AI and cutting edge tech can make football great again !


Football needs to invest in R&D and content quality instead of tryin to sell more units of it’s standard product

We can see that FIFA, UEFA, clubs and associations do everything they can to stop football from declining. They have tough competitors as Today’s teens might prefer social media, streaming services and video games instead of watching 2x 45mins of a real match where if you are lucky you may see 2 goals (possibly one of them from penalty). Unfortunately they do it in a clumsy manner, by simply trying to add more units of the same content (yes — they want all those players to keep playing more and more games) or by selfish initiatives like Superleague or Champions league for giants which preety much centralises football economy around less clubs -leaving lower leagues poorer and poorer. They have already exceeded players health capabilities as well as general football games demand, but still they don’t change their approach. Idea of those organisations was to promote sustainable development of that sport and by sport — peoples life quality around the globe. As you can see it doesn’t really happen that well.

My proposition would be to focus on content quality and form instead of quantity of games. These days people are closer to actual idols than every: by their social media content. Unfortunately football associations are a little behind when it comes to trends and hence there is no investment in cutting edge content. Football has great potential when it comes to data. Data can come from games stats, training, players recover and even from football itself — like data coming from sensors during game. All that is needed is making it more accessible like instagram photos are. Football should be trying to steal peoples attention from youtubers back towards sport idols.

Data is the gold of XXI century, all it needs is nice presentation

Data, data … more exciting data.

First major idea is about giving very detailed players stats on the custom social media platform — data that could be used in making bet predictions for games. I would like to be able to see the current performance stats of my favourity club updated on a weekly basis. Stats that could be standarised and compared to other clubs. Open data, basic endurance and agility tests, just like in video games: something that I could read and use to compare squads and likelyhoods of the games. That would probably also help me in discovering new people to follow, maybe the players from the 2nd shelf in the club — which might have some attributes that i don’t know about. I would like to see how my idol recovers from tough match or injury. I would like to be with them. It’s much more engaging than seeing their static picutres on Instagram. That would require some investment and commitment from players and clubs. Nobody could cheat. No big names could hide their weaknesses, coaches would be judged on those stats as well, if they put somebody on the pitch just for their surname. This would create lots of discussions and this is what we want to make it transparent and exciting again. Live. Not only for 90mins a week. Live 24/7/365. Numbers make the magic. I would like to see at least 1 smaller football league to trial this sort of idea, with nice mobile app. I bet it would boost their fan engagement by double figures as well as betting revenue for their games. I am 100% sure of it.

Trained AI algorithm can provide real time recommendations to each individual player

Is Hybrid football going to be football of XXI Century ?

You must have seen football simulation in form of video game. Computer simulates match with programmed players statistics etc. Google has recently made it’s own research environment for training AI to play football: much better than anyone ever did, it’s called: Google Research Football. It’s all open source and there are tones of already trained AIs for that environment. It’s got one hidden potential: it can provide real time recommendations for each individual player on the pitch about their move direction, pace and type of action (e.g. long pass, tackle, shot, dribble etc.). Every second it predicts best possible move to increase likelyhood of winning the game. It can be really sophisticated. Imagine that it could do the same for real players — during trainig or game. As long as we provide real time observations data from the pitch (e.g. players positions, score, ball position etc.) we could do exactly the same simulation. That could actually up the game, make it looking smoother and smarter. Arsene Wenger said that players already reached their maximum athletic stat and now is time to train the brain, and yes there is no better way than trainig with super AI. I think it could be trialled during training sessions as supplementary thing for match preparations — especially because you can try to emulate specific opponent setup and behaviour. Let’s pretend you that you play with Real Madrid next week in UCL. You could load it’s trained up to date setup — which is algorithm that simulates Real Madrid and than on the training session you could make 2nd squad to listen to headphone recommendations which would aim to replicate real Real madrid. Than first squad could either play without AI support or play with their own custom tactic in hybrid mode as well. I believe that training approach would make those games much more attractive by brining more action and goals to those matches.

This is super exciting and I appreciate there would be some resistance from community and players themselves as it would show that they don’t see as good as computer and cannot play as good without that 3rd eye from the PC.

I think initially hybrid football could co-exist in paralell to conventional one, but sooner or later it couldn’t be ruled out by mainstream stock holders (introduced as train aid, but later as part of the game). Imagine how many fans would like to see first hybrid football game played by famous clubs ?AI and top players of the world: you just cannot fail.

Altough I speak in the football context, those ideas are quite universal and sooner or later they will be adapted to other sports. We see it already to some extent in other disciplines, but there is no holistic apprach in mainstream sports. I think team sports are the ones which could benefit the most of those applications as numebr of combinations of lineups, strategies and players moves are almost limitless and this is where AI comes to maximize it’s potential by aligning all those factors in perfect setup.

If you like this article please give a clap — if I see 500 claps there will be second article with coding examples of those ideas as well as … real implementation of hybrid training with AI aid ! Yes we can do it together !

P.S. Let me know if you have ever heard somewhere else about hybrid-football, as I think it’s the first article about that concept.

AI/ data science engineer