You sized that position by how sure you felt, not by your account.

Conviction is the worst input to a size decision

Certainty measures how completely a story has been told to you. It rises with minutes spent on the chart, with how many people are saying the same thing, with how clean the narrative is. None of that is information about the next candle. By the time the evidence is strong enough to make you sure, it is public.

The deeper problem is that conviction carries no loss term. It answers how likely you are to be right. Size is not a question about being right, it is a question about what happens when you are not, and in memecoins that case is not minus 15 percent. Conviction is also uncalibrated: you have never scored it, so you do not know whether the trades you felt surest about beat the ones you took with a shrug. It is still the only variable you let move your size.

What an oversized position does while it is open

At your normal size you watch the chart. At three times that you watch the P&L. Those are different instruments, and the second one moves with your balance instead of with the token.

A 30 percent drawdown you would have sat through without noticing becomes a number with a meaning attached, so you close it. Green works the same way in reverse: 1.4x on a position three times your normal size pays the same as 2.2x would have at your normal size, so you take it and watch the thing run without you. Both get filed later as reading the chart. Neither had anything to do with the chart.

Run far enough against you and the loss becomes too large to accept, so it stops being a trade and turns into something that has to come back. The position too big to cut is the one you hold to zero.

The arithmetic of getting back to flat

A loss and the gain that cancels it are not the same size, and the gap widens the deeper the hole gets.

What bites is not the hole, it is the next decision. The account that now has to produce 233 percent is smaller than the one that lost it, so you size up to make the recovery arrive sooner, which is this same mistake with a deadline on it.

Your results are not the average of your trades. They are the average weighted by size, so your P&L is mostly a record of the trades you were most emotional about, not the ones you read best. One 2 SOL loss cancels ten 0.2 SOL wins.

• Down 20 percent: you need 25 percent to get back to flat.
• Down 40 percent: you need 67 percent.
• Down 50 percent: you need 100 percent.
• Down 70 percent: you need 233 percent.

How to tell it apart from a legitimate big position

A large position is not automatically this mistake. What makes it the mistake is that the number got chosen after you saw the chart. Three questions separate them.

• Did the number exist before the token did? A size that comes from a rule you set earlier is a decision. A size that changed because this one felt different is not.
• Can you finish this sentence out loud before you send it: if this goes to zero in forty minutes I am down 1.8 SOL, and that is fine.
• Look at the SOL amounts on your last twenty entries. If they are scattered, you do not have a size. You have a mood.

Fix the number before the chart

Set the size before you open the chart, and set it in SOL, not percent. Percent is abstract, SOL is what you feel. The order matters more than the number: any fixed size chosen in advance beats a smart size chosen at entry, because entry is when your judgment is worst.

Sizing is one of the few mistakes visible in the ledger alone, because the SOL amount sits in the swap. Dossier will not tell you what to size. It shows what you did size, what the chart did around it, and how far that entry sat from your own baseline.