I run a boring sheet. When top-quartile scores stop producing cleaner trades than bottom, the weights get rewritten from mistake codes — not from vibes or Twitter.
### Columns
Code: Select all
ticker
gap_%
rel_vol_pre
float_M
short_interest_%
catalyst_score_0_3
spread_cents_0935
trend_align_SPY_0_2
liquidity_ok_bool
total
trade_allowed_bool
Code: Select all
gap_% magnitude clipped: 0–2 pts
rel_vol_pre: 0–2
catalyst_score: 0–3
spread penalty: -2 if spread > your max
float sweet-spot: 0–2 (your range)
SI: 0–1 (only if your playbook uses it)
SPY align: 0–2
1. `trade_allowed_bool` is a hard gate (spread, halt risk, earnings binary, etc.).
2. Sort by `total` **after** gates.
3. Cap names you'll actually trade (3–7). Ranking 50 names you can't watch is self-entertainment. I've done the entertainment version. Didn't help fills.
4. Recompute after the first 15 minutes — premarket rank ≠ 10:00 rank.
### Review
Once a week: did top-quartile scores produce cleaner trades than bottom? If not, weights are cosplay — rewrite from the journal.
What weights matter most in *your* sheet right now?
Hard gates — what's your automatic kill (spread, float, earnings)?
How many names do you allow on the wall at the open?
Anyone re-rank at 09:45 and find the premarket list was mostly wrong?