> ## Documentation Index > Fetch the complete documentation index at: https://docs.pathprotocol.finance/llms.txt > Use this file to discover all available pages before exploring further. # Liquidity impact > Predicted APY after a deposit, not just predicted APY. Closes the question every institutional allocator asks first. The single most common question we get from institutional allocators is some version of: > "If this pool has \$19M in TVL and I deposit \$50M, the APY will crunch. Does your model predict that?" It is the right question. Predicting that a pool's APY trends from 6% to 7% over the next 7 days is useless if a \$50M deposit drops it back to 4% on entry. Liquidity-impact modelling is therefore a first-class concept inside Signal Mesh, not an afterthought. ## What we surface Every prediction comes with a liquidity-impact context block that answers four questions for an allocator considering a deposit: For a hypothetical deposit of size *X*, what percentage of pool TVL would you represent? Above 10% of TVL is a yellow flag; above 25% is a red flag. Using the protocol's actual rate mechanism (utilisation curve for lending, swap-fee math for AMM-style yield, governance-set rate for stability rails), we compute the APY the pool would settle at immediately after your deposit lands. For pools that route through an AMM at deposit time, the expected slippage on entry, derived from the live order-book / curve depth snapshot. Combining the prediction-direction signal with the post-deposit APY estimate and the gas cost of moving, the daemon outputs a break-even horizon: "this move pays back in N days" where N depends on direction confidence and deposit size. ## How we predict it The post-deposit APY estimate is derived analytically from the rate mechanism rather than learned end-to-end. We invert the protocol's own rate formula given the current pool state and your deposit size. For lending pools that means the utilisation curve maps deposited supply to a borrow rate, which maps back to a lender APY via the reserve factor. For AMM-style yield pools the math is different but the principle is the same: the protocol's mechanism is public, so the post-deposit APY is computable, not predictable. What is *predictable* is the second-order effect: other allocators will respond to your deposit. The directional model captures that through the cohort-flow signal block (the percentage of deposits in the next 24h that historically follow large entries on similar pools). ## Current performance The post-deposit APY estimate matches realised post-deposit APY within \~1 percentage point in 6 cases out of 10 on backtest data, and the gap shrinks materially when restricted to pools where Signal Mesh has high signal coverage. We continue to improve this surface with two specific work streams: * Wider cohort-flow data ingestion, particularly for institutional allocators whose deposit clustering tells us more than the average retail flow. * Per-protocol rate-mechanism specialists for the long tail of pools with non-standard rate formulas. The honest framing for an institutional partner is that **Path is opinionated about whether you should move and gives you a defensible estimate of what your APY will look like after you move**. The combined directional + liquidity-impact view is what a Strategy Manager would consume.