The whole ledger, not a highlight reel
Every rating ChopArm publishes is logged the moment it fires and graded against what the market did next. This page recomputes from those logs on every load — the same numbers that drive the internal tuning loop. If a rating stops working, you’ll see it here before we do anything about it.
Singles ratings · every call, graded one scan later
Every rating is logged the moment it’s published and graded against what those cards actually did by the next scan (roughly a week). One row per rating — the same Strong Buy / Buy / Watch / Hold / Trim language you see everywhere on the site (Strong Buy grades inside the Buy row). Nothing cherry-picked; only data artifacts (moves beyond ±300%) are excluded.
RatingCallsAvg next moveHow it’s doing
buy591+1.08%Doing its job. 59% of Buy-rated cards were up by the next scan; a Buy is right when they rise.
trim1659-2.40%Doing its job. 74% of Trim-rated cards were down by the next scan; Trim (take profit) is right when they fall.
watch11220-0.08%Benchmark row — Watch makes no promise. Buy has to beat this line to be worth anything.
hold5201+0.57%Benchmark row — the do-nothing baseline.
new1204-1.98%New sets still settling — these should look roughly random.
Rip verdicts · open it or keep it sealed, checked 14+ days later
“Rip edge” = how the cards inside did minus how the sealed box did over the same window. “Rip now” is right when the edge is positive or flat (opening cost nothing); “hold sealed” is right when it’s negative (the box out-earned its cards).
VerdictnBox moveContents moveRip edge
either70-2.74%+0.69%+3.43ppNo claim — either choice was fine
hold sealed405+0.33%+2.10%+1.78ppWrong so far — the cards did better
rip now202-0.66%-0.04%+0.61ppRight so far — opening cost nothing