Squintz monitors 2,007 tickers every hour. Most of them do nothing worth your time. A few every day do something interesting. And maybe one or two a day clear all eight High-Conviction gates and earn a text.
The full gate list
Every filter below comes straight from config/risk-parameters.json. If you're getting the text, the setup cleared every one of them:
- GRIPS total ≥ 75 — HIGH tier or better across all five dimensions.
- G (Growth) ≥ 14 — real velocity, not a slow drift.
- I (Intensity) ≥ 10 — enough volume that the move isn't a couple retail orders.
- S (Spread Quality) ≥ 10 — tight enough book that execution doesn't eat the edge.
- Change ≥ 8% — meaningful move size inside the scoring window.
- Velocity ≥ 0.20%/min — the move has to be active, not decaying.
- Direction = up — Squintz is long-only. Shorts are a different beast entirely.
- Window ≤ 45 minutes — the move happened recently. Stale rallies don't fire.
- Time of day: 9:30–11:30 AM ET — the institutional entry window.
- Entry price ≤ $300 — affordability for normal account sizes.
Miss any one of these by even a hair and the alert doesn't fire. That's the deal.
Why each gate exists
GRIPS 75 and the sub-dimension floors
A 75 total can theoretically pass with one strong dimension and four mediocre ones. The sub-floors (G ≥ 14, I ≥ 10, S ≥ 10) prevent that. You can't sneak through on ML probability alone.
The 9:30–11:30 window
Historical backtests across ~14,000 labeled signals showed the same pattern every time: morning setups work. Afternoon setups are a coin flip. The 11:30 cutoff isn't arbitrary — it's where the edge statistically collapsed.
The $300 ceiling
Two reasons. First, affordability — a $5/month user can actually buy shares. Second, PDT exposure. Higher-priced stocks produce bigger P&L swings per share, which pushes account equity closer to the $25k minimum rule if a trade goes wrong.
The 8% move floor
Below 8%, the reward-to-friction ratio gets ugly fast. Take a 3% winner, subtract spread both ways, subtract slippage, and you're looking at a 2% net for the same execution risk as an 8% setup. Squintz demands a move worth showing up for.
Why HC alerts are deliberately rare
On an average day, Squintz scans ~2,000 stocks every hour. Of the top 30 movers per hour, maybe 1–3% end up passing all eight gates. That's the design. The value of a text isn't frequency — it's how much you trust it when your phone buzzes.
If the gates fired 30 times a day, you'd ignore them by Wednesday. Fired 1–3 times a day, you'll stop whatever you're doing and look.
The contract
When you see the orange HC pill on a signal card, Squintz is saying: every filter passed, the math checked out, and the odds are on your side for the next 45 minutes. That's not a guarantee. It's a shot worth taking.
You'll still get losers. The gates tilt the distribution — they don't rewrite probability. What they do is guarantee every trade you take was worth taking.