Notes to myself, kept in public.

Every time a read goes wrong in a way that annoys me, I trace it back and find the same missing step: I never said out loud whether I was in a range or a breakout. I went straight to the setup.

The setup is downstream. A High 1 in a strong breakout and a High 1 at the top of a range are the same three bars and opposite trades. If I cannot name the market type in one word, I have not started reading yet.

I have a habit of entering on the first leg because it feels decisive, and the first leg is exactly the part that feels decisive to everyone.

Waiting for the second leg costs me the best entries in the strongest trends and saves me from most of the ranges. On balance I think that trade is worth making, but I want to be honest that it is a trade, not a free improvement.

A count is checkable and a prediction is not. Six bars is six bars whatever I think about the economy.

The days I do this well, my notes are almost boring: bar four, bar five, pause, bar six. The days I do it badly, the notes are full of reasons.

For a while every chart I saved had eight lines on it, and I could make any of them the important one afterwards. That is not analysis, it is a way of always having been right.

Now I draw the range edges and, if there is a trend, one channel line. If a level matters it will matter without me decorating it.

The clearest signal I get all day is the feeling of not wanting to look at a bar. It arrives before the reasoning does.

What I do with it is the whole of digit zero. Usually the bar is telling me something specific: the follow-through did not come, or the correction has gone one leg too far. I have never once found that the feeling was noise and the position was fine.

A breakout bar tells me somebody was willing. The bar after it tells me whether anybody else was.

Big body, small tails, closing beyond — the move has participants. An inside bar or a doji, and I am looking at a probe that will probably come back. I get more out of that one bar than out of anything I do before the breakout.

I have never identified a wedge on the second push with any confidence. On the third push it is obvious, and by then a good part of the move is gone.

So I have stopped trying to call them early and started treating the third push as the event. Late and correct beats early and hopeful, at least in my notes.

The textbook version is: if I had to be in the market in one direction right now, which would it be. In practice I use it as a tripwire.

When three bars in a row argue against the position I am holding, the answer to that question has changed, and my position has not. That gap is where most of my bad days live.

The rule I am strictest about with myself: the read goes in the notebook before the next bar prints. Afterwards it is not a read, it is a memory with the edges filed off.

It is uncomfortable how different the two are. Roughly a third of what I would have sworn I saw in real time turns out to be something I assembled later.

Putting the wrong reads up is the only reason any of the right ones mean anything. A page with only good calls on it is a marketing document.

It is also the part I have to force. The entries I most want to quietly not publish are reliably the most useful ones to re-read three months later.

I spent two years watching charts for six hours a day and learned less than I did in the first three months of writing one paragraph about each session.

Watching produces a feeling of familiarity. Writing produces something I can check. Only one of those has ever corrected me.