Every projection here was recorded before its outcome could exist; we then compare what the plan projected to what your scale actually showed, near each 28-day target. A miss can reflect your intake differing from the projection's assumption, not just the model. Weights in lb. Not medical advice.
Demo profile — this ledger is seeded for illustration from the same projection engine. A weigh-in logged below won't be saved for a demo profile; sign in to build your own.
Across 10 resolved projections, the typical miss was 1.3 lb (good). On average your real weight came in 0.2 lb heavier than projected. 100% of projections landed within 2.2 lb. Assuming no change would have missed by 3.7 lb, so the projection is adding real signal.
Log today's weight
Feeds the ledger above — one weigh-in per day. After two weigh-ins ≥ 2 weeks apart (plus a few logged meals), projections start opening and grade themselves as each 28-day target passes. Or connect WHOOP to auto-fill the days you don't log (a manual weigh-in always wins for its day).
Your daily calorie & macro target is a versioned assumption. v1 is the formula plan set at onboarding (Mifflin–St Jeor × activity); v2 is the adaptive plan inferred from your logged weigh-ins + meals — it tracks your measured metabolism, so it can differ from the textbook number.
Coverage: 10 resolved · 6 open · 0 expired (a closed target with no nearby weigh-in expires — shown, not hidden).
Each point is realized − projected weight (lb). The dashed line is zero — a perfect projection.
See where these projections come from on the What if…. The ledger grades the model against your self-reported weigh-ins, so it's only as honest as your logging.
Not medical advice. Demo data only.
| Made | Target | Projected | Realized | Error | Assumptions | Status |
|---|
| 2026-08-06 | 2026-09-03 | 199.5 lb | — | — | v2 | pending — resolves 2026-09-03 |
| 2026-07-30 | 2026-08-27 | 201.7 lb | — | — | v2 | pending — resolves 2026-08-27 |
| 2026-07-23 | 2026-08-20 | 200.8 lb | — | — | v2 | pending — resolves 2026-08-20 |
| 2026-07-16 | 2026-08-13 | 201.2 lb | — | — | v2 | pending — resolves 2026-08-13 |
| 2026-07-09 | 2026-08-06 | 201.4 lb | — | — | v2 | pending — resolves 2026-08-06 |
| 2026-07-02 | 2026-07-30 | 202.3 lb | — | — | v2 | pending — resolves 2026-07-30 |
| 2026-06-25 | 2026-07-23 | 205.9 lb | 204.8 lb | −1.1 lb | v2 | matched |
| 2026-06-18 | 2026-07-16 | 204.6 lb | 205.0 lb | +0.4 lb | v2 | matched |
| 2026-06-11 | 2026-07-09 | 207.8 lb | 206.0 lb | −1.8 lb | v2 | matched |
| 2026-06-04 | 2026-07-02 | 208.2 lb | 206.7 lb | −1.5 lb | v2 | matched |
| 2026-05-28 | 2026-06-25 | 208.2 lb | 208.8 lb | +0.6 lb | v2 | matched |
| 2026-05-21 | 2026-06-18 | 209.5 lb | 208.7 lb | −0.7 lb | v2 | matched |
Head-to-head: Across 10 weeks resolved by both versions, the challenger's typical miss was 1.3 lb vs 1.3 lb for the active plan — too close to call.
This grades the Twin's weight projections (not the calorie plan). We run two adaptive estimators side by side: v2 uses a 28-day maintenance window and is the active plan that steers your daily target (the headline above grades v2 only); v3 is a shadow challenger on a shorter 15-day window — more responsive, but noisier — that is graded against the same weigh-in history yet never changes your food. Comparing their typical miss is an honest test of whether a more responsive estimator would actually track better before we'd ever promote it. The challenger needs two weigh-ins a full 14 days apart inside its 15-day window, so it legitimately opens fewer weeks than v2 under sparse or mid-cadence logging — its open and expired weeks are counted above so that gap stays visible. The head-to-head line compares the two only on weeks where both were graded against the same weigh-in — an exact, like-for-like comparison — so its figures can differ from, and occasionally order differently than, the per-version rows above, which also count weeks only one version resolved or where each landed on a different nearby weigh-in (those shared-but-split weeks are disclosed in the line itself). If too few weeks share the exact same weigh-in to score precisely, the line says how many exact weeks are still needed and then adds an approximate comparison over every week both versions resolved (clearly labelled, since some of those weeks were graded against different weigh-ins), so an occasional logger still sees a signal. The separate v1→v2 history of your daily plan is in the plan-assumptions section above.