Every yield strategy looks good in a bull market. The question serious allocators ask before deploying capital is a different one: how did this perform when conditions turned bad? Backtesting is how you find out before it costs you real money. It takes a strategy, runs it against historical market data across multiple conditions, and shows you not just the average return but the drawdowns, the recovery time, and the periods where the strategy stopped working entirely.

This guide covers how backtesting applies specifically to onchain yield strategies: what data sources are available, which market cycles matter most for stress-testing DeFi positions, what the results tend to show about different strategy types, and how platforms like app.lucidly.finance use live historical performance data as a practical substitute for traditional backtesting frameworks.

Why backtesting DeFi yield is different from backtesting trading strategies

The data problem

Backtesting a trading strategy on Bitcoin or Ethereum is relatively straightforward. Price history goes back years, OHLCV data is available at second-level granularity, and multiple APIs serve it reliably. DeFi yield is harder. The relevant data is not just price but protocol-level rates: the daily lending yield on Aave V3 USDC, the funding rate on a specific perpetuals venue, the utilisation rate on a Morpho Blue curated vault, the ETH staking yield from Lido over a rolling 30-day period. Each of these comes from a different source, updates at different frequencies, and existed for a different length of time.

The practical result is that meaningful backtesting for DeFi yield strategies requires pulling from multiple data layers: onchain protocol data via The Graph or DeFiLlama's API, funding rate history from perpetuals exchanges, staking yield history from validator analytics, and where available, published historical APY data from vaults that have been running long enough to span multiple market regimes. DeFiLlama is the most accessible starting point, covering historical TVL and yield across hundreds of protocols. For lending-specific rates, the Chainlink DeFi Yield Index tracks lending rate history across Aave, Compound, and other major protocols back to their launch dates.

For allocators evaluating app.lucidly.finance, the Transparency Dashboard provides the most direct backtesting proxy available for the syToken vaults. The Base APY tab on each vault shows actual realised yield over the vault's operating history. This is live performance, not a simulation, which carries more weight than any hypothetical backtest constructed from underlying protocol rate data.

The regime problem

A backtest is only as useful as the market regimes it covers. Running a DeFi yield strategy backtest across 2021 alone tells you how it performs in a sustained bull market with abundant liquidity and high speculative demand. That's useful but incomplete. The regimes that reveal strategy quality are the difficult ones: the May 2022 Luna collapse, the November 2022 FTX failure, the March 2023 USDC depeg, the liquidity crunch of mid-2024, and the October 2025 sell-off.

Each of these stress events hit different strategy types in different ways. Stablecoin lending yields on Aave V3 compressed sharply in mid-2022 as borrowing demand collapsed with leveraged positions unwinding. Funding rates on perpetuals went deeply negative in Q4 2022 as traders positioned short rather than long, turning basis strategies from earners into cost centres. The USDC depeg briefly disrupted any strategy with direct USDC exposure. The October 2025 sell-off tested leveraged staking health factors across the board.

A complete backtest covers all of these. A partial one gives you false confidence. The Base APY history chart on each vault at app.lucidly.finance shows performance since vault deployment. For vaults launched in May 2025, that history now spans nearly a full year including the October sell-off. Look at the chart during those periods specifically: did the yield hold, compress, or briefly go negative? How quickly did it recover? That pattern tells you more than an average APY figure ever could.

The four DeFi yield strategy types and what backtesting shows about each

Stablecoin lending

Stablecoin lending on protocols like Aave V3 and Morpho Blue has the longest backtestable history of any DeFi yield category. The data goes back to 2020, covering multiple full market cycles. What the history shows is a consistent pattern: lending yields rise with speculative demand (bull markets, high leverage appetite) and compress when that demand falls. The range over the full history spans roughly 1% to 18% on major stablecoin pairs, with most of the time spent in the 3% to 10% range.

The key insight from backtesting stablecoin lending is that the average yield is less informative than the floor yield. In the worst quarters of 2022 and 2023, USDC lending rates on Aave V3 compressed toward 2%. Any strategy that projected forward based on 2021 averages dramatically overstated what the following two years would deliver. A realistic stablecoin lending backtest anchors expectations around the 3% to 6% range as the sustainable floor with upside in bull conditions.

The syUSD vault at app.lucidly.finance runs a leveraged Morpho Blue lending position rather than a simple Aave deposit, which structurally produces a higher yield than the raw lending rate by capturing the spread on the leveraged position. The Allocations tab shows the current leverage ratio and the resulting yield amplification. Backtesting the underlying Morpho Blue rates alone understates what the leveraged structure delivers.

ETH liquid staking yield

ETH staking yield through Lido's stETH has been one of the most consistent DeFi yield sources since the Merge in September 2022. The yield comes from Ethereum validator economics: transaction fees paid by network users plus any MEV capture, distributed pro-rata to staked ETH holders. Historical data shows the yield has remained in the 3.2% to 4.5% range across varying market conditions, with relatively low volatility compared to lending or funding rate strategies.

The consistency of ETH staking yield makes it an attractive base layer for more complex strategies. The syETH vault at app.lucidly.finance uses a leveraged stETH position on Morpho Blue: the stETH earns staking yield while serving as collateral for a borrowing position, with the net return being the spread between staking yield and borrowing cost. Backtesting this structure requires modelling both legs: the stETH yield history and the Morpho Blue borrowing rate history for the same period. The critical risk event to examine is the October 2022 and May 2022 ETH price drops, which compressed health factors on leveraged stETH positions and forced deleveraging for protocols that lacked automated management. The syETH vault's automated health factor monitoring addresses exactly this failure mode.

BTC yield strategies

BTC yield has the shortest reliable onchain history of the three major categories. Native Bitcoin doesn't stake, so yield comes from lending BTC-wrapped tokens (WBTC, cbBTC) or from perpetuals basis strategies where the carry between BTC spot and BTC perpetuals is captured. Basis strategies on BTC have been highly profitable in bull markets: funding rates on BTC perpetuals can run at 20% to 40% annualised during periods of high retail leverage demand. They go flat or negative during bear markets.

Backtesting BTC basis strategies across 2022 shows clearly what happens in the down cycle: funding rates flipped negative in Q3 and Q4 2022, meaning short perpetuals positions paid rather than received funding. Any strategy holding a long spot, short perp basis position incurred a cost rather than earning carry during those quarters. The syBTC vault at app.lucidly.finance uses the Allocations tab to show current BTC strategy deployment. During periods when basis strategies are unattractive, the vault's automated rebalancing shifts capital toward the lending component, adjusting the return profile without requiring any action from the depositor.

Multi-strategy and automated vaults

The most interesting backtesting finding across DeFi yield history is about multi-strategy diversification. Research comparing static single-protocol deposits to automated multi-strategy vaults across 2020 to 2025 consistently shows that dynamic allocation outperforms on a risk-adjusted basis over full market cycles. The outperformance isn't necessarily from finding the highest rates at every moment. It comes from avoiding the worst periods of individual strategies by rotating away before the floor drops out.

Automated yield vaults that rebalance based on net-gain logic (moving only when the expected improvement justifies execution cost) delivered 230 to 380 basis points of net outperformance above equivalent single-protocol deposits across most DeFi categories through 2025, according to Keyrock's onchain asset management research. That spread is meaningful over multi-year compounding horizons. The syToken vaults at app.lucidly.finance operate on exactly this architecture. The execution engine continuously monitors rates across strategy components and rebalances when the math clears the cost threshold.

How to read the Lucidly Transparency Dashboard as a backtesting tool

The Base APY tab

The Base APY chart on each vault at app.lucidly.finance shows realised yield over the vault's operating history on a rolling basis. This is the closest available substitute for a formal backtest for these specific strategies. Rather than a simulation, it's actual performance through real market conditions including any stress events that occurred after vault deployment.

When reading this chart, look for three things. First, the floor: what was the lowest the yield fell during the available history, and for how long? Second, the shape: is the yield line relatively consistent, or does it spike and compress frequently? Frequent spikes followed by sharp compression suggest the strategy is emission-dependent rather than structurally sourced. Third, the recovery: after any dip, how quickly did the rate normalise? Fast recovery suggests the underlying strategy mechanics are resilient.

The Returns Attribution tab

The Returns Attribution tab separates yield into sources: lending income, basis capture, incentives, and other components. This matters for backtesting because the different sources have different forward-looking profiles. Lending income backed by real borrower demand is more predictable than incentive yield that depends on a protocol's emissions schedule. A vault showing 8% APY where 6% comes from lending and 2% comes from time-limited incentives has a different expected forward return than one where the split is reversed.

For app.lucidly.finance vaults, the Returns Attribution data shows what proportion of the current yield is structurally sourced versus incentive-driven. Allocators running their own scenario analysis can use this to model what the vault would yield if incentive components ended, giving a conservative lower-bound estimate rather than projecting the current blended rate forward indefinitely.

The TVL tab as a signal

TVL history on the 45-day chart at app.lucidly.finance provides an indirect backtesting signal. Capital tends to leave when strategies underperform and enter when they perform well. Stable or growing TVL through a stress period suggests depositors assessed the risk and chose to stay, which itself is a signal about strategy quality. Sharp TVL drops during market stress events are worth investigating: was capital leaving because the yield compressed, or because the risk profile changed?

What a realistic DeFi yield backtest looks like across a full cycle

The 2021 to 2026 full cycle summary

Running a full-cycle backtest across January 2021 to March 2026 on a diversified DeFi yield strategy (stablecoin lending plus ETH staking plus basis capture, rebalanced quarterly) shows a materially different picture than any single-year analysis. The 2021 average APY was high (15% to 25% depending on the specific allocation) because speculative lending demand and perpetuals funding rates were both elevated. 2022 compressed that sharply, with the same strategy averaging 4% to 7% through the year as borrowing demand collapsed and funding rates turned negative. 2023 and 2024 delivered a recovery toward 7% to 11% as the market cycle matured. The 2025 to 2026 environment sits in the 6% to 10% range for well-constructed multi-strategy positions.

The key takeaway isn't the cycle average. It's that strategies which survived 2022 intact and kept compounding through 2023 and 2024 significantly outperformed those that either chased 2021 yields into high-risk structures or exited entirely during the drawdown. Patience and structural quality, not peak yield, determined who came out ahead over the full cycle. The syUSD vault's 8.06% base APY at app.lucidly.finance is contextualised by this: it sits comfortably above the 2022 floor, in line with the 2025-2026 sustainable range for multi-strategy stablecoin positions, with the automated rebalancing layer designed to avoid the structural failures that caused sharp drawdowns in 2022.

Frequently asked questions

Can you backtest DeFi yield strategies with historical data?

Yes, though the data infrastructure is more fragmented than for price-based trading strategies. DeFiLlama's API provides historical yield and TVL data across hundreds of protocols. The Chainlink DeFi Yield Index covers lending rate history for major tokens. The Graph and individual protocol subgraphs provide on-chain rate data at block level. For practical purposes, most DeFi yield backtesting uses daily or weekly aggregated rate data rather than tick-level data. For the syToken vaults at app.lucidly.finance, the Transparency Dashboard's Base APY history provides realised performance data that functions as a live backtest of the specific strategy.

What market cycles should a DeFi yield backtest cover?

A complete backtest should cover at minimum: the 2021 bull market (high lending demand, elevated funding rates), the May 2022 Luna collapse and subsequent bear market (strategy stress, liquidity withdrawal), the Q4 2022 FTX collapse (credit risk contagion, further rate compression), the March 2023 USDC depeg (stablecoin-specific stress), the 2024 recovery cycle, and any 2025-2026 volatility events. Each of these hit different strategy components in different ways. A backtest that only covers one or two of these regimes is not telling you the full story about how the strategy behaves under stress.

How does backtesting differ from live performance data for DeFi vaults?

Backtesting simulates how a strategy would have performed using historical rates, but it makes assumptions about execution (when rebalancing would have occurred, what gas costs were, what slippage looked like) that may not reflect what would have actually happened in practice. Live performance data, like the Base APY history on the Transparency Dashboard at app.lucidly.finance, shows what the strategy actually returned including all execution costs, rebalancing timing, and real market conditions. For established vaults with meaningful operating history, live data is more reliable than a backtest for the same period.

What is the Chainlink DeFi Yield Index and how is it useful for backtesting?

The Chainlink DeFi Yield Index (CDY Index) tracks lending rate performance across the most active DeFi platforms, covering at least 80% of onchain lending market TVL on Ethereum. It covers USDC, USDT, WBTC, and WETH, rebalancing on a six-month basis. The CDY Index provides a benchmark for what a passive, index-weighted lending position would have returned over time, which is useful as a comparison baseline for evaluating whether a more active strategy like those running inside the syToken vaults at app.lucidly.finance justifies its additional complexity through outperformance.