Posts: 32
Threads: 2
Joined: May 2024
Hey everyone!
I’ve been diving into the idea of using AI to predict sports outcomes, and I’m curious—what are the key factors I should consider when building or using an AI model for this?
Should I focus more on historical data like team performance and player stats? Or are things like player health, weather, and even psychological factors just as important?
What data points do you think are crucial for making accurate predictions?
Posts: 10,305
Threads: 216
Joined: Jun 2023
Political factors too. Keep us updated on this, I'd love to know what logic goes into this!
"Universal appeal is poison masquerading as medicine. Horror is not meant to be universal. It's meant to be personal, private, animal"
Posts: 23
Threads: 0
Joined: Jan 2025
There are many examples of this. I don't think it works very well.
Posts: 10,305
Threads: 216
Joined: Jun 2023
(01-28-2025, 11:28 AM)moritz1 Wrote: There are many examples of this. I don't think it works very well.
I mean, if it did, it wouldn't be a secret.
"Universal appeal is poison masquerading as medicine. Horror is not meant to be universal. It's meant to be personal, private, animal"
Posts: 23
Threads: 0
Joined: Jan 2025
(01-28-2025, 12:27 PM)DredgenSun Wrote: (01-28-2025, 11:28 AM)moritz1 Wrote: There are many examples of this. I don't think it works very well.
I mean, if it did, it wouldn't be a secret.
If the striker playing that day had a problem with his girlfriend 1 hour before the match, the analysis is nothing but garbage. In my opinion, the luck factor is 90%, analysis is 10%.
Posts: 10,305
Threads: 216
Joined: Jun 2023
(01-28-2025, 12:36 PM)moritz1 Wrote: (01-28-2025, 12:27 PM)DredgenSun Wrote: (01-28-2025, 11:28 AM)moritz1 Wrote: There are many examples of this. I don't think it works very well.
I mean, if it did, it wouldn't be a secret.
If the striker playing that day had a problem with his girlfriend 1 hour before the match, the analysis is nothing but garbage. In my opinion, the luck factor is 90%, analysis is 10%.
Completely correct in every respect.
"Universal appeal is poison masquerading as medicine. Horror is not meant to be universal. It's meant to be personal, private, animal"
Posts: 32
Threads: 2
Joined: May 2024
It was an idea that had been going around in my head, but apparently, it's not worth wasting my time on.
Maybe the hardest part is identifying the factors that have the most weight in the final results, because the range of factors is vast, and those factors aren’t always present or representative. So sad
Posts: 3
Threads: 0
Joined: Jan 2025
(01-29-2025, 03:26 PM)white_morpheus Wrote: It was an idea that had been going around in my head, but apparently, it's not worth wasting my time on.
Maybe the hardest part is identifying the factors that have the most weight in the final results, because the range of factors is vast, and those factors aren’t always present or representative. So sad
It’s an intriguing concept with potential, but sportsbooks already invest heavily in similar predictive modeling. The effectiveness likely varies by sport, but from my experience as a professional hockey player, expected goal differential (xGF - xGA) is a key metric. Teams with consistently strong differentials tend to outperform betting lines over time. For example, NHL teams ranking in the top five for xG differential over a season often exceed expectations in both standings and betting outcomes.
HMU if you're interested in discussing it further, but don't give up on the idea yet.
Posts: 217
Threads: 33
Joined: Oct 2023
Following on this as i have tried different strategies but i keep losing on my bets.
Posts: 41
Threads: 3
Joined: Feb 2025
(01-29-2025, 10:48 PM)itsguccibro Wrote: (01-29-2025, 03:26 PM)white_morpheus Wrote: It was an idea that had been going around in my head, but apparently, it's not worth wasting my time on.
Maybe the hardest part is identifying the factors that have the most weight in the final results, because the range of factors is vast, and those factors aren’t always present or representative. So sad
It’s an intriguing concept with potential, but sportsbooks already invest heavily in similar predictive modeling. The effectiveness likely varies by sport, but from my experience as a professional hockey player, expected goal differential (xGF - xGA) is a key metric. Teams with consistently strong differentials tend to outperform betting lines over time. For example, NHL teams ranking in the top five for xG differential over a season often exceed expectations in both standings and betting outcomes.
HMU if you're interested in discussing it further, but don't give up on the idea yet.
That’s super interesting! I had no idea xGF and xGA were such game-changers for predictions. I’m really curious to hear more about your pro experiences and any other key stats you think are must-know!
|
