Our analysis determines our fate.

[Ryu_Ar1]

I've always been interested in some posts/discussions that relate AI can help us in betting especially in sportsbetting because some of us even quite confidently say that AI predictions can be greater than our own analysis in terms of percentage. like some post This which is got me interested do some steps testing.

Can AI really predict the outcome of a sports match?
If personal prediction is failing you try AI prediction it works 65%

because I'm participating in a contest on Sportsbet.io's English Premier League Pools So this could be something that encourages me to get a test whether the AI analysis is that good or not.
Last week in week 16 of the EPL I tried asking several AIs such as google assistant and Chat GPT (I also tried deepseek but the data became less meaningful so I didn't enter it) and this is what I got.

Predictions for week 16 EPL chat GPT                                                
 


Predictions for week 16 on google assistent


My personal analysis on Superbru


I tried to compare with the system in superbru where when the score is correct it will be assessed with 3 points, if the score is close it will count 1.5 points and if it is wrong then it is 0 points.
and here are the results of the whole.



From this I have concluded that in the end AI is still AI where it cannot be used as a benchmark for making decisions and can only be an auxiliary reference rather than a direct decision.

Although my personal analysis is also not very good because from 10 matches only 11 points (which should be if the perfect point means 30 points) but my personal analysis is slightly better than the predictions made by several AIs where there are only 5 and 7.5 points if I calculate with the format that superbru does.
So from this even though in the end AI becomes an important element for now but we should not make this too much as a benchmark because in the end in gambling especially in betting of course our own decisions determine our fate.

[Sticky Bomb]

These predictions are about correct scores and not a dynamic prediction pattern we follow, I hardly predict scores. I base more on point spreads. AI is not a perfect prediction channel, not even near perfect actually. I cannot trust it's predictions, I record a better success rate when I do the prediction myself compared to using AI.

AI predicts strictly with performance data and results, but predicting as a person, beyond data and research, you'll need to apply some rational assessment based on knowing the teams involved, individual players and observations from actually watching them play in previous matches to form your decision in selecting some picks.

[$crypto$]

No matter how much you analyze using AI from different tools it still won't be a benchmark --- even when the results are worse than we think making your own decisions is best in sports betting coupled with a strong feeling can determine your betting confidence.

I use AI only as a tool for analysis and some absent players, the rest is looking at h2h etc.
Sometimes I feel my own bets are better than AI.

[Odohu]

I don't believe so much in using AI for prediction because I add a little bit of my personal sentiment in my analysis which make me completely ignore some teams irrespective of what the data says. For instance, I will never bet on Borussia Dortmund, Manchester United and Liverpool now irrespective of who they place. If you put these teams in AI, they will give clear signals but in reality, they are super inconsistent to me and have caused me more losses than I have won with them. In this case, avoiding them is the best decision to take.

[Proty]

As long as gambling is concern not even AI can supercede luck. If the luck isn't on our side even if we are to make use of different AI to predict the outcome of a bet it will still not be a win. The only thing AI will do is to guide us while doing our analysis but to think that the prediction given by AI is %100 sure , it is wrong. If AI is actually this powerful to predict the outcome of a bet perfectly then I don't think if casinos will ever be operating.

[Pablo-wood]

No matter how much you analyze using AI from different tools it still won't be a benchmark --- even when the results are worse than we think making your own decisions is best in sports betting coupled with a strong feeling can determine your betting confidence.

I use AI only as a tool for analysis and some absent players, the rest is looking at h2h etc.
Sometimes I feel my own bets are better than AI.

I think AI put most factors into consideration before coming up with most analysis, like h2h stats, team strength, winning streak, current form and teams overall performance, although they are not always accurate but they can serve as a good tool to boast our confidence and help adjust certain predictive moves to make better predictions. I am not saying they are always accurate but I am only saying their guesses combined with our experience can be a big advantage.

[Oshosondy]

So from this even though in the end AI becomes an important element for now but we should not make this too much as a benchmark because in the end in gambling especially in betting of course our own decisions determine our fate.

Good work OP, it is good for people to know that AI can not help in gambling and trading if someone is looking for money from them, especially gambling that people are gambling against the house, people will only use AI and lose more money as they continue to gamble. AI are good but they are used for something different and not used for having accurate analysis in gambling.

[Julien_Olynpic]

Of course, artificial intelligence can only do what the industry can do on average, as this is common knowledge and common statistics. Perhaps some breakthrough will come with customized artificial intelligence, but we can't say for sure. Better forecasts require better methods, more advanced methods than currently available. And artificial intelligence can only quickly perform mechanical tasks, which humans are poor at.

[danherbias07]

I am not a fan of AI, and that is why I don't use it.

When it comes to predictions in our favorite sports, I still believe that the power of the human mind is the better choice for decision-making. In case of losing, we only blame ourselves for not doing it right, and we won't be mad with anything else.
Plus, the satisfaction level when a bet wins from our own prediction is way too satisfying compared to relying on the prediction of others or an AI.

Still, I believe the AI can only give us facts, but it won't help when it comes to the results of the game.

[Sim_card]

AI can only give the answers based on what is iin the database and not about what is to happen in future. Human programmed AI and human cannot predict what will happen tomorrow, therefore, is incapable to predict football matches and is useless to depend on. AI can only assit you in your research and analysis to make it faster for you.

[Botnake]

AI also has some limitations. We may see it as helpful to other things, but that doesn't mean it fits everything.  It could be used for predicting which team would win, but it wouldn't be correct. The same with us, and even if we spend the whole day analyzing the data, we can't assume it will be right as well.

We can use all the resources we have, including AI, but this is gambling, and it guarantees nothing. So I think we leave that way and rely on ourselves rather than relying on AI.

[MAAManda]

This is a good test to make a decent comparison, mate. In the end, nothing is certain in gambling. In week 16 of the PL, you were better than all the AI ​​predictions, but in week 17 or 18, it could be that the AI ​​predictions are better & more accurate, that's the reality. Personally, I'm not convinced by all the predictions I see. IMO, that's the stupidest way to lose money at sportsbook. I'd rather lose money based on my analysis & confidence, it gives me experience & lessons.

[rbynxx]

From this I have concluded that in the end AI is still AI where it cannot be used as a benchmark for making decisions and can only be an auxiliary reference rather than a direct decision.

Although my personal analysis is also not very good because from 10 matches only 11 points (which should be if the perfect point means 30 points) but my personal analysis is slightly better than the predictions made by several AIs where there are only 5 and 7.5 points if I calculate with the format that superbru does.
So from this even though in the end AI becomes an important element for now but we should not make this too much as a benchmark because in the end in gambling especially in betting of course our own decisions determine our fate.

They should be treated as such, they are just a guide for your decisions, more like advisors to me and a betting assistant especially if I want some stats that only they can articulate within minutes. Using AI is also our decisions and everything revolves us and we don't have to say it that it comes from an AI's fate, it's on us.

[crwth]

So from this even though in the end AI becomes an important element for now but we should not make this too much as a benchmark because in the end in gambling especially in betting of course our own decisions determine our fate.

Nice experiment, OP. Well, one thing is for sure: the personal analysis would have some shred of emotion. It could have influenced how it went through, and I agree that AI is valuable, and it's for reference, not for making decisions immediately and maybe pointing to AI sports data for real-time data stats? I think this could also improve the prediction.

[Zigabel]

From this I have concluded that in the end AI is still AI where it cannot be used as a benchmark for making decisions and can only be an auxiliary reference rather than a direct decision.

Although my personal analysis is also not very good because from 10 matches only 11 points (which should be if the perfect point means 30 points) but my personal analysis is slightly better than the predictions made by several AIs where there are only 5 and 7.5 points if I calculate with the format that superbru does.
So from this even though in the end AI becomes an important element for now but we should not make this too much as a benchmark because in the end in gambling especially in betting of course our own decisions determine our fate.

Relaying on AI to make decisions can be of some really huge disadvantage, it should not be our benchmark of decision, it should only assist with some data collection. gambling with AI prediction can come off good sometimes and in the other time it may cause you losses because AI does work with data made available to it beyond which it cannot assert expectations and mind you, AI may not be bale to understand the human factors that will come in if you are betting on sport games, this is why you sometimes may have even a better prediction than some of these AI if you are diligent with you data and follow-ups with the happenings around the team to get enough information that will be used to make predictions.

[cryptomaniac_xxx]

I think we have several threads already trying to used AI in sports betting and it didn't went well. So I guess for now we can truly say that AI is useless if we are going to used it for sports betting as it might not capture several factors as to why certain things can pull a upset.

So just like here in what the OP posted, doesn't look good for his prediction. With that, it's better for us to really make our own analysis. We all know that sports betting is one of the most exciting games and so I don't think that we should allow ourselves not to make our own bet base on what we see and depend on AI to make predictions for us.

[X-ray]

AI is good for the past game research only, while it can't be used to determine the upcoming result as it's relied upon the past game data. So theoretically, using AI for the future prediction is possible, but it's gonna be less accurate caused by it used past game data. Meanwhile, game result determination involved so many variables that can happen anytime such as injury, gameplay change, subs, and etc.

This is the reason using AI for the sportbetting is just wasting of time.

[rdluffy]

I agree with you, especially on this point:

... can only be an auxiliary reference rather than a direct decision.
...

AI can help us a lot in analyzing a match
You can request a history of matches between the two teams, the results over the last 10 years, the percentage of wins or losses for that team when playing away from home, etc

But the final analysis is up to us, and that “feeling” in sports is very important
There are some games that I can identify and win good bets on, especially in soccer in my country, which I have been watching for several years
Human perception is still more valuable than AI for betting

[Kasabus]

From this I have concluded that in the end AI is still AI where it cannot be used as a benchmark for making decisions and can only be an auxiliary reference rather than a direct decision.

Although my personal analysis is also not very good because from 10 matches only 11 points (which should be if the perfect point means 30 points) but my personal analysis is slightly better than the predictions made by several AIs where there are only 5 and 7.5 points if I calculate with the format that superbru does.
So from this even though in the end AI becomes an important element for now but we should not make this too much as a benchmark because in the end in gambling especially in betting of course our own decisions determine our fate.

This is to prove that there is no tool to defeat the casino. AI technology is helping a lot. However, for some reason, we can't say that it will dominate in all things. To those who think that AI makes them win, I will say that it was just a big coincidence. The same thing we do when speculating who will win. Sometimes we're right, and sometimes we're wrong.

In gambling, everything is unpredictable because even a favorite team doesn't always win. How much more just relying on data?

[alani123]

Why would AI be good at predicting matches when most of the data is not even quantifiable.

How is an LLM going to know which players are sick, find scorecards etc. These materials aren't exactly built to be parsed. If you want to build your own model and have subscriptions to data sources it might be better at creating results based on data rather than just AI slop as most LLMs would do, but really it's not something that can guarantee results. Because even data driven predictions fail at gambling due to how odds are set.