Strategies I designed and built at Lucidly Labs, 2024–26

Yashish Khurana
LinkedIn · X · GitHub

These are DeFi yield strategies I designed and built over two years: a fixed-yield delta-neutral book, a stablecoin loop, a lending allocator and a hedged LP. The production code belongs to Lucidly, so each one here is a rewrite: the allocation and risk rules in plain Python, a replay on bundled data, and a page of results.

The strategies

dnENA — fixed yield plus a short — Pendle PT-sENA, ENA short on Hyperliquid at 1x, 15% reserve.
Real data. 422 rolling 88-day windows, start dates Jun 2025 – Jul 2026: 85% positive, median +4.5% annualised, worst −27%. The hedge was lost in 105 windows, when the Jun–Sep 2025 ENA rally outran the reserve. In the reserve sweep a 50% reserve loses it in none, at a lower yield. The Feb–Apr 2026 window on its own made +1.45%.

Leveraged siUSD/USDC loop — Morpho Blue, 1x to 8x.
Real data, 22 Mar – 22 Jun 2026. Every row was positive, but leverage didn't help: +1.31% at 1x, +0.85% at 8x, because the borrow rate ran above the collateral yield on many days. The gate levers on net yield after borrow, not the raw spread, and de-levered four times at 8x.

USDC lending cluster — Morpho Blue, 3 markets.
Real data, same window. +0.48% over 93 days. The allocator is mean-variance on paper, but the hurdles and caps are what bind: on average only 38% of the book was deployed.

WBTC/WETH concentrated liquidity with a hedge — Uniswap v3 and Aave, Arbitrum.
One real day, 9,940 swaps, fees measured from the pool's events: about 5× what a flat 12%/yr fee assumption gives. +0.0108 BTC against holding 10 BTC, and the hedge halved equity vol. One day is one day.