About
About BananaPoint
In tennis there is no team to hide behind: one player, one surface, one day's form. Our model measures all of it, and every prediction is locked the moment it is published.
Why we built it
Tennis is decided by details a ranking list cannot see: how quick the court is, how much a player has already played this week and whether one style happens to suit another. BananaPoint was built to measure those details instead of guessing them.
Our model goes well beyond win–loss records. It reads serve and return statistics, break-point conversion and form on each surface, and updates its ratings every day as results come in from the ATP, WTA, Challenger and ITF circuits.
From data to graded pick
Collecting the data
Before each match the model updates what it knows about both players: how they fare on this surface, against this kind of opponent, on serve and on return, and how much tennis they have played lately. Those measures become form and fatigue ratings, and every probability is built on top of them.
Locking the prediction
As soon as a prediction is created it is stored with a timestamp and can no longer be changed. When the match ends, the result is pulled in automatically and the pick is graded as won, lost or void.
Learning from results
Every finished match is fed back in. With that feedback the model rebalances its inputs, for instance court speed against an opponent's return errors, before it forecasts the next round.
Nothing gets deleted
A hit rate is only worth something if you can check it. That is why every BananaPoint.ai prediction is kept, including the ones that lost, and the whole archive can be downloaded for free in CSV and Excel, each forecast next to its result, for anyone who wants to recount.
What we stand by
The model makes the forecasts; our technical team looks after the model. It monitors how the model performs and makes sure injury news, surface changes and crowded schedules are reflected in its inputs. Personal opinion never goes into a pick.
Independent by design: no bookmaker owns, pays or advises us. Every forecast comes straight out of the model, with nobody editing it, so there is no one who could tilt a pick one way or the other.
How BananaPoint is funded
Running the model every day costs money, and we chose not to pay for it with ads: BananaPoint is completely ad-free. Subscriptions fund the site instead, which keeps us independent of advertisers and bookmakers alike.
Several picks are free every day, and when a match ends its prediction and result go into the public archive permanently. The subscription is deliberately priced low, because good tennis analysis should not be a professionals-only product.
The companies behind BananaPoint
BananaPoint is run by two registered companies, each with its own role:
EUNerdyTech Hub SRL
Based in Romania (EU), it builds and maintains the model and the data behind it and answers for GDPR compliance.
USKickOff Ventures LLC
A US company, it takes care of commercial deals abroad, partnerships and our relationship with customers.
Each company is accountable in its own jurisdiction for the part it runs. For legal or data protection questions, use our contact page.
Founder: Daniel Ionescu · LinkedIn
18+ · Forecasts made by AI, for information only · BananaPoint neither takes nor places bets