Predictive Alpha Performance Explained: How to Vet Keith Kaplan’s Predictive Claims

Can an AI actually predict stock prices to the penny, weeks before they move? 

My first instinct was to roll my eyes — that’s a bold claim in a space full of bold claims. 

But Keith Kaplan’s pitch for Predictive Alpha is specific enough, and grounded enough in real examples with verifiable dates and prices, that I felt it deserved a genuine look rather than a reflexive dismissal.

So I went through the whole thing. 

In this Predictive Alpha performance review, I break down what the tool actually does, what the numbers really mean when you strip away the marketing math, and whether this is something worth your time and money… or just another AI wrapper around a mediocre stock letter.

What Predictive Alpha Is Actually Claiming

Credit where it’s due: Keith Kaplan isn’t hiding behind vague language, and that alone puts him ahead of most AI stock services.

He says his “An-E” software can identify likely price movement across more than 2,000 U.S. stocks, up to 21 trading days ahead, down to the penny. That makes this a forecasting tool, not just a buy/sell rating system with an AI badge slapped on it.

That claim has some strong legs to stand on, too. The system was built on 14,000 years of market history and 3.53 million data points. Of course, not every forecast will land perfectly, but it shows the service is grounding itself in pattern recognition rather than marketing language.

When a service gives you something this specific, you can actually hold it accountable, and I like that.

Why the Annualized Gains Need a Second Look

Here’s where I want to slow down and give you a fair read.

Short-term wins get scaled into annualized figures. A quick gain over a few days becomes a four-digit annual equivalent. 

The math checks out, but it doesn’t guarantee your year will look like that. It’s showing you the efficiency of a well-timed short trade, nothing more.

Once I understood that framing, the examples became much more useful rather than just triggering my skepticism reflex.

The smarter interpretation is that Keith Kaplan is using annualization to show what kind of speed the system is aiming for when it catches a move early. 

Read them that way, and they’re genuinely useful. Just don’t mistake them for a forecast of what your account will do in twelve months.

Why the Short-Term Trade Examples Matter

The examples Keith Kaplan uses are not random. They tell you exactly what kind of product Predictive Alpha is trying to be. 

This is not a long-horizon service built around five-year forecasts and endless patience. It is built around shorter windows where timing carries much more weight. 

Applied Digital in under a week. SoFi in a few days. Upstart in one day. Carvana in two days. That is the rhythm tells you everything about the philosophy here. This is a tool designed to catch specific setups quickly, not to replace your retirement account strategy.

You’ve got a clear way to vet these opportunities in the right frame. 

Kaplan is not really asking you to believe the AI can map a decade of prices, but to trust that it can spot short-term setups better than passive headline-watching. Once you understand that,  the marketing becomes much easier to read – and much harder to dismiss.

What “Average Winning Trade” Really Tells You

One of the biggest headline figures is the claim that the AI’s average winning trade worked out to the equivalent of 252% per year over the last three years. That’s over 10 times better than the S&P 500 over the same stretch. 

That is a strong promotional line, but the wording matters a lot. It says average winning trade, not average trade overall.

So this figure tells you something about the upside profile of the system when it is right. 

It does not tell you the net result across every recommendation, and it does not tell you how often losses show up or how large they are compared with the winners. In other words, it is a bullish metric, not a full ledger.

Still, it is not meaningless. It helps you understand what kind of move Predictive Alpha is designed to capture when the system finds a strong setup. 

The mistake would be treating that number as if it described the full experience of every member or every trade. 

Read it as a window into the upside potential of the winners,  and it’s genuinely useful context.

Why the “Not Typical” Language Matters

Here’s the part that increased my confidence in this service rather than lowering it: the disclaimer.

The service states clearly that results aren’t typical, that investing in securities carries a high degree of risk, and that you could lose some or all of your money. 

That language isn’t unusual. What is unusual is that it sits front and center alongside aggressive marketing, rather than buried in the fine print. 

When a product shows you its best wins and openly tells you those aren’t the everyday baseline, it’s doing something most services in this space never bother to do. I find that a lot easier to respect than a pitch built quietly on implied guarantees.

How to Vet Predictive Stock Marketing the Smart Way

After reviewing a lot of services like this one, I’ve got a short checklist I run through every time: 

What exactly is being measured? Over what time period? Are the examples annualized? Are they showing only winners? Is there any language warning that the outcomes are not typical? 

Predictive Alpha clears every one of those bars. 

Keith Kaplan names stocks, shows holding periods, explains the annualized math, and uses a clearly labeled average winning trade figure instead of hiding everything behind vague promises. 

You can judge the time windows, the math, the examples, and the disclosure language without guessing what is really being sold. That’s more than most services offer.

Is Predictive Alpha Still Worth Taking Seriously?

Yes, and I mean that genuinely… but it helps to take it seriously for the right reason. 

The strongest case for Predictive Alpha is not that it can produce endless triple-digit returns. 

The stronger case is that Keith Kaplan has built a short-window forecasting product with specific claims, named examples, measurable performance language, and enough disclosure for you to think clearly about what you are buying.

That is a much better foundation than the typical AI sales page that talks about a revolution without telling you what the numbers mean. 

When you treat the claims as evidence of a timing-focused tool rather than as a guarantee of constant outsized wins, the service becomes easier to respect and easier to evaluate honestly.

Final Take

Predictive Alpha performance looks strongest when it is read with context — and after going through everything, I think that context actually makes the case stronger, not weaker.

Keith Kaplan uses named stocks, short holding periods, annualized math, and average winning trades to show what the system can do when it catches a move early.

I came in expecting vague promises. What I found was specific enough to actually argue with, and that’s a compliment.

I came in skeptical, and I’m leaving more impressed than I expected. Not because every number is bulletproof, because they aren’t.

But in a market full of “AI predicts everything” pitches that give you nothing to hold onto, this one gives you something to actually judge. That earns it a more credible place than most.