Topic: Artificial Intelligence Gambling Using Grok! (Results: 5/9) (Read 589 times)
I had to ask Grok about the Mike Tyson fight tonight. I got somewhat of a strange answer, but it is what it is. Grok said to bet on Mike Tyson to KO Jake Paul and to also bet on Jake Paul to win. I found that a bit fascinating. I don't have the ability to bet on Mike Tyson to win by KO as far as I can see (I'm not the most familiar with gambling platforms) so I don't know what the odds would be if you made both of those bets, but I assume maybe it would be profitable either way? Anyway, I don't really know how you'd say a win or a loss on this one, so I'm going to give Grok the day off in honor of Iron Mike.
Add the odds for those bets it is telling you to place, it's one thing to track them when it's 1x2 but if you're playing handicap it's a pain in the ass to look them up, especially since they are SP bets in most archives, not the ones you placed maybe one day before the game.
Indiana Pacers was a win at 2.32
Thunders beat Clippers at 1.37
Two more matches lost -2
Those 4 were not any better and you're still down in the long run, funny enough if that handicap option in Dallas-Utah which I see at 1.90 (?) comes out you break even.
But overall I would say that the current result shows the same randomness any bettor would do, in the long run I doubt it would turn any profit over 10% let's say at maximum in a best case scenario but breaking even maybe the actual range where it lands.
Indiana Pacers was a win at 2.32
Thunders beat Clippers at 1.37
Two more matches lost -2
Those 4 were not any better and you're still down in the long run, funny enough if that handicap option in Dallas-Utah which I see at 1.90 (?) comes out you break even.
But overall I would say that the current result shows the same randomness any bettor would do, in the long run I doubt it would turn any profit over 10% let's say at maximum in a best case scenario but breaking even maybe the actual range where it lands.
Today Grok says that the bet to make is between the Dallas Mavericks and Utah Jazz. Grok is expecting the Mavericks to not only win but to cover the spread and is recommending to bet on the Dallas Mavericks (-8.5) to cover the spread over the Utah Jazz. That’s a pretty aggressive bet, but I like it.
Looking like Grok might get itself another W today. I feel a little bit bad that I didn’t do a testing phase before starting this. It took me a minute to get the prompt right to where Grok was using it’s insight instead of just picking a game’s winner based on odds. The 3/7 start before I fine tuned the prompt has hurt the 5/6 performance since (probably 6/7 in another hour).
Edit: I guess I spoke too soon. Grok took another hard loss on a good bet.
Edit: I guess I spoke too soon. Grok took another hard loss on a good bet.
Grok has surprised me yet again today. When asking it to select a matchup it had unique insight into and then select a bet, it gave me the response that Luka Doncic is going to get more than 7.5 assists and that is what I should bet on. I didn't even know it would be able to select player prop bets, but here we are. Luka on the over for assists in the Dallas Mavericks vs Golden State Warriors matchup (aka the return of Klay Thompson).
Today Grok has not only selected a winner, but given me a score. I won’t bet based on the spread and score because Grok didn’t tell me to, but I am betting the Thunder to beat the Clippers. Final score will supposedly be 116-111 Thunder.
Nice innovation, but it still depends on what actually happens right?, especially if it's a sports bet. I thought it was more of a game made in a gambling platform. If you a gambler doesn't really know how the teams perform or how strong they are, AI can be a good option. But I doubt it’ll consistently give you an 70-80% win rate, especially since it lacks context on team dynamics, recent player performance, injuries, and other factors that human analysts typically evaluate. AI can look at historical data and make educated guesses, but sports are unpredictable, with real-time factors that can impact outcomes. It’s tough to get a high win rate, and without in-depth, real-time insights, even human experts struggle to keep up.
What do you think would happen if you just relied on your own understanding of the game? Might even get you above 6/10, but still, it’s an impressive innovation. I'm really interested to see how AI can do this much already, especially for minimizing risks in betting. I might try it too with the combination of my own insights and AI, could increase my chance.
What do you think would happen if you just relied on your own understanding of the game? Might even get you above 6/10, but still, it’s an impressive innovation. I'm really interested to see how AI can do this much already, especially for minimizing risks in betting. I might try it too with the combination of my own insights and AI, could increase my chance.
Ohhh… Today Grok picked its first underdog to win. I find that interesting. It said in todays matchup with the Pacers and the Knicks, that the Pacers were going to beat the odds and win the game. I’m very curious how this will turn out, as it is the first bet against the odds.
Indiana Pacers beating the New York Knicks
Indiana Pacers beating the New York Knicks
So let's see the results:
Dallas Mavericks beating the Orlando Magic +1.37
New Orleans Pelicans beating the Atlanta Hawks -1
Brooklyn Nets beating the Detroit Pistons -1
Boston Celtics beating the Hawks +1.19
Oklahoma City Thunder beating the Magic +1.13
Los Angeles Lakers beating the Pistons -1
Boston Celtics beating the Golden State Warriors -1
Milwaukee Bucks beating the Utah Jazz +1.22
Cleveland Cavaliers beating the Golden State Warriors +1.64
Chicago Bulls (+3.5) covering against the Atlanta Hawks +1.88
If my math is right you're down $1.57 for $10 bet despite a win rate of 60%.
Dallas Mavericks beating the Orlando Magic +1.37
New Orleans Pelicans beating the Atlanta Hawks -1
Brooklyn Nets beating the Detroit Pistons -1
Boston Celtics beating the Hawks +1.19
Oklahoma City Thunder beating the Magic +1.13
Los Angeles Lakers beating the Pistons -1
Boston Celtics beating the Golden State Warriors -1
Milwaukee Bucks beating the Utah Jazz +1.22
Cleveland Cavaliers beating the Golden State Warriors +1.64
Chicago Bulls (+3.5) covering against the Atlanta Hawks +1.88
If my math is right you're down $1.57 for $10 bet despite a win rate of 60%.
A 60% winning rate using AI that choose the heavy favorite is something that is not impressive at all considering that the pick is not that complicated.
A gambler can haves same result by just picking 10 matches with same type of pick and enjoy the game at the same time. I doubt using this AI to predict matches is worth it for now because its pick has less accuracy while still choosing normal pick.
I’m sure this will get worse on larger bet sample.
A gambler can haves same result by just picking 10 matches with same type of pick and enjoy the game at the same time. I doubt using this AI to predict matches is worth it for now because its pick has less accuracy while still choosing normal pick.
I’m sure this will get worse on larger bet sample.
Of course we have to fine tune any AI chatbot before expecting from them to give us a 25 game winning parlay with a really good odd, that is asking to much from them right now in the phase where we are now. It may look like nothing special the fact that this chatbot predicted 6 out of 10 games correctly as a start yet I would celebrate such happening. Now I just checked the website regarding Grok and in the premium X service where you are able to get the services it says that the developers are making sure the Grok second version makes sure gives users a lot more than the actual version is capable so we have to see. Machine learning and data models is what feeds the AI so we have a long way before using these chatbots to make us winners in the gambling sector.
I've decided to let Grok do some gambling as a test to see how artificial intelligence would perform if given the opportunity. I am only using simple bets, like who will win each game. I will track the results here.
Even though I'm not sure the chatbot/Al can work well in predicting the NBA sports bets you are betting on, I probably don't want to forbid you from using the Grok-1 AI, because the Grok-1 AI chatbot is relatively new and was created by the Elon company, but there's no harm in testing the tools you use in gambling, but my little advice is don't place too big a bet on the team you want to bet on.Like the basketball team I saw between the Brooklyn Nets vs. Detroit Pistons, there said the Nets beat the Pistons. I'm not sure the result could be what the AI Grok-1 said, because last Monday's results 11/4/25 (92-106), as well as the Dallas Mavericks vs. Orlando Magic the same day also produced points (108-85) 108 to the Mavericks and so on several other basketball teams.
For me it's not a problem if you want to test the artificial intelligence of AI Grok-1, but you shouldn't bet large amounts, for $5 I think it's not a problem, unless you predict it yourself, that's another story.
A 60% winning rate using AI that choose the heavy favorite is something that is not impressive at all considering that the pick is not that complicated.
A gambler can haves same result by just picking 10 matches with same type of pick and enjoy the game at the same time. I doubt using this AI to predict matches is worth it for now because its pick has less accuracy while still choosing normal pick.
I’m sure this will get worse on larger bet sample.
A gambler can haves same result by just picking 10 matches with same type of pick and enjoy the game at the same time. I doubt using this AI to predict matches is worth it for now because its pick has less accuracy while still choosing normal pick.
I’m sure this will get worse on larger bet sample.
Of several groups are on the same slip and you lose even one, you can't count the rest as a success. So I wouldn't think this strategy is that sane to be honest. Seems so far it would result in a loss of funds in most cases sooner than later.
AI isn't particularly capable of predicting matches, it doesn't losses any knowledge we're unaware of especially for ongoing events. It's just going to hallucinate its way toward making these predictions. But really at best it's like a random choice.
AI isn't particularly capable of predicting matches, it doesn't losses any knowledge we're unaware of especially for ongoing events. It's just going to hallucinate its way toward making these predictions. But really at best it's like a random choice.
I am not convinced of the results. 5 or 6/10 is something a good bettor should get anyway, without using AI. Considering that it usually picks big favorites to win with small odds, you might still be in a loss overall even if it guessed 6/10 games so far. It would be good to know if Grok takes into account and has data on injuries and players who won't play. Perhaps you should ask it to provide you sources that it's basing its predictions on and list everything that has influenced it to make such prediction.
I think we cannot rely on the AI at all for placing the bets. It’s just choosing a team on the basis of fan favorites and randomly claiming to be the winner. It’s just 50-50 chances, and hence we are able to see 6/10 matches are passed. I think the AI models needed to be further trained with more data and how a player behaves at a particular playing condition. Then we might see some improvement in the result. I would not risk my hard-earned money on a bot. Not in the current situation, at least.
Yeah it’s just not there yet, but give it another 10 years or so and then revisit this conversation! But honestly the initial bets that OG had weren’t bad for 6/10 60% winning is pretty decent
A 6/10 isn't bad already because if you bet a dollar on each of them, you will end up winning (of course the odds are a factor as well).TBH, threads like this are the ones that I'm waiting. I mean I've seen threads about trading using AI, but not gambling using AI. I will follow this thread by OG, and TBH, it would be better if he will give some prompts that he is using so that we can also do an experiment on it or make it better as well. Overall, I will only try this if it goes to 100 bets coming from AI already. I'm expecting nothing, but this is very interesting.
Or TBH, I might try doing it myself as well in the future because I don't have extra funds for now that're for gambling. I know that there are some who are still skeptical about bets being done by AI, but this is just an experiment, and isn't it interesting that 3-5 years from now, we will be using AI to increase our chance in our sports bets?
Don't expect too much from this experiment knowing that we are still in the early stages of AI, but it's good that we are seeing a development already. 
From the very beginning I believe testing an artificial intelligence to see whether it could be considered to be a reliable mechanism for gamblers to improve their yield on bettinghouses and casinos is a fruitless experiment. Most artificial intelligences do what any person is already able to (if we have enough time to analyze data in bulk).
If there is any advantage when using artificial intelligences for betting an gambling, then it would be minimal and casino would end up adjusting their fees for them to continue to be profitable, in spite of having a significant amount of bettors opting to read what a machine has to say when comes to where one is supposed to put money on and when to avoid to bet, all for the sake of better chances.
Nevertheless, it is interesting but I am already aware of what the result is going to be... it is going to be lackluster.
If there is any advantage when using artificial intelligences for betting an gambling, then it would be minimal and casino would end up adjusting their fees for them to continue to be profitable, in spite of having a significant amount of bettors opting to read what a machine has to say when comes to where one is supposed to put money on and when to avoid to bet, all for the sake of better chances.
Nevertheless, it is interesting but I am already aware of what the result is going to be... it is going to be lackluster.
Yeah it’s just not there yet, but give it another 10 years or so and then revisit this conversation! But honestly the initial bets that OG had weren’t bad for 6/10 60% winning is pretty decent
From the very beginning I believe testing an artificial intelligence to see whether it could be considered to be a reliable mechanism for gamblers to improve their yield on bettinghouses and casinos is a fruitless experiment. Most artificial intelligences do what any person is already able to (if we have enough time to analyze data in bulk).
If there is any advantage when using artificial intelligences for betting an gambling, then it would be minimal and casino would end up adjusting their fees for them to continue to be profitable, in spite of having a significant amount of bettors opting to read what a machine has to say when comes to where one is supposed to put money on and when to avoid to bet, all for the sake of better chances.
Nevertheless, it is interesting but I am already aware of what the result is going to be... it is going to be lackluster.
If there is any advantage when using artificial intelligences for betting an gambling, then it would be minimal and casino would end up adjusting their fees for them to continue to be profitable, in spite of having a significant amount of bettors opting to read what a machine has to say when comes to where one is supposed to put money on and when to avoid to bet, all for the sake of better chances.
Nevertheless, it is interesting but I am already aware of what the result is going to be... it is going to be lackluster.
Grok throwing us a curveball today. For the first time it didn't select a winner, but selected a bet to win. It first chose the game between the Chicago Bulls and the Atlanta Hawks, then it said that the Bulls were likely to cover the spread and may even win the game outright. I'm taking this as a sign that it wants me to bet the Bulls to cover the spread against the Hawks. A first for my Grok predictions, but I like it.
Chicago Bulls to cover the spread against the Atlanta Hawks.
Atlanta Hawks - Chicago Bulls
Handicap (Incl. Overtime)
Chicago Bulls (+3.5)
Odds
1.88
Chicago Bulls to cover the spread against the Atlanta Hawks.
Atlanta Hawks - Chicago Bulls
Handicap (Incl. Overtime)
Chicago Bulls (+3.5)
Odds
1.88
Perhaps I've found the special sauce by asking it what game it wants to bet on instead of asking it for the winner of a game. It's also possible that Grok just got lucky 2 days in a row. Tonight's pick of the Cavs over the Warriors wasn't a cake walk though. We'll see what Grok has in store for us tomorrow, but I for one am glad that it is back above 50% accuracy. Considering it hasn't bet against the odds once yet, it should at least be performing this well.
If you want get some realistic result from your experiment please add ODD of the event you're going to gamble.
Since you're already selecting the matches to gamble in this way you can avoid any kind of bias.
I think that you must also add in your counter the odd of each bet. I mean you have a result of 3/6 = 50% of your predictions were profitable.
But what was the odd? It's not the same guess 50% of bets at @1.01 or ... @10.00 . You can calculate the "profit" or ROI and not only how many times your prediction was correct.
Maybe this system as a better rate with certain odd and not other one... I can help you with these calculations if you provide odds in decimal format
Personally I am still skeptical a system like this can be something useful for an avarage gambler but who knows if
it is not the same result if a bot is able to give prediction of easy odds... its not the same guess a result 1.01 or 10! But what was the odd? It's not the same guess 50% of bets at @1.01 or ... @10.00 . You can calculate the "profit" or ROI and not only how many times your prediction was correct.
Maybe this system as a better rate with certain odd and not other one... I can help you with these calculations if you provide odds in decimal format
Personally I am still skeptical a system like this can be something useful for an avarage gambler but who knows if
Of course if you will count both as "winner" but it's not the same
if you don't consider also the odd of the event... this experiment is just useless it's more "toss a coin" and the fact that you're close to 50% well... Jump to: