False Positives & False Negatives

Even the best AI screener won’t be 100% accurate.

Sometimes it will flag a stock as an “opportunity” when it’s not.

Other times, it will miss a stock that could have been a winner.

These errors are called false positives and false negatives. Understanding them is key to managing risk when using AI.

What is a False Positive?

A false positive happens when the screener includes a stock that looks promising but turns out to be weak.

Example:

AI shows you a mid-cap IT stock because it meets all your filters.

But then you check the news and see the company facing a big regulatory issue.

The screener gave a positive signal — but it was “false.”

What is a False Negative?

A false negative is the opposite. It happens when the screener misses a good stock that actually fits your strategy.

Example:

A bank stock is trading above its 200-DMA and has strong fundamentals.

But due to a data gap or a wrongly set prompt, the screener leaves it out.

You lose out on spotting a genuine opportunity.

Why Do These Errors Happen?

• Incomplete data: Not all sources are updated or accurate.
  
• Overfitting: Too many filters can reject good stocks.
  
• Loose prompts: Vague wording can make AI include irrelevant names.
  
• Market surprises: A sudden event can break any screener’s logic.
  

How to Handle False Positives

• Always cross-check shortlisted stocks with news and filings.
  
• Add extra filters: For example, remove companies with high promoter pledges.
  
• Don’t treat every stock the screener shows as a buy signal. It’s just a shortlist.
  

How to Handle False Negatives

• Avoid making prompts too strict.
  Instead of “ROE above 18%,” try “ROE above 12%.”
  
• Run variations of the same prompt.
  One for P/E, another for debt, another for profit growth.
  
• Remember, missing one stock is better than chasing the wrong one.
  

Final Takeaway

AI screeners are not perfect.

They can show false positives — stocks that look good but fail.

And they can create false negatives — missing out on real opportunities.

The way to manage risk is simple:

Cross-check results, keep prompts balanced, and never rely on one output alone.

A screener is your filter, not your fortune teller.