Short version: predictions are based on past frequency. They are for entertainment only. No prediction system can beat a random draw.
Step 1 — collect draw history
We fetch the most recent results from each lottery operator via an API. For each operator we collect up to 100 of the latest draw results, covering all sub-games (e.g. Golden Treasure, Golden Sunrise for Golden Chance; Premier International, Premier Metro for Baba Ijebu, and so on).
Step 2 — count frequency
Every winning number from every draw is counted. If the number 34 appeared in 14 of the last 100 draws, it gets a frequency score of 14. The result is a ranked list of how often each number has been drawn.
Step 3 — rank hot & cold
- Hot balls — the numbers with the highest frequency in the analysis window.
- Cold balls — the numbers with the lowest frequency (or that haven't appeared recently at all).
You can view the full hot & cold breakdown on the Hot & cold page.
Step 4 — generate predictions & hot picks
- Predicted numbers — the 5 highest-frequency numbers from the current analysis window, presented as a suggested selection.
- Hot picks — the 3 single numbers with the strongest recent frequency, shown as the most prominent picks.
- Confidence level — a rough indicator of how much data was available: High (80+ draws analysed), Medium (40–79 draws), Low (under 40 draws).
What the predictions do not mean
Important — please read this
Lottery draws are independently random. Each draw is mechanically or computationally isolated from every previous draw. This means:
- A number that has appeared frequently is not more likely to appear again.
- A cold number is not "overdue" — it has the same probability as any other number.
- No algorithm, formula or system can increase your odds of winning a random lottery.
Our predictions are a form of entertainment — the same way people enjoy looking at star signs or lucky numbers. They represent what has happened, never what will happen. Please treat them accordingly.
Updates
Predictions and hot/cold statistics update automatically after new draw results are fetched (results are cached for 15 minutes and refreshed on the next page load). There is no manual curation — everything is computed directly from the latest available data.