01
DATA
Collect and normalise team form, match events, market prices and contextual information from multiple sources.
ABOUT US / OUR STORY
Our story: building football analysis around expected value.
EV Bet Engine did not begin with a pitch deck, a brand strategy or a plan to build a betting business. It was a private tool, made for one person, in the extension of a hobby.
Its founder simply loved football, spent an unreasonable amount of time analysing it, and wanted to know whether betting decisions could be made with more discipline: by treating odds as prices, matches as probability distributions, and every judgment as something that must be tested.
The more the tool was corrected and improved, the more enjoyable building it became. Somewhere along the way, it felt less like work and more like childhood again: becoming completely absorbed in a favourite game, only now the toy was made of football, data, models and code.
01 / THE ORIGIN
A different question.
Most football predictions begin with the same question: who will win? EV Bet Engine began somewhere else. What is the probability of each outcome — and is the price offered by the market worth buying?
A strong opinion about a team is not automatically a good bet. A likely outcome can be overpriced. An unfashionable market can hold value. Separating probability from price became the first principle of the engine.
The earliest version was not elegant. Odds were collected manually. Match data from FootyStats, Soccerway, Sofascore and other sources had to be cleaned and reconciled. Models were run, results were checked, and mistakes were documented one by one. It was slow, but it created the discipline that still defines the project.
02 / WHAT WE BUILT
Not a tip sheet. A decision system.
The engine connects the full path from raw information to a decision that can be explained, sized and reviewed.
02
MODELS
Estimate outcome distributions through Poisson, Dixon–Coles, Monte Carlo and market-specific statistical models.
03
VALUE
Convert model probability into fair odds, compare it with the bookmaker price, and calculate Edge and EV.
04
RISK
Use the Kelly criterion with practical adjustments, market risk factors and exposure limits to determine Stake.
05
JUDGMENT
Check tactics, match context, sample quality, market behaviour and model consistency before the final human decision.
06
REVIEW
Compare the thesis, actual bet and result; record failures as carefully as wins and return the learning to the next analysis.
03 / THE TURNING POINT — WORLD CUP 2026
The private tool produced public-scale evidence.
The 2026 FIFA World Cup became the first full proving ground. EV Bet Engine was used across pre-match analysis, live decisions and post-match review. The results went far beyond what the original personal tool had been expected to deliver.
What changed the future of the project was not one spectacular bet. It was the repeated connection between model output, market selection, disciplined Stake design and live re-evaluation. The process held together across different matches and different market types.
The easy choice would have been to protect what had worked, keep the engine private and preserve the profit. Instead, while the excitement of building was at its strongest, we chose the harder option: challenge the result, rebuild what could be better and take the tool to its next stage.
One reader became two. Comments and reviews began to arrive. Eventually, the work caught the attention of investors. A personal tool was becoming an expected-value analysis project — not because that was the original plan, but because people found something honest and interesting in watching it grow.
That was the moment a private engine became a project.
64.72%
Hit Rate
Recorded engine recommendations
+71.96%
ROI
Recorded engine performance
3 PHASES
One Cycle
Pre-match, live and post-match
Performance figures describe the recorded World Cup engine sample at the point the project was formalised. They are historical results, not a promise of future returns.
04 / A LIVING ANALYSIS CYCLE
The analysis does not stop at kick-off.
A match changes continuously. Line-ups, tempo, tactics, injuries, cards, substitutions and the market itself can invalidate a pre-match assumption in seconds.
PRE-MATCH
01
Design the conditions for value.
Collect the evidence, build the probability model, compare fair price with market price, and define what must be true before a position becomes valid.
LIVE
02
Find value before it disappears.
Re-test the thesis against xG, shots, big chances, corners, cards, substitutions, tactical changes and price movement. Live value can disappear in seconds.
POST-MATCH
03
Turn every decision into evidence.
Separate the quality of the decision from the result. Record what worked, what failed and what the system must learn next.
05 / OUR OPERATING PRINCIPLE
Speed and quality that can keep up with overwhelming information.
Live betting is unforgiving. A goal, a card, a substitution or a market suspension can erase a price before an explanation is finished.
But speed is not an excuse to publish a betting slip without context. While others may show only the transaction, we want to show the decision: Market, Odds, model probability, fair price, EV, Stake and the reason the value exists.
We do not make the work faster by removing the work. We make it faster by designing the analysis, communication and visual production as one operating system.
Fast, but deep. The value of the moment, with the reason behind it.
06 / WHAT WE BELIEVE
We do not sell certainty.
Football is uncertain and betting markets are noisy. A good decision can lose; a poor decision can win. That is why we judge the quality of the process over the drama of one result.
PRICE OVER OPINION
A football opinion becomes a betting decision only after it is tested against price.
PROCESS OVER OUTCOME
One result cannot validate or destroy a model. The complete decision process is the unit of evaluation.
TRANSPARENCY OVER HYPE
Recommendations, Stakes, outcomes, errors and improvements should leave a record.
RESPONSIBILITY OVER AUTOMATION
AI can expand scale and speed. The final betting decision and its consequences remain human.
07 / THE PEOPLE AND SYSTEMS BEHIND IT
A founder, an AI team, and one shared standard.
EV Bet Engine is led by its founder and developed with specialised AI systems across research, data collection, modelling, engineering, execution, design and review.
This is not a story about replacing human judgment. Machines process a scale of information that one person could never handle alone. The founder holds the match context, challenges abnormal outputs, owns bankroll risk, sets product direction and makes the final call.
The goal is a form of collaboration where overwhelming information does not defeat speed, and speed never defeats quality.
FOUNDER
Thesis, final judgment, risk ownership, product vision and the values that cannot be traded away.
AI SYSTEMS
Research, data processing, modelling, engineering, monitoring, design support and structured review.
THE ENGINE
A growing body of models, records, rules and learning that makes each new analysis stronger than the last.
08 / BUILDING IN PUBLIC
No record is worth keeping if it is not honest.
When the first few people began watching this unusual personal project, we made a simple choice: there was no point creating a false record to look impressive. More importantly, we wanted to be honest with anyone willing to spend time following a self-made tool that had never been intended for an audience.
That is why match reviews have shown losses as openly as wins. A good decision can lose, and a bad decision can win. Hiding either one would make the record useless and betray the reason for building it.
Match analysis shows what the engine believed before and during a match. Post-match reviews test whether that reasoning survived contact with reality. Development journals preserve the uncertainty too: the corrections, failed ideas, difficult decisions and moments when the product refuses to work as intended.
EV Bet Engine is building a database not only of bets, but of decisions. Wins are evidence. Losses, false confidence, bad prices and missed conditions are evidence too — but only if they remain visible.
09 / THE FIRST PROMISE
What we refuse to lose as we grow.
EV Bet Engine is now a real project, and profit matters if it is to continue. But profit and recognition are not why we build. We build because the process itself is joyful — because we are still completely absorbed in making the best version of this strange, favourite toy.
What matters more is gratitude to the people who stayed, read, commented, reviewed and helped shape it. We believe this is being built with them, not merely shown to them. One day, we hope that belief becomes a community where users can participate in the platform itself.
Growth is not success if it requires us to place profit above those people or betray their trust. The technology will evolve. The team will grow. The platform may become global. Match reviews will still include losses, and development journals will still show the doubts, mistakes and unfinished work.
If you are willing, please watch over this journey with us. We will not forget the people around us. That is our first promise.
10 / WHERE WE ARE GOING
A global football analytics platform, built one verified decision at a time.
The next chapter goes beyond individual match predictions. Data, models, market prices, live signals, transparent performance records and editorial research will connect in one place — creating an analysis platform that continues learning before, during and after every match.
01More leagues, competitions and market types
02Professional data APIs and stronger automation
03Transparent, filterable performance records
04Live analytical signals with human judgment
05Multilingual research for a global audience
06A platform designed for continuous learning
This is still the beginning. But it is no longer just an idea.
EV Bet Engine now stands at the starting line of the platform it was never originally meant to become.