Yo Pit Crazies,

My new scanner just ran through the entire market and spit out 17 stocks that met every single criterion. 

I bought one that wasn’t on the list.

Let me explain why the human element still matters, even when the algorithm is doing exactly what you built it to do.

The scanner methodology is simple. We start with stocks making 52-week lows. 

Once they bounce 10 percent off the bottom, they move to the next column. Then we check for earnings improvement. 

Finally, we look for a 40 percent spike in open interest on the options, which tells us institutional money is positioning for a move.

When the Machine Says One Thing

Last week, the scanner flagged ONEOK (OKE) as a buy. All four criteria lined up perfectly. The open interest on January 2027 calls jumped 40 percent in a single week. Earnings improved. The 10 percent bounce happened right on schedule.

Here’s where it got interesting. The scanner wanted me to buy the 97.5 calls because that’s where the math showed the best potential return. But those markets were thin. Goofy bid-ask spreads. Hard to get in, harder to get out.

I bought the 82.5s instead because I could actually get filled at a reasonable price. The scanner doesn’t know about liquidity. It just runs the numbers.

When to give a second look

Then there’s Fiverr International (FVRR). This one wasn’t even close to passing the scanner criteria. But I bought it anyway.

Here’s the thing. Fiverr is an Israeli technology company sitting at a two-year low while institutions are quietly loading up on calls. The open interest grew about 50 percent in a short window. That’s the fingerprint I’m looking for.

But the scanner missed it because the stock made an earlier 52-week low months ago and then made a second one. The algorithm saw the earlier bounce and moved on. It wasn’t watching for the second setup.

A human was.  Me.

The Tool Is Not the Trade

StockBrokers.com just published research showing that AI-powered scanners like Trade Ideas can identify setups faster than any human could manually search. But they also found that scanner alerts require confirmation with your own analysis and that sometimes even a perfect-looking alert should be disregarded if it contradicts your broader market view.

That’s exactly right. The scanner does the searching. You do the thinking.

I’ve been trading options for 35 years. The best tools I’ve ever used didn’t replace judgment, they freed up time so I could apply more of it. When the scanner shows me 17 names, I’m not blindly buying all 17. I’m looking at each one and asking: Does this make sense right now? What’s the liquidity? What risks is the algorithm not seeing?

The Takeaway

Build the scanner. Trust the scanner. But don’t follow it off a cliff.

Every tool has blind spots. The best scanners can filter thousands of stocks in seconds, but they can’t assess geopolitical risk, liquidity, or whether the options market is thin enough to trap you. That’s your job.

The scanner narrowed my universe from thousands of stocks to 17 worth watching. That’s valuable. But deciding which ones to actually buy, and which to ignore even when the math says yes, that’s where the edge lives.

Does that make sense?

Andrew