PROS :
It looks tidy.
The data is interesting, having both estimations and odds, and an attempted bet with "confidence".
All my inverstiagtion was done on desktop version. UX and displays seemed good.
CONS :
When you offer a recommendation outside W-D-W you do not offer the appropriate odds - eg. Over 1.5 goals still shows WDW.
The estimating and confidence numbers appear naive at best.
WIN ESTIMATES :
The AI estimates for W-D-W are fairly generic at best and very sloppy at worst. I cannot work out whether you are basing it on form, bookmaker odds, or something more proprietary but I'd say most estimates are +/-10% for wins and +/- 5% for draws - basically little more than a child would do by looking at the league table. How are you directing the AI to do this ?
Example :
Bayern Munich v Cologne over 1.5 goals is estimated at 68% / 32% .. which I think means 68% to hit?
Bayern Munich are scoring 3.5 goals per game themselves ... the bookmakers are offering 1.03 for Over 1.5 goals .. suggesting a 95% likelihood ???
ODDS CALCULATIONS :
And even when you get the win estimates accuarate there is no reference or thought to implied value or EV.
Example :
Leverkusen estimated 74% to win vs Hamburg - my own sheet has them at 73% - so I am impressed with that pick.
But the price offered is just 1.25 at the bookies ... where as fair value would be about 1.33 to 1.35
So the bet is -EV and taken long term if betting at 1.25 you'd lose money on it.
OVERALL :
It looks pretty. It has some promise. But that's the candy floss covering on top.
It feels like an "AI" bandwagon - with little real thought to actually finding good bets.
As a tool - it's just going to encourage bad value betting with inaccurate data and recommendations on poor bets.
---------------------
EDIT :
I read it back after posting and it sounded too harsh.
The work you've done on display and presentation is very good. I could use the UX immediately.

It's "just" the data and bet recommendations that are dubious - but for someone like me - THAT IS THE MAIN THING.
