Home
The Lab
The Graded Tape
Archive
About
Home
The Lab
The Graded Tape
Archive
About
Subscribe
Sign in
Latest
Top
Discussions
I built a model to say when to trust my trading model
A small model built to say when the big one could be trusted, and a FAIL I published.
Nov 5
•
Verbatim
1
Pre-registration for traders
The one-page design doc that decides a verdict before the results exist, field by field, free to copy.
Nov 3
•
Verbatim
1
October 2026
Weekly note 2: four questions to ask any backtest
What I check before I trust a backtest, mine included. Paid subscriptions now start with a 14-day free trial.
Oct 30
•
Verbatim
1
Why I froze three months of data I'm not allowed to look at
The winner's curse, a holdout used exactly once, and the Sunday evening the sealed months were finally read.
Oct 29
•
Verbatim
1
Walk-forward testing, explained for people who lost money after a good backtest
How a model gets graded only on days it has never seen, what that costs, and the one failure it still cannot catch.
Oct 27
•
Verbatim
1
Weekly note 1: a positive week on the Tape, and two ideas that only won in hindsight
Three sessions up on the Tape at paper sizes before costs, and two older ideas that beat the index only when chosen with hindsight.
Oct 23
•
Verbatim
1
The research log, September 2026
The allocator failed its full-history test, Lookout's paper losses traced to the fills, and a faster data feed would not have helped. This first issue…
Oct 22
•
Verbatim
1
Arc two: The Machinery
Five checks I built to keep from fooling myself that my neural network is good at trading.
Oct 20
•
Verbatim
1
2
My model is 6x more precise on low-priced stocks
Same model, same test, six times the precision. Then I checked the base rates. The arc one working papers are below, free.
Oct 19
•
Verbatim
1
A logistic regression scores 0.35. My deep model scores 0.36. Is that worth it?
A plain logistic regression got 96 percent of the way to my deep learning score. What the trees scored told me where the edge comes from.
Oct 19
•
Verbatim
My only clean win cost fifteen dollars
Nine ideas and $423 produced one clean win: a fifteen-dollar average of four tied models and a plain logistic regression.
Oct 19
•
Verbatim
1
Four models tied. That was the right answer.
Written in July 2026.
Oct 19
•
Verbatim
This site requires JavaScript to run correctly. Please
turn on JavaScript
or unblock scripts