Our AI Stock Picking Methodology: How Every Pick Is Made
Our AI stock picking methodology is built on one idea: no black boxes. This page explains how AI picks stocks on our platform, from the raw data to the final monthly list. If you want to know why a stock made the cut, the answer starts here.
The Short Version
Here is our AI stock picking methodology in five steps.
We collect the data. Financial reports and price history for 5,000 companies across 25 markets.
We score each stock. Five data pillars roll up into one score from 0 to 100.
We test before we launch. Every strategy is backtested on years of past data.
We pick and rebalance. The top scorers make the list. Once a month, the list is refreshed.
People check the work. Our team reviews every strategy before picks go out.
The 5 Data Pillars
Our stock scoring model looks at each company from five angles. No single number decides a pick.
Profitability
Profit margins, return on equity, cash from operations
Shows if a business actually makes money
Growth
Sales and earnings growth over 1, 3 and 5 years
Shows if the business is getting bigger
Financial health
Debt levels, interest cover, cash on hand
Shows if the company can survive a bad year
Valuation
Price compared with earnings, sales and cash flow
Helps avoid paying too much
Momentum
Price trend and earnings revisions over recent months
Shows if the market is starting to notice
In total, each stock is checked on 80+ data points.
From Data to a Single Score
Each pillar gets its own score. Those five scores are then combined into one overall score from 0 to 100.
The weights depend on the strategy. A value strategy cares more about price. A growth strategy cares more about sales growth.
| Strategy | Profitability | Growth | Health | Valuation | Momentum |
|---|---|---|---|---|---|
| Tech Leaders | 25% | 35% | 10% | 15% | 15% |
| Steady Value | 30% | 10% | 25% | 30% | 5% |
| Mid-Cap Movers | 15% | 35% | 10% | 15% | 25% |
| Dividend Builders | 30% | 15% | 30% | 10% | 15% |
| Global Picks | 25% | 20% | 20% | 15% | 20% |
Only stocks that also pass our basic filters can be picked. These include a minimum company size of $2 billion and enough daily trading so you can buy and sell easily.
How We Backtest Every Strategy
Backtesting a stock strategy means running it on past market data to see how it would have done. It is the main way we check if a strategy is worth launching.
Train on one period
The model learns from data between 2005 and 2019.
Test on a different period
We then run it on years it has never seen. This is called out-of-sample testing, and it helps stop the model from simply memorising the past.
Include real costs
Results include estimated trading costs of 0.1% per trade.
Check the bad years
We look closely at how the strategy held up in market drops, such as the 2020 crash and the 2022 bear market.
Pass rules: a strategy only launches if it beats its benchmark over the full test period, recovers from drops within 6 months, and does not rely on just one or two lucky stocks.
Backtests are useful, but they are not a promise. Real results can be different, so we also publish the live results in our full track record.
Monthly Rebalancing
Markets change, so the list changes too. On the first trading day of each month, every strategy is updated.
Stays
stocks that still score near the top keep their place.
Drops out
stocks whose score falls below 60, or that fail a filter, are removed.
Comes in
new top scorers take the open spots.
Each stock in a strategy has an equal weight. Every change is listed with its date, so you can follow along step by step.
The Fair Value Check
A great company can still be a bad buy if the price is too high. Before a stock is added, the AI estimates what the company is worth based on its earnings and cash flow.
If the price is far above that estimate, the stock’s score is cut. This keeps strategies from chasing stocks that have already run too far.
Risk Controls
Good AI stock picking is not only about finding winners. It is also about limiting damage when things go wrong.
| Control | Rule |
|---|---|
| Sector limit | No more than 25% of a strategy in one sector |
| Country limit (Global Picks) | No more than 20% in one country |
| Number of stocks | 15 to 25 stocks per strategy |
| Minimum size | Company value of at least $2 billion |
| Trading volume | Average daily trading of at least $5 million |
| Risk score | Every pick shows a volatility-based risk score from 1 (low) to 5 (high) |
Human Oversight
The AI does the number crunching. People make sure it is working as it should.
Our team of analysts and data scientists reviews each strategy every month before picks go out.
We check the data for errors, such as wrong prices or missing reports.
We review and retrain the models every 6 months.
Want to know who is behind the models? Meet the team.
What Our AI Does Not Do
No AI stock picking methodology can remove market risk. Being clear about limits is part of being trustworthy.
- It does not predict the future. It ranks stocks based on data and past patterns.
- It does not know your personal goals, income or risk level.
- It does not place trades or manage your money.
- It cannot react to surprise events, like sudden news, between monthly updates.
Please read our Risk Disclosure before acting on any pick.
Our Data Sources
Data is updated daily, and scores are recalculated before every monthly rebalance.
Frequently Asked Questions
How does ProPicks AI choose stocks?
Each stock is scored on five data pillars: profitability, growth, financial health, valuation and momentum. The highest overall scores make each strategy’s list.
How often is the stock scoring model updated?
Scores are recalculated before every monthly rebalance. The model itself is reviewed and retrained every 6 months.
Are backtest results the same as real results?
No. Backtests show how a strategy would have done in the past. Live results can differ, which is why we publish our full track record.
Can I see why a stock was picked?
Yes. Every pick comes with a short note explaining its strongest scores and its main risks.
See the Method in Action
Now you know how AI picks stocks on our platform. See this month’s list and the reasons behind every pick.
Get This Month's Picks
Join 12,000+ investors. Create your free account in under a minute.
Investing involves risk. Picks are for information only and are not financial advice.