ANY was an even cleaner signal. The scanner scored it 55 on March 27th, and at that point borrow fee was already elevated at 98.6%. But here's what made it interesting: over the following week, borrow fee kept climbing. It hit 144%. That's not a typo. Shorts were paying 144% annualized to hold their positions. With 12.94% short interest, this was a textbook squeeze setup. The scanner caught the acceleration in borrow costs before the move.
BETR showed a different pattern. Borrow fee jumped from 2.8% to 13%, which doesn't sound dramatic until you realize that's an 83% spike in the cost to short. The scanner is designed to catch these accelerations, not just high absolute numbers but rapid changes. With 25.49% short interest, BETR had the ingredients.
SOUN was one of our best calls. The scanner flagged it with 34% short interest and a borrow fee staircase pattern, climbing from 4.7% to 12.8% in steady steps. But the share availability data was what really stood out. At one point, availability dropped to zero. Then a 100,000 share refill showed up and got absorbed in 34 minutes. When new short inventory gets eaten that fast, someone is aggressively shorting into strength or covering positions. Either way, it signals pressure.
The miss that made us better
Now for the important part. The time we got it wrong.
UGRO scored 95. Critical. The borrow fee was 199%, the highest we'd ever seen. By every metric in our original model, this should have been the squeeze of the year.
The stock dropped 37%.
We went back and studied what happened. The squeeze had already occurred. By the time borrow fees hit those extreme levels, the price had already run and was on the way back down. Shorts were winning, not losing. The high borrow fee was a lagging indicator of past tightness, not a predictor of future squeeze.
This miss taught us something crucial: squeeze conditions without momentum confirmation can be a trap. If the stock is already in a sustained downtrend, all those bullish squeeze indicators are just showing you a battleground where shorts already won.
So we added momentum scoring to the model. Now the scanner weighs price action alongside the short data. A stock with 200% borrow fee but a 37% decline gets flagged differently than one with the same borrow fee but building higher lows.
UGRO cost us nothing except ego, but it made the scanner meaningfully better. That's how you build something real.
What we're building next
The current version works. It catches setups, it surfaces opportunities, and most importantly it gives us the data early enough to do our own research before everyone else is talking about it.
But we're not done. The next version will incorporate catalyst proximity scoring, because squeeze setups near earnings or FDA announcements behave differently. We're also working on integrating SEC failure to deliver data more granularly, since FTD spikes often precede squeeze activity by days or weeks.
The goal isn't to build a system that tells you what to buy. The goal is to surface the names worth researching so you can make better decisions faster.
Why we're sharing this
Some people guard their edge jealously. We think differently about it. The market is massive. A scanner that helps identify squeeze setups doesn't stop working because other people know about it. If anything, understanding the mechanics makes everyone a better trader.
We also believe in building in public. When we miss, like UGRO, we want to explain why and show how we improved. When we hit, we want to share the receipts. This isn't about appearing smart. It's about being useful.
The scanner runs every morning. The data updates daily. If you want to see the current scores, the tickers being flagged, and the historical performance, the data is at visionboardfinance.com.
We built this because we wanted it to exist. Now it does.