Each case is one accrual: a balance, a rate and the seconds since the last update. The deployed integer maths, recompiled from pinned sources and checked byte for byte against mainnet, gives the new balance. 200-digit arithmetic gives the formula the protocol states: Aave's erate·years for borrows and per-second compounding for supply, and Morpho's e∫rate along its adaptive curve's own stated rate path. FP127 and FP19 compute that same formula from the same integers.
Normal operation is accurate: the deployed interest is right to within the token's own decimals. The digits go where the stated formula and the shipped shortcut diverge: a fully used market left alone for a day or a week, a rate at its cap.
| case | true interest | deployed off by | deployed digits | token grid allows | FP127 digits | FP19 digits |
|---|---|---|---|---|---|---|
| Aave v3 | ||||||
| Aave v3 · USDC, 100% used, borrow · 1 d | 397.339 USDC | +1.23e-07 USDC | 9.51 | 8.90 | 35.42 | 67.97 |
| Aave v3 · USDC, 100% used, supply · 1 d | 357.598 USDC | -0.0639 USDC | 3.75 | 8.85 | 30.83 | 63.28 |
| Aave v3 · USDC, 100% used, borrow · 7 d | 2,784.69 USDC | -2.16e-06 USDC | 9.11 | 9.75 | 35.98 | 69.06 |
| Aave v3 · USDC, 100% used, supply · 7 d | 2,505.87 USDC | -3.13 USDC | 2.90 | 9.70 | 30.83 | 63.27 |
| Aave v3 · WETH, 100% used, borrow · 7 d | 1,573.84 WETH | -2.55e-07 WETH | 9.79 | 21.50 | 35.90 | 68.66 |
| Aave v3 · LINK, 100% used, borrow · 1 h | 350.518 LINK | -6.29e-10 LINK | 11.75 | 20.85 | 35.05 | 67.96 |
| Aave v3 · LINK, 100% used, supply · 1 h | 280.405 LINK | -0.0393 LINK | 3.85 | 20.75 | 31.57 | 64.80 |
| Aave v3 · LINK, 100% used, borrow · 1 d | 8,446.43 LINK | -0.000209 LINK | 7.61 | 22.23 | 37.25 | 69.36 |
| Aave v3 · LINK, 100% used, supply · 1 d | 6,751.46 LINK | -22.7 LINK | 2.47 | 22.13 | 31.57 | 64.80 |
| Aave v3 · LINK, 100% used, borrow · 7 d | 60,644.5 LINK | -0.507 LINK | 5.08 | 23.08 | 37.67 | 70.20 |
| Aave v3 · LINK, 100% used, supply · 7 d | 48,228.3 LINK | -1.13e+03 LINK | 1.63 | 22.98 | 31.56 | 64.79 |
| Aave v3 · LINK, 100% used, borrow · 30 d | 287,019 LINK | -178 LINK | 3.21 | 23.76 | 38.17 | 70.90 |
| Aave v3 · LINK, 100% used, supply · 30 d | 223,680 LINK | -2.18e+04 LINK | 1.01 | 23.65 | 31.52 | 64.75 |
| Aave v3 · 5%, borrow · 365 d | 51,271.1 USDC | -0.263 USDC | 5.29 | 11.01 | 37.20 | 70.17 |
| Morpho Blue | ||||||
| Morpho Blue · at target, 4% · 1 d | 1,095.95 USDC | -6.39e-07 USDC | 9.23 | 9.34 | 30.10 | 67.42 |
| Morpho Blue · at target, 4% · 7 d | 7,674.18 USDC | -8.18e-07 USDC | 9.97 | 10.19 | 30.10 | 69.89 |
| Morpho Blue · 95% used, 10% · 1 d | 7,091.84 USDC | +0.577 USDC | 4.09 | 10.15 | 30.29 | 68.30 |
| Morpho Blue · 95% used, 10% · 7 d | 61,708.5 USDC | +275 USDC | 2.35 | 11.09 | 30.29 | 70.21 |
| Morpho Blue · 100% used, 10% · 1 d | 11,751.9 USDC | +3.08 USDC | 3.58 | 10.37 | 30.29 | 68.54 |
| Morpho Blue · 100% used, 10% · 7 d | 129,539 USDC | +2.48e+03 USDC | 1.72 | 11.41 | 30.29 | 70.00 |
| Morpho Blue · 100% used, 4% · 7 d | 51,615.5 USDC | +984 USDC | 1.72 | 11.01 | 30.09 | 69.48 |
| Morpho Blue · 100% used, 4% · 30 d | 2.07329e+06 USDC | +3.74e+05 USDC | 0.74 | 12.62 | 30.13 | 71.08 |
| Morpho Blue · 100% used, at the cap · 1 d | 221,598 USDC | -0.0966 USDC | 6.36 | 11.65 | 31.97 | 71.03 |
| Morpho Blue · 100% used, at the cap · 7 d | 1.6582e+06 USDC | -238 USDC | 3.84 | 12.52 | 31.95 | 71.44 |
| Morpho Blue · 100% used, at the cap · 30 d | 9.30027e+06 USDC | -8.94e+04 USDC | 2.02 | 13.27 | 31.84 | 71.75 |
| Morpho Blue · 80% used, 10% · 7 d | 16,689.7 USDC | +5.03 USDC | 3.52 | 10.52 | 30.29 | 69.22 |
| Morpho Blue · 5% · 365 d | 512.711 18-dec token | -0.00263 18-dec token | 5.29 | — | 37.81 | 70.30 |
| Morpho Blue · 800% · 7 d | 1,658.2 18-dec token | -0.238 18-dec token | 3.84 | — | 37.74 | 70.94 |
Hover a case for its parameters. Off by is deployed minus the stated formula, in tokens: negative is less interest than the formula. Most of Morpho's error after a day or a week idle is not in a library call: its adaptive curve integrates the rate with a two-panel trapezoid, which overcharges while the rate at target moves.
Every live market
The pending accrual of every Aave v3 reserve side (74) and every Morpho Blue market (252) at the survey block, with the harness reproducing each deployed result exactly. Live intervals are short, so these are high. The provenance, the claims quoted from each protocol's source and the split of every error into approximation and rounding: accrual-deployed.md.