This week the board changed jobs.
It used to rank ten coins and then sit still. The order rebuilt every eight minutes and nothing on a card moved between rebuilds. A coin could fall 20 percent and the card would still carry the old number for the next seven minutes.
Now it reads like a market. The order rebuilds about every three minutes. Market cap and pool size refresh every 22 seconds. Each card carries its own move since the build, green or red, grey when it is under half a percent so sideways drift never gets painted as a signal. A coin down 8 percent since the build says COOLING. One down 20 percent says FADED. The first FADED card seen in the wild read minus 27 percent.
Every card also carries an EXIT chip. It states what a $500 sell costs at the visible pool depth, recomputed from the live pool on every 22 second tick. Amber when the cost passes 3 percent, salmon past 5. That is the number that decides whether a paper gain is a real one, and it sits on the card instead of in the small print.
The name changed too. The board is “the ten coins with the most live buying”, not “coins we would buy”. That is more honest about what the ranking is: a measurement of live buying pressure, not a recommendation.
Two new penalties in the grade
A coin under 60 minutes old now loses 10 points on both grade and buy count. The first hour is where the numbers are least trustworthy, so the grade docks it instead of rewarding the spike.
The second penalty is for crowded names. A ticker with 10 or more launches in the latest sweep loses 3 points. One with 25 or more loses 5. When a name is being printed over and over, the newest copy does not get extra credit for the crowd around it.
Cards also show market cap and pool size side by side, with the pool flagged THIN when it is under 3 percent of the cap. A big number standing on a small pool is not a big number, as an earlier post on thin pools showed with a 151 million dollar cap standing on a 1 million dollar pool.
The record now counts the touch and the close
The bigger change is the record. A new picks record page lists every day’s calls and how each one finished: entry, 24 hour result, peak, and the clock.
Peaks used to stop being counted too early. Pricing stopped at 50 hours, so a high set on day two or day nine was never seen at all. Now every pick is repriced for 30 days, and the peak is taken from the coin’s own 1-minute chart rather than from polled samples. A high print that arrives days later still counts. The supply scale was corrected too, calibrated to the cent against a chart screenshot, so the published dollars match what the chart actually says.
Recomputed from the ledger this morning, the entry-to-peak record across 170 picks reads like this: median peak 1.53x, 51 picks (30 percent) at 2x or better, 31 at 3x, 12 at 5x, and 4 at 10x. The best is LAZARUS, listed at $13,862 and printed at $277,552 about 34 hours later.
Put the 24 hour record next to it and the two numbers tell the honest story together. Of 93 scored picks, the median 24 hour result is 0.24x and 13 of them, 14 percent, closed above entry. Most coins on this board touch a high and give it back. The peak record counts the touch. The scored record counts the close.
What this means for a reader
The board tells you what is being bought right now, how far it has moved since the ranking was built, and what a $500 exit costs before you care about exiting. The record page shows what happened to every call after it was made, including the ones that faded. And the checker still grades any pump.fun coin you paste into it on the same measures the board uses.
None of this tells you what the next coin will do. It tells you what happened to the last 170, and what the current board is doing as of the last tick.
Method
The changes listed here come from the site changelog, entries shipped October 30 and 31, the only changes made in the last seven days. The peak numbers were recomputed this morning from the stored entry-to-peak ledger of 170 board picks, and the 24 hour numbers from the 93 scored picks in the same ledger. Limits: peaks are the highest 1-minute print a coin reached, not prices anyone realized. A pick’s peak can still rise after publication, because the 30-day sweep keeps repricing. The sample is one board over its first days of operation. It measures the record, not the future.
