▶ THE AUDIT — WHAT WE FOUND, AND WHAT KARMA IS FOR
Trade from receipts, not vibes
karma exists to help you find good coins and, more to the point, good people, or at least people who won't rug you or dump on you, so you can open pump.fun with a real base to start from. it will not make you safe. the space is brutal and you can still get wrecked. it just means you start from what people actually did, instead of whatever the group chat is screaming.
1 · WE READ THE WHOLE ROOM
733 wallets · 5661 scored calls · 3925 coins
we pulled the caller space wide and looked at what actually happened to every call: did the coin survive, did anyone get a real chance to sell, did the caller cash out into the people they called in. that whole history is the ground we score on.
read that and the temptation is to call everyone a scammer. don't. this is the asset class, not the crime. the baseline is savage on its own, and grading anyone against zero would just be grading memecoins. so we don't. every score is measured against these numbers, against what a caller in this space normally does.
2 · WE GRADE CONDUCT, NOT LUCK
the thing everyone wants to sell you is a hit rate. it doesn't hold up. how well a caller picked this month barely tells you how they'll pick next month, so we show picking on every card and score none of it. what does hold up is character: whether they dump on the people they called in, and whether their coins die. we split each caller's history in half and check whether the first half predicts the second. that's the test below.
| TRAIT | HOLDS UP OVER TIME | GAP, WORST → BEST | SCORED? |
|---|---|---|---|
| dumps on the people they called in | yes | +14.1pp | scored |
| their coins die on you | yes | +25.0pp | scored |
| picks a 2× | no | +5.0pp | shown, not scored |
picking barely predicts itself, so it stays off the score. conduct predicts itself, so conduct is the score. a high grade doesn't mean someone picks winners. it means someone measurably less likely to turn you into their exit liquidity.
3 · A COIN THAT GAVE YOU A CHANCE ISN'T HELD AGAINST THEM
this is the part that makes the grade fair. a coin that ran and gave you a real window to sell at a profit before it faded is not counted against the caller. dying is not the crime. almost everything here dies. we only punish two things: dumping on your followers, and coins that were dead on arrival, that never gave anyone a chance to get out ahead.
the whole difference is “gave you a chance” versus “made you the exit.” this fairness rule is why the collapse rate above sits where it does and not higher: coins that died fair aren't charged as rugs.
4 · WE GRADE OURSELVES, AND PUBLISH WHERE WE'RE WRONG
a rating you can't check is just an opinion with a logo. so here is where our own board misses. a “blowup” is a wallet we ranked among the very best that then went on to behave worse than the market anyway. we count them, out loud, on this page, because that's the number that tells you whether to trust the rest.
| OUR TOP | RANKING | HARM VS MARKET | WE GOT WRONG |
|---|---|---|---|
| top 10 | raw | 39.9% | 3 of 10 |
| top 10 | shipped | 35.7% | 2 of 10 |
| top 25 | raw | 43.8% | 6 of 25 |
| top 25 | shipped | 43.5% | 6 of 25 |
| top 50 | raw | 48.7% | 13 of 50 |
| top 50 | shipped | 48.4% | 13 of 50 |
read the last column plainly: of the callers we ranked highest, this many still behaved worse than the market's 65.2% average. the shipped board puts 2 in its top 10, and we're not hiding it. our claim is “measurably less likely to hurt you.” it is never “safe,” and this table is exactly how far off “measurably” runs.
5 · MOST CALLERS GRADE POORLY, ON PURPOSE
48% of scored callers land D or F. that's not a harsh curve, it's what the chain says: most callers are exit liquidity, so most callers grade poorly. these are fixed cut points, not a quota. nobody gets bumped up to fill a slot, and the bar doesn't bend for anyone. the handful up top earned it.