Audit Report — Entry Sizing and Fill Tracking
Date: 2026-06-07 Scope: All market-maker strategies in the strategy pack
Question 1 — Does the strategy scale into a 50% position, or buy it all at once?
Most trend / mean-reversion MMs already scale in (DCA ladder gated by ATR or % drop). The strategies that DO NOT scale in are the symmetric-quoting market makers — they post bid and ask simultaneously each cycle at full target size, so when only one side fills, the bot can land an unintended lump position.
Lump-sum candidates that need a scale-in patch:
- cartaJaiMicroMM.js
- glostenMilgrom.js
- inventoryBalancer.js
- microPriceMM.js
Already scaling in (verified): adxTrendMM, atrChannelMM, bbSqueezeMM, cciMeanRevMM, donchianBreakoutMM, emaBounceMM, ichimokuCloudMM, keltnerChannelMM, macdMomentumMM, obvMomentumMM, parabolicSarMM, rsiMeanRevMM, stochMeanRevMM, superTrendMM, twapScalperMM, vwapMM, williamsPctRMM.
Question 2 — Does the internal fill counter actually match real fills?
Strategies where the internal counter increments on order PLACEMENT (not on actual fill confirmation), causing the bot to act as if it has more on than it does:
- harrisScalpMM.js
- hawkesMM.js
- kalmanGrid.js
- ofiFlashMM.js
- queueReactiveMM.js
- rollModelMM.js
Fill-accurate (verified): all DCA-ladder strategies that gate the counter behind a validated-order block, plus the symmetric MMs that reconcile against inventory delta (cartaJai, glostenMilgrom, inventoryBalancer, microPriceMM).
Patch plan
Two passes, one strategy at a time, each with a layman patch note:
- Pass A — Scale-in conversion (4 files)
- Pass B — Fill-counter reconciliation (6 files)