The Statistics tab on Whale Alpha opens with six tiles, and two of them are the ones everybody screenshots. Best Performing Whale, reading $1.00B. Worst Performing Whale, reading $0. They are the most eye catching cells on the page and they are the two you should learn to skip.
Not because they are wrong, though there is something wrong with them, but because of what they are. They are the maximum and the minimum of a sample of 26,687 tracked wallets. A maximum is by definition the least typical observation in a dataset. It tells you about the edge of the sample and nothing whatsoever about what a tracked whale actually does, which is the thing you were presumably trying to find out.
Two cells out of 26,687
Start with the scale. The Total Whales Tracked tile on the same row reads 26,687. The Best and Worst tiles are two rows out of that. If you picked two wallets at random from 26,687 you would learn very little; picking the two extremes teaches you less than that, because extremes are where measurement problems collect.
This is a general property rather than a criticism of one dashboard. Every "best" cell on every dashboard is a maximum over a large set, and every maximum attracts whatever is broken, because a broken row is much more likely to produce an extreme value than a typical one. Bugs sort to the top of a maximum. Once you have that heuristic you start to see it everywhere.
A round number that three wallets share
Here it is not even a general worry, it is checkable. The Best tile reads exactly $1.00B, which is a suspiciously round figure for a measurement.

Switch to the Finder tab and the reason becomes clear. Three separate wallets, on Ethereum, Base and BSC, all displayed a profit of exactly 1.00 B USD and a win rate of exactly 95 percent. Three unrelated wallets on three different chains do not land on the same figure to the penny. That is a ceiling in the display, which means Best Performing Whale is not identifying a best performer at all, it is reporting a tie at a cap.
The Worst tile has the mirror problem. It reads $0. Zero is not the worst possible outcome for a trader, it is what an untracked, unresolved or never active wallet shows. A wallet that lost heavily would show a negative number. So neither tile is measuring what its label promises, and the two of them between them are the entire reason people quote this page.
When two tiles in the same row disagree
Two more cells on that row are worth looking at together, because they demonstrate the check you should be running on any dashboard.
Total Volume reads $2889244.43B, all time. Average Bet Size reads $0. Those two cannot both be describing 26,687 wallets. If there were any volume at all across that many wallets, the average bet size is not zero, and if the average bet size were genuinely zero there is no volume to total.
The useful move when two cells in one panel contradict each other is not to decide which one you believe. It is to stop using derived cells from that panel entirely and drop down a level to the raw distributions, which is exactly what sits underneath.
The histograms underneath are the real answer
Two charts, and they are worth more than all six tiles combined.
The Win Rate Distribution splits wallets into five buckets and at capture read $2K for the 0 to 20 percent bucket, 935 for 20 to 40, $13K for 40 to 60, $3K for 60 to 80 and $3K for 80 to 100. The labels are inconsistent, four carrying dollar signs and one not, and the bars sum to somewhere around 22 thousand against 26,687 tracked wallets. Read together, that says these are wallet counts with a currency formatter applied to them rather than dollar amounts, since the one bar small enough to escape the thousands rounding prints as a bare 935.
Read the shape, not the labels. The mass is in the 40 to 60 percent bucket, which holds something over half of all tracked wallets. Roughly 3,000 wallets, a bit over one in ten, sit in the 80 to 100 bucket. That shape agrees with the Average Win Rate tile at 44.3 percent, and that agreement is the point: when a tile and the chart underneath it reconcile, you can use the tile. When they do not, as with the volume and bet size cells above, you cannot.
So the honest headline number for this page is 44.3 percent, and the honest reading is that a typical tracked whale wins somewhere between four and six trades in ten. That is a completely different mental picture from the one the billion dollar tile gives you, and it is the one that should inform how you treat any individual wallet you find.
Picking your size threshold off the bet size bars
The second chart is the one that changes what you actually do. Bet Size Distribution reads 451 in the $100K to $250K band, 193 in $250K to $500K, 94 in $500K to $1M and 25 above $1M. Those are clean counts with no formatting oddity, summing to 763.
Notice the halving. Each step up in size roughly halves the population, which is the ordinary shape for this kind of data and is useful because it lets you choose a filter threshold on evidence rather than on which round number sounds serious. The feed's size chips run $100K, $250K, $500K and $1M. Setting $1M leaves you 25 things to look at, which is a manageable evening and a very thin sample. Setting $100K leaves 763, which is more than you will read. Setting $250K leaves 312, being the top three bands added together, which is the setting I would start from.
That is the practical output of this whole page. Not a wallet to follow, a threshold to filter at, chosen from a histogram rather than from a habit.
What to carry away from a page like this
Stop screenshotting the Best tile. It is a tie at a display ceiling and it is not identifying anybody. Write down the two distribution shapes instead, and use them as priors whenever you look at an individual wallet.
Concretely, that means the next time you find a wallet advertising a 95 percent win rate, you already know it sits in a bucket holding roughly 3,000 others, that the population it belongs to averages 44.3 percent, and that the first question is how many trades the 95 percent was measured over. That single reframe, from this wallet is exceptional to this wallet is in the top bucket along with three thousand others, is most of the value on this tab.
And keep the general rule. On any dashboard, check the tile against the chart underneath it, use the ones that reconcile, and treat a maximum as decoration. The cells that survive that test are few, and they are the ones worth building a habit on.