AI football scanners: tips and predictions85hidden gems found by AI
We gave the AI 8 years of football to remember, minute by minute: our own data, saved match after match since we were still LivePick. It hunted for patterns no human could ever spot in that much data and built scanners from the best of them. Every one passed our backtest on all matches since 1 January 2023 and now collects results live.
Sign up freeIn the free trial you can have up to 3 AI Scanners switched on, with a paired device
8 years, minute by minute
Our own match data from the last 8 years, minute by minute. We have saved every match since we were still LivePick. We loaded all of it into the memory of the AI and it cost us a lot of credits.
The AI found patterns
It went through all that data and found patterns that keep repeating, the kind nobody would ever spot by hand in so many numbers.
Scanners, backtested since 2023
From those patterns it built scanners and ran every one through our backtest on all matches since 1 January 2023. We start there on purpose: the COVID seasons before, with empty stadiums and no home crowd, were an anomaly, not patterns worth betting on. They went live on 12.11.2026.
Live ever since
Since that day the scanners have been running: they pick live and keep their own statistics. The automatic Winstrike settles every pick.
€
Stake per match
We show the profit of every AI Scanner in euros, as if you bet this amount on each selected match. Change it and all amounts update at once.
Filters
How many picks the scanner makes in an average week. You can choose several ranges.
Average odds of all settled picks. You can choose several ranges.
AI descriptionThis is the plainest version of my early BTTS idea, with no team-form context: in goalless openings it backs Both Teams to Score: Yes at long odds, without narrowing the field much. Two to three picks a day and 35 of 46 months in profit show how widely the pattern holds.
AI descriptionA form-based BTTS scanner for goalless openings: it needs both teams to come from matches where both sides usually score, then backs Both Teams to Score: Yes at mid-range odds. It picks around eleven times a week and returned a Yield of +17 % in the Backtest.
AI descriptionThis scanner bets Match Goals UNDER on at most one more goal when a level match with goals has gone quiet in the middle of the second half, with little danger near either goal. It was one of my steadiest UNDER scanners, in profit in 37 of 46 months.
AI descriptionNo goals yet, but plenty of attacking pressure: in exactly these lively openings this scanner backs Both Teams to Score: Yes at long odds. About two picks a day and 36 of 46 months in profit in the Backtest.
AI descriptionThis scanner waits for a goalless opening in which the goalkeepers have already been tested, then backs Both Teams to Score: Yes at long odds. Roughly one pick a day, with a Yield of +17 % in the Backtest.
AI descriptionA little into a goalless match that has already produced a shot on target, this scanner backs Both Teams to Score: Yes at long odds. Two to three picks a day and 37 of 46 months in profit.
AI descriptionEarly in goalless matches between teams with high-scoring recent games, this scanner backs Both Teams to Score: Yes over a broad range of odds. Two to three picks a day, more than a third of them winners.
AI descriptionThis scanner covers goalless league matches a few minutes after kick-off and backs Both Teams to Score: Yes at mid-to-long odds, about four times a day. With nearly 5 400 picks and 37 of 46 months in profit, it is one of the most robust league versions I have.
AI descriptionMost of my BTTS scanners act early, but this one waits until a goalless first half is well advanced and the price is long, in matches between goal-heavy teams. Three to four picks a day and 38 of 46 months in profit in the Backtest.
AI descriptionMidway through a goalless first half between two goal-heavy teams, this scanner backs Both Teams to Score: Yes once the price has drifted long. About two picks a day, with +16 % in the Backtest.
AI descriptionThis prematch scanner backs Both Teams to Score: Yes in league matches whenever the market prices that outcome as clearly unlikely. It finds about one pick a day, and in the Backtest since January 2023 it returned a Yield of about +16 % with 38 of 46 months in profit.
AI descriptionThis scanner backs Both Teams to Score: Yes a little into goalless matches where the prematch market saw goals at both ends as at least an even chance. Under two picks a day with a Backtest Yield of +16 % and one of the smallest drawdowns in its family.
AI descriptionA looser calm-tempo version for goalless first halves: it accepts a little more activity than its strict sibling and bets 1st Half Goals UNDER towards the interval. That nearly doubles the number of picks to about five a week, at a Yield of around +16 % in the Backtest.
AI descriptionThis scanner backs Both Teams to Score: Yes at longer odds a few minutes into any goalless match, without limiting itself to leagues. It picks almost twice a day, with a Backtest Yield of around +16 %.
AI descriptionShort odds, the very end of the match and no sending-off: when a narrow lead remains, this scanner bets Next Goal: No goal after an extra check on the live market. About one pick a day and 61 % of them won.
AI descriptionThis prematch scanner backs Both Teams to Score: Yes in men's football when the price is long, covering a slightly wider price range than my league-only version. It picks about nine times a week, and its 39 of 46 Backtest months in profit are one of the best monthly records among my prematch scanners.
AI descriptionThis prematch scanner backs Both Teams to Score: Yes when the prematch prices suggest the market expects the goals to come from one side only. It picks about nine games a week, and with 36 of 46 Backtest months in profit it has a steady monthly record.
AI descriptionThe Match Goals twin of the low-volatility scanner: very last minutes, no sending-off, a narrow lead and an extra check on the live market, bet as UNDER on the current score. About one pick a day, with a Backtest drawdown of around ten stakes.
AI descriptionThis scanner backs Both Teams to Score: Yes in goalless openings between teams whose matches usually see both sides score, over a wider range of odds than my other form-based versions. About three picks a day and 35 of 46 months in profit.
AI descriptionMy most consistent kick-off scanner backs Both Teams to Score: Yes in the first minutes of goalless matches at mid-range odds, five to six times a day. In the Backtest 44 of 46 months finished in profit, the best record in this family.
AI descriptionThis scanner follows narrow leads through the closing minutes and bets Next Goal: No goal once the trailing side has gone quiet, at mid-range odds. About one pick a day, with a Backtest Yield of about +14 %.
AI descriptionA busier Match Goals scanner for the closing minutes of matches without a sending-off: with a narrow lead and an extra check on the live market, it bets UNDER on the current score at moderate odds. One to two picks a day and 36 of 46 months in profit.
AI descriptionOne of my highest-volume BTTS scanners: during the opening phase of any goalless match it backs Both Teams to Score: Yes at mid-range odds, about seven times a day. It returned a Backtest Yield of around +13 %, with 42 of 46 months in profit.
AI descriptionA middle ground between the broad and the shot-less versions: it bets 1st Half Goals UNDER in goalless first halves that have produced only a handful of shots on target. About one pick a day and a Yield of around +13 % across more than 1 300 picks.
Profit+9 094 €
Yield
+13.2 %
Win rate
59.4 %
Picks
1.3K
Avg odds
1.91
No AI Scanner matches these filters
Try a wider odds range, another market or a different time in the match.
Why the conditions are hidden
These are our hidden gems. If the conditions were public, competitors would copy them within a day. You see what each scanner bets, how often and how it performs, and you get its picks like those of your own Scanners.
Why only in Ultra
We keep it a small club. If crowds bet the same matches, the odds would drop and the edge would be gone. That is why AI Scanners are only for Ultra members.
How we count the results
One bet on every selected match at the odds of the moment of the pick, at your stake per match (set above the list, 50 € unless you change it). Profit is in euros, yield is profit divided by stakes. Only settled picks count, the automatic Winstrike settles them.
Sign up and try Ultra free for 4 daysWith Ultra you switch any AI Scanner on with one click and its picks come to you with notifications. The conditions stay hidden. No card needed.