbers with the methodology below: backtests are optimistic by construction (full-size fills, no market impact). They rank strategies well; they overstate absolute returns. Real out-of-sample live returns on liquid stablecoin books have been regime-bound around 8.5–10% APR.
USD funding
The deepest Bitfinex funding book. Stablecoin-style dollar lending against margin traders; rates track the Flash Return Rate (FRR), Bitfinex’s hourly base funding rate.
| Strategy | Avg APR (net) | Order fill rate | Utilization |
|---|---|---|---|
| Yield Curve Best in backtest | 10.75% | 21.7% | 97.0% |
| Adaptive Carry | 9.63% | 5.9% | 73.1% |
| Adaptive Ladder | 8.36% | 64.6% | 98.2% |
| Rate Ladder | 7.77% | 61.1% | 98.4% |
| Maturity Ladder | 7.21% | 48.2% | 95.7% |
| Carry Optimizer | 6.54% | 64.0% | 98.1% |
| Spike Catcher | 6.03% | 17.6% | 64.0% |
| FRR + Offset | 4.83% | 82.0% | 99.0% |
| Simple FRR | 4.83% | 82.0% | 99.0% |
| Adaptive Spread | 4.08% | 72.0% | 98.1% |
| Rate Spread | 3.97% | 54.3% | 96.2% |
USDT (UST) funding
Tether funding, listed as UST on Bitfinex. Slightly thinner than USD but typically a comparable rate regime; the same strategies apply.
| Strategy | Avg APR (net) | Order fill rate | Utilization |
|---|---|---|---|
| Yield Curve Best in backtest | 9.80% | 17.8% | 95.6% |
| Rate Ladder | 7.91% | 54.8% | 98.0% |
| Adaptive Ladder | 7.54% | 62.9% | 98.1% |
| Maturity Ladder | 7.06% | 46.4% | 95.0% |
| Carry Optimizer | 6.63% | 53.8% | 97.1% |
| Adaptive Carry | 6.52% | 2.5% | 49.4% |
| Spike Catcher | 6.26% | 11.5% | 53.8% |
| FRR + Offset | 5.17% | 72.8% | 98.3% |
| Simple FRR | 5.17% | 72.8% | 98.3% |
| Adaptive Spread | 4.25% | 68.7% | 97.8% |
| Rate Spread | 4.17% | 52.4% | 95.9% |
BTC funding
Crypto funding rates are far lower than dollar rates — borrowing demand for BTC is structurally weaker. Absolute APRs here are small; the interesting signal is the ranking between strategies.
| Strategy | Avg APR (net) | Order fill rate | Utilization |
|---|---|---|---|
| Yield Curve Best in backtest | 0.25% | 17.3% | 83.9% |
| Spike Catcher | 0.22% | 1.6% | 36.2% |
| Adaptive Ladder | 0.19% | 48.4% | 94.0% |
| Carry Optimizer | 0.18% | 52.0% | 95.6% |
| Rate Ladder | 0.10% | 43.5% | 95.0% |
| Maturity Ladder | 0.08% | 45.0% | 79.3% |
| Adaptive Carry | 0.08% | 0.7% | 13.7% |
| FRR + Offset | 0.04% | 69.7% | 94.2% |
| Simple FRR | 0.04% | 69.7% | 94.2% |
| Rate Spread | 0.03% | 42.0% | 91.9% |
| Adaptive Spread | 0.03% | 57.1% | 94.9% |
ETH funding
Like BTC, ETH funding earns much less than dollar funding and is more episodic — returns cluster around demand spikes rather than a steady carry.
| Strategy | Avg APR (net) | Order fill rate | Utilization |
|---|---|---|---|
| Yield Curve Best in backtest | 1.49% | 7.4% | 78.8% |
| Spike Catcher | 1.45% | 1.1% | 31.3% |
| Rate Ladder | 1.22% | 28.6% | 89.8% |
| Adaptive Ladder | 0.93% | 38.0% | 90.5% |
| Carry Optimizer | 0.66% | 29.8% | 88.6% |
| Maturity Ladder | 0.63% | 30.7% | 80.0% |
| FRR + Offset | 0.52% | 46.2% | 91.3% |
| Simple FRR | 0.52% | 46.2% | 91.3% |
| Adaptive Carry | 0.48% | 0.6% | 15.3% |
| Rate Spread | 0.42% | 39.1% | 91.9% |
| Adaptive Spread | 0.40% | 46.9% | 93.6% |
Avg APR is the annualized average return over the full window, net of Bitfinex’s 15% fee on earned interest. Order fill rate is the share of simulated offers that filled. Utilization is the share of capital out on loan, on average. Rows with a high APR but a very low fill rate depend on rare fills and are the most exposed to backtest optimism.
How these numbers are computed
Each figure comes from the Stratum backtest engine replaying 1,095 days (~3 years) of historical Bitfinex funding-market data — recorded rates and funding candles — per currency. The engine simulates each strategy placing real offers: it accrues interest continuously per loan, respects loan tenors, models early returns, and reports APR net of Bitfinex’s 15% fee on earned interest. The replay window and fill/fee model are identical for every strategy, so rows within a table are directly comparable.
Be blunt about the limits: these backtests are optimistic. They assume offers fill at full size at the recorded rate, with no market impact — your own size never moves the book and never loses queue priority to other lenders. Across regimes we have observed, real out-of-sample live returns on liquid stablecoin books are regime-bound around 8.5–10% APR at sensible defaults, regardless of the headline backtest number. Treat the tables as a ranking instrument, not a forecast.
Two more honesty notes. First, strategies with a high APR but a very low fill rate (event-driven approaches that wait for rate spikes) extrapolate from few fills — they carry the widest gap between backtest and live. Second, Bitfinex funding can be returned early by the borrower, so long-tenor offers do not reliably earn their full term. Platform basics for context: the FRR (Flash Return Rate) is Bitfinex’s hourly base funding rate, the minimum offer is roughly $150 equivalent, and a daily rate annualizes as daily rate × 365.
Free to cite — with attribution and a link
You are welcome to use these tables in articles, research notes, comparison pages or videos. The only condition: attribute Stratum and link to this page, so your readers can check the methodology and the most recent snapshot. The dataset is regenerated from the same engine when strategies or the replay window change; this page always carries the snapshot date.
Or link naturally in text, e.g. “according to Stratum’s 3-year Bitfinex lending backtest data” with the anchor pointing at this URL.
License: free to cite and quote with attribution to Stratum and a link to this page. Do not republish the full dataset without the methodology caveats.
Embed this data
Add this strategy leaderboard card to your own site or blog with a single static iframe — no JavaScript required. Free to use as long as the "Data: Stratum" attribution link stays intact.
FAQ
How were these backtest results computed?
The Stratum backtest engine replays 1,095 days of historical Bitfinex funding-market data per currency and simulates each strategy placing offers against the recorded rates. Interest accrues per loan, tenors and early returns are modeled, and reported APR is net of Bitfinex’s 15% fee on earned interest. Every strategy uses the same window and the same fill and fee model, so results are directly comparable within a currency.
Why are backtest APRs higher than what I should expect live?
Backtests assume every offer fills at full size at the recorded rate, with no market impact and no competition for queue position. Live, your size moves the book and other lenders undercut you. Out-of-sample live returns on liquid stablecoin books have been regime-bound around 8.5–10% APR at sensible defaults. The backtest is reliable for ranking strategies against each other, not for predicting your absolute return.
Can I reproduce these numbers?
Largely, yes. The underlying funding-market history is public via the Bitfinex API (funding candles and stats endpoints), and the methodology is described on this page: a 1,095-day replay, continuous per-loan interest accrual, net of the 15% fee. Stratum subscribers can run the same backtests in-app on the same engine. Independent reimplementations will differ in fill assumptions — which is exactly why we publish fill rate and utilization next to APR.
Run these strategies on your own account
Stratum places funding offers through your scoped, withdrawal-disabled Bitfinex API key — non-custodial, flat $9–$99/mo, no cut of your earnings. Paper mode lets you watch a strategy run before committing real capital.