In Part 1, We Analysed 8,439 Bets. Here’s What We Found., we showed you the headline stat: +398 points profit from 8,439 AI bet recommendations. But we also said the real story isn’t just that Gecko Edge is profitable — it’s also where and how the profit is delivered.
We’re focusing on the 1,747 unique AI pre-match bet recommendations in this part. These are deduplicated — one entry per fixture per bet type — so we’re looking at pure model performance, not inflated by multiple users running the same analysis.
The Bet Types That Make Money
Here’s how every major pre-match bet type performed:
The winners:
- Under 2.5 Goals — +50.07% ROI | 73.8% strike rate | 172 bets
- First Half Over 0.5 Goals — +22.33% ROI | 69.6% strike rate | 194 bets
- Second Half Over 1.5 Goals — +13.26% ROI | 54.5% strike rate | 101 bets
- Asian Handicap — +7.90% ROI | 53.8% strike rate | 380 bets
- BTTS Yes — +4.05% ROI | 56.5% strike rate | 262 bets
- Match Result Home — +2.22% ROI | 47.1% strike rate | 87 bets
The under-performers:
- Over 2.5 Goals — -5.08% ROI | 48.1% strike rate | 316 bets
- Match Result Draw — -7.55% ROI | 22.7% strike rate | 22 bets
- BTTS No — -9.39% ROI | 46.4% strike rate | 125 bets
- Over 3.5 Goals — -75.53% ROI | 13.3% strike rate | 15 bets
- Match Result Away — -54.66% ROI | 13.7% strike rate | 73 bets
This is not to suggest that ‘under-performers’ should be ignored entirely. There can be any number of reasons for a bet to under perform over a sample, such as volume of bets, variance and freshness of available data.
Under 2.5 Goals: The AI Betting Superpower?
+50.07% ROI. 73.8% strike rate. 172 bets.
When Gecko Edge says a match is going under 2.5 goals, it’s been right almost three-quarters of the time in our sample. At average odds of 2.10, that combination of strike rate and price generates serious returns.
Why does the model excel here? Under 2.5 markets tend to be undervalued when two defensively solid sides meet or when league-wide scoring rates are lower than the market assumes. The model picks up on these structural factors — things like low combined xG, disciplined defensive records, and leagues where 1-0 and 1-1 scorelines occur more frequently than the odds suggest.
In our sample, this has been the single most profitable pre-match bet type in the dataset by a significant margin.
First Half Over 0.5 Goals: The Consistency Machine
+22.33% ROI. 69.6% strike rate. 194 bets.
Nearly 70% of these land. The odds are generally lower, so the returns per bet aren’t as high as Under 2.5 — but the consistency is remarkable. You’re collecting small, regular wins at a rate that compounds quickly.
The model identifies fixtures where early goals are structurally likely — aggressive pressing sides, leagues with high first-half scoring rates, or matchups where both teams tend to start fast. The market underprices this more often than you’d expect.
If you want a bet type that delivers steady, reliable returns without the variance swings, this could be it.
Build in a drip-fed or delayed staking plan could extract even greater returns.
(Side note: The Gecko Edge team tends to focus on the First Half Goal market, even when the +EV value is higher on FHG 1.5).
Asian Handicap: The Volume Play
+7.90% ROI. 53.8% strike rate. 380 bets.
This has been the highest-volume profitable bet type in the dataset. The ROI isn’t as eye-catching as Under 2.5 or First Half 0.5, but the volume matters. 380 bets generating a consistent +7.90% return gives you a large, repeatable edge with lower variance than smaller sample sizes.
It makes sense that the AI modelling handles handicap pricing even more effectively on the Asian Handicaps as it provides more flexibility and efficiencies compared to the traditional match odds.
You can spread risk when variance is high and leverage returns when confidence is strong by applying the right Asian Handicap per match.
Odds Band Analysis
The key finding is that the Gecko Edge Betting Assistant is profitable at every odds band level, excluding bets above 4.00, which are higher variance and with a much lower sample size of just 56 bets.
- 1.01-1.50: +4.85% ROI | 75.7% strike rate | 140 bets
- 1.50-2.00: +2.56% ROI | 58.1% strike rate | 800 bets
- 2.00-2.50: +11.41% ROI | 51.6% strike rate | 517 bets
- 2.50-3.00: +6.90% ROI | 40.4% strike rate | 138 bets
- 3.00-4.00: +29.70% ROI | 39.3% strike rate | 90 bets
- 4.00+: -21.02% ROI | 16.1% strike rate | 56 bets
The 2.00-2.50 range is the sweet spot. It generates the highest total profit (+58.99 pts), strong ROI (+11.41%), and a strike rate above 50%. This is where the model’s edge and the market’s pricing inefficiency overlap most consistently.
Below 2.00, you’re winning often but the prices are too short to generate definitive returns. However, do not discount these out of hand – you can feed these into your football trading plan and look for higher +EV Opportunities when betting InPlay.
The 3.00-4.00 range looks impressive (+29.70%), but from only 90 bets — treat that as a signal to watch or research further on a match by match basis, rather than a confirmed edge.
The Sanity Score: Your Confidence Filter
Gecko Edge produces a Sanity Score — a confidence rating — for most recommendations.
Here’s how the three segments perform:
- Sanity Score 6 and above: +9.98% ROI | 55.1% strike rate | 999 bets
- No Sanity Score available: +3.91% ROI | 54.4% strike rate | 537 bets
- Sanity Score below 6: -7.58% ROI | 45.0% strike rate | 211 bets
The Sanity Score is a quick indicator when conducting your analysis to understand how confidence the Gecko Edge AI Betting Assistant is in its analysis.
Gecko Edge delivers probabilities, not tips. So it stands to reason that the more accurate the data, the freshness the odds, and clear the analysis, the stronger the probability recommendation.
Anything rated 6 or above has shown to be the strongest segment — nearly +10% ROI.
There are occasions where no Sanity Score was able to be produced at the time of the analysis. You should not ignore these, it does not indicate an issue in the analysis and they have proven to still perform solidly at +3.91%.
The model edge exists even when the confidence rating couldn’t be generated. Don’t discard a recommendation just because it doesn’t have a Sanity Score.
The only segment that loses money is below 6. If the model tells you it’s not confident, listen.
Going deeper, the individual scores tell a clear story:
- Sanity 4: -9.72% ROI
- Sanity 5: -14.32% ROI
- Sanity 6: +7.31% ROI
- Sanity 7: +13.96% ROI
- Sanity 8: +13.35% ROI
The jump from 5 to 6 is where everything flips. Below 6: losing money. At 6: making money. At 7: making even more. The gap between “below 6” and “6 and above” is 17.5 percentage points of ROI.
Which AI Betting Prompt Should You Use?
Two pre-match prompts dominate the data:
+EV Goal Analysis: 1,200 bets | +118.39 pts | +9.87% ROI | 56.5% strike rate
+EV Match Odds & Asian Handicap: 516 bets | -16.53 pts | -3.20% ROI | 45.9% strike rate
The +EV Goal Analysis prompt is easily the most popular with Gecko Edge users and it is easy to see why.
It generates 69% of all unique pre-match recommendations and returns nearly +10% ROI. It covers goal markets (Over/Under, BTTS, First Half, Second Half) and Asian Handicap — the bet types that the model deliver consistently.
The Match Odds & Asian Handicap prompt runs at a loss. Despite showing higher average EV% (19.5% vs 12.9%), it includes the outright match result markets — Home, Away, Draw — where the sample size has been smaller and the higher odds recommendations have dragged down performance.
As noted earlier, the Asian Handicap selections within this prompt have proven exceptionally consistent.
The Mid-Tier League Edge
One of the more interesting findings: the model performs best in competitions where market efficiency is lowest.
Top pre-match leagues by ROI (minimum 25 bets):
- Brazil Serie A: +40.87% ROI
- Bulgaria First League: +34.88% ROI
- Germany 3. Liga: +33.43% ROI
- France Ligue 2: +28.98% ROI
- Romania Liga I: +26.85% ROI
- Italy Serie A: +21.08% ROI
- Turkey Süper Lig: +18.94% ROI
The Premier League: -19.28% ROI.
This makes complete sense. Bookmakers pour resources into pricing the Premier League, La Liga, and the Bundesliga accurately. The margins for error are much smaller. But in Brazil Serie A or Bulgaria’s First League? The pricing is softer, the data advantage of a model like Gecko Edge is larger, and the miss-pricing opportunities are more frequent.
This doesn’t mean you should ignore major leagues entirely — Italy Serie A (+21.08%) and Spain La Liga (+12.75%) both perform well. But if you’re only betting on the Premier League, you’re fishing in the most competitive pond. The edge is in the leagues the market pays less attention to.
The Key Takeaways
Back these pre-match:
- Under 2.5 Goals — the model’s standout market
- First Half Over 0.5 Goals — the consistency play
- Asian Handicap — the volume play
- Anything with a Sanity Score of 6+
Avoid these pre-match:
- Match Result Away — structurally broken
- Over 3.5 Goals — extreme predictions don’t land
- Anything with a Sanity Score below 6
- Odds above 4.00
The sweet spots:
- +EV Goal Analysis prompt
- Odds range 2.00-2.50
- Mid-tier leagues over top 5
- EV% in the 10-20% range
What’s Next
Part 3: The 15-Minute Window That Changes Everything Everything you just read applies to pre-match. In-play is a different game — and the data reveals one timing window that generates 90% of all in-play profit. Plus which in-play prompts to use, which to skip, and the single most profitable in-play angle in the entire 8,439-bet dataset.
Part 4: The Gecko Edge Playbook The complete decision framework — pre-match and in-play rules combined into an actionable system.
Part 1: We Analysed 8,439 Bets. Here’s What We Found.
See what Gecko Edge finds in tonight’s fixtures.
