Changelog
Scoring rules change in public.
The CopyGrade Score model is tuned continuously against real wallet outcomes. Every material change — to the scoring math, the farming detectors, or the product — is recorded here, dated, so the way a score is computed is never a moving target you can't see.
July 29, 2026
Telegram push now sits with the Discord/Slack webhook on the Pro side of the line. Nothing changes about which alerts a free account receives — entry, exit and farming-risk alerts still fire on a free watchlist — only where they are delivered: email and the in-app feed.
- Telegram push delivery requires an active Pro subscription. The free tier keeps its three alert kinds, by email and in the in-app feed.
- Connecting a Telegram chat is available on Pro. Free accounts see the option with an upgrade link rather than a control that would deliver nothing.
- If a Pro subscription lapses, the linked chat is kept, not disconnected: pushes pause and resume automatically on re-subscribe, with nothing to reconnect.
- The free public Telegram channel is unaffected — it stays open to everyone.
July 28, 2026
Pro drops from $25/month to $15/month ($12/month billed annually, $144/year). The price now reflects what Pro actually is after the repositioning: the real-time signal tier — per-minute alerts, confluence and specialist signals, unlimited watchlist, saved screens, baskets, CSV export — on top of an analysis layer that is free for everyone.
- New subscriptions are $15/month or $144/year (a 20% annual saving, as before). The 14-day first-subscription trial is unchanged.
- Existing subscribers keep their current price automatically — nobody is charged more, and anyone who prefers the new price can cancel and resubscribe or contact us.
- Nothing about the product or the model changed with this entry — pricing only.
July 27, 2026
The score paywall is gone: the full Copy Verdict — the 0–100 CopyGrade Score, five sub-scores, farming-risk forensics, category edge, and realistic post-fee return — is now free on every wallet, for everyone, including the public wallet pages and leaderboards. Pro is now about acting on the analysis: alerts, unlimited watchlist, saved screens, baskets, and CSV export.
- The free tier's 3-lifetime-unlock model is retired. Nothing is gated on a verdict any more, signed in or not.
- Public wallet pages, leaderboards, share cards and the embeddable badge now show the real numeric score (previously band-only).
- The Copy Simulator is now available to every signed-in account, on any wallet.
- Nothing about the model changed — no factor, weight, detector, or threshold moved. This is a packaging change only.
July 25, 2026
An ingestion defect had been writing duplicate copies of the same Polymarket fill, leaving roughly a quarter of stored trades as phantom rows. Every wallet has been re-scored on the corrected tape, and scores moved — mostly downward in farming risk.
- What was wrong: the sync's de-duplication key was built from a timestamp that changed shape once it had been written to and read back from the database, so an already-stored fill did not match itself on the next run and was inserted again. About 24% of rows in the trades table were duplicates.
- Why it moved scores: farming detectors read repetition in the trade history — self-trade wash, iceberg accumulation, copy-bait cadence. Duplicated fills look like repeated activity, so they pushed wallets toward a flag that their real history does not support.
- Measured effect, holding the model constant: across the same 2,016 wallets, the farming-flagged rate fell from 61.7% to 47.8% and the severe rate from 52.6% to 37.4%. 281 wallets moved from flagged to clean; 2 moved the other way. Median CopyGrade Score rose about 3 points.
- If you looked at a wallet before 2026-07-24 and it carried a farming flag, re-check it — there is a real chance the flag was ours, not the wallet's.
- No weight, threshold or detector changed. This was a data-quality fix; the model is the same one described on /methodology.
July 24, 2026
A rename, not a model change: the 0–100 metric formerly called the Copy Score is now the CopyGrade Score. Same inputs, same weights, same numbers — only the name is new.
- Why: “Copy Score” turned out to be a crowded, generic name in the copy-trading space; naming the metric after CopyGrade itself makes it unambiguous whose methodology produced the number.
- Nothing about the model changed with this entry — no factor, weight, detector, or threshold moved. Scores computed before and after the rename are directly comparable.
- Older entries on this page now use the new name for readability; where they say CopyGrade Score, the metric was called Copy Score at the time.
July 18, 2026
The fee-adjusted (“realistic”) edge previously used a simplified fee assumption. It now applies Polymarket's actual published rates, which makes realistic edge lower — and more honest — for most wallets.
- Realistic edge is what we estimate a copier keeps after costs. Charging the real schedule rather than an approximation moved the median realistic edge down across the scored population.
- This is a methodology change, and it makes our own headline numbers worse. Realistic-edge figures are therefore not directly comparable with editions published before this date.
- Headline (pre-cost) edge is unaffected — only the fee-adjusted figure moved.
July 17, 2026
A wallet whose stored history does not reach back across a full window used to have its windowed stats rendered as if it did. Those figures are now labelled as floors, from the wallet's real tape start.
- Per-wallet trade history is fetched up to a ceiling. For very active wallets the tape can be truncated, which means a “90-day” figure was really “as far back as we have.”
- Windowed profit, volume and trade counts are now derived from the wallet's actual coverage and presented as floors where the tape is short, so a truncated history reads as a lower bound rather than a complete one.
- This can change windowed numbers on deep, high-frequency wallets. Nothing about the score's weights changed.
July 1, 2026
The first 30-day forward cohort matured. The honest read: no measurable rank relationship yet between CopyGrade Score and the realized 30-day return of a simulated copier. We committed to publishing this either way, before we saw it.
- Across 336 matured, active 30-day outcomes, the rank correlation between CopyGrade Score and realized copier return is Spearman ρ = −0.001 — indistinguishable from zero, and consistent with the 7-day read (ρ = −0.039). The live band table on /methodology#track-record shows the full breakdown.
- The highest-scored band's negative average is driven by roughly four outlier wallets out of ~25 (its median outcome is 0.00%), which is why the verdict is read on rank correlation, not averages — and the rank correlation is ~0.
- What did not change: the model weights. We do not retune a documented v1 heuristic on one thin, outlier-heavy read — that would be fitting noise. The calibration job runs propose-only, and any re-weighting must improve a held-out check and clear a deliberate operator promotion before it ships.
- The CopyGrade Score keeps its 'documented v1 heuristic, openly stated as unvalidated' label. Next step is a monthly re-run of the same public check; a signal has to persist across reads before we call it one.
June 10, 2026
Farming flags are heuristic assessments of public on-chain data, not verified findings — the wording now says so, and wallet operators get a visible way to dispute one.
- The severe flag now reads “High farming risk” with “N signals detected” (previously “Farming detected” with “N confirmed signals”) — detectors flag patterns past a threshold; nothing is manually confirmed.
- The watch-level flag reads “Emerging signals — not conclusive” (previously “Under review”, which implied a human review that doesn't happen).
- The lowest verdict band reads “Avoid — farming risk” (previously “Avoid — farms followers”), and the severe farming alert email is headlined “…is now flagged for farming” (previously “…is farming its copiers”) — both asserted conduct rather than risk. Already-scored wallets show the new band label as they re-score.
- Every full flag card links to the methodology, and every verdict page now carries a standing note that scores, verdicts, and flags are algorithmic opinions — with a dispute contact for wallet operators.
June 2, 2026
The first pass over real Polymarket data showed the v1 model — tuned on synthetic patterns — over-flagged farming and over-rewarded thin, no-edge records. Recalibrated.
- Farming detectors tightened against real behaviour: self-trade wash now needs 8 near-price buy/sell pairs (was 3) and manufactured-volume churn 350 trades at <1.5% edge (was 200 at <2%), and churn is weighted lower — so an active-but-unprofitable wallet reads “watch”, not “severe”.
- A wallet with too little recent history (under ~20 trades) or near-zero realistic edge can no longer read as a copy candidate, however clean its risk metrics — copying a no-edge or unproven wallet nets nothing.
- Fixed the return denominator to include open-position cost basis and clamped the displayed edge, drawdown, and Sharpe — so a large held position can no longer render an impossible figure like -500% edge.
June 1, 2026
The simulator's equity curve was a deterministic synthesis that never touched the wallet's trades. It now replays the wallet's actual fills.
- The backtest is a FIFO replay of the wallet's real, timestamped trades — scaled to your capital and discounted by modelled fees, slippage, and copy latency.
- Wallets with too few synced fills fall back to the prior heuristic, flagged plainly in the UI rather than presented as a real replay.
- Inputs are URL-encoded so any run is shareable, and two wallets can be compared side by side under one identical set of conditions.
May 29, 2026
Farming risk moved from a flat count of signals to a severity-weighted rollup, and gained two single-wallet detectors plus a cross-wallet pass.
- Each finding is now weighted by severity (how bad if real) and confidence (how far past its threshold the pattern sits), so one unmistakable signature can veto a score while two borderline ones only warn.
- Two new single-wallet detectors: pump-and-dump and copy-bait.
- A cross-wallet pass flags decoy clusters and coordination by timing and opposite-side correlation, folded into each member's farming level.
May 26, 2026
The CopyGrade Score's return and risk factors were rebuilt to reflect what a copier actually keeps, not the headline number.
- Returns are now net of Polymarket's taker fee, dollar-weighted, and matched FIFO — not gross headline PnL.
- Risk-adjusted return uses an annualized, frequency-aware Sharpe ratio with shrinkage for short trade histories.
- A latency haircut discounts edge that only exists if you fill instantly — which a copier, arriving after the leader, never does.