Long term profitability for a successful manual trader now that AI is here

[Big Bad Barney]

This is why I stick to laying/betting because I am in complete control of what I decide to lay/bet .

Ain't that the same thing as trading? just doing it once per market instead of multiple times?

Pretty sure that Elon dude reckons we're all gonna have AI doing everything for us by 2030 ...a meme I saw today said that we don't need to save for our retirement according to him... he's a smart dude right? So yeah... you won't have to lift a finger to do anything...just wait 4 years...

[Euler]

While on a short break recently, I read a load of academic papers on the use of AI on sports betting and / or Betfair and was surprised at how far off the mark they were.

Trying to solve problems that made no sense, a lack of understanding of the way the markets work, etc.

Some were quite extensive, but seemed to completely miss the mark. It seemed like they were trying to apply ML to things for the sake of it, rather than seeing what the problem was and trying to solve it.

I see this a lot, and I think Betfair trading has long suffered from 90% of advice coming from people who don't actually do it, and it now appears to be baked into LLM's as well. It's curious.

[wearthefoxhat]

While on a short break recently, I read a load of academic papers on the use of AI on sports betting and / or Betfair and was surprised at how far off the mark they were.

Trying to solve problems that made no sense, a lack of understanding of the way the markets work, etc.

Some were quite extensive, but seemed to completely miss the mark. It seemed like they were trying to apply ML to things for the sake of it, rather than seeing what the problem was and trying to solve it.

I see this a lot, and I think Betfair trading has long suffered from 90% of advice coming from people who don't actually do it, and it now appears to be baked into LLM's as well. It's curious.

+1

Makes sense, as the LLM train, is trained on the stuff us humans shovel into the A.I. stove. Of course it does depend on what's been shovelled...

LLM Express.png

[Sonny]

While on a short break recently, I read a load of academic papers on the use of AI on sports betting and / or Betfair and was surprised at how far off the mark they were.

Trying to solve problems that made no sense, a lack of understanding of the way the markets work, etc.

Some were quite extensive, but seemed to completely miss the mark. It seemed like they were trying to apply ML to things for the sake of it, rather than seeing what the problem was and trying to solve it.

I see this a lot, and I think Betfair trading has long suffered from 90% of advice coming from people who don't actually do it, and it now appears to be baked into LLM's as well. It's curious.

https://arxiv.org/abs/2303.06021

Some get at least the basics right.

[tico]

Ain't that the same thing as trading? just doing it once per market instead of multiple times?

Pretty sure that Elon dude reckons we're all gonna have AI doing everything for us by 2030 ...a meme I saw today said that we don't need to save for our retirement according to him... he's a smart dude right? So yeah... you won't have to lift a finger to do anything...just wait 4 years...
Top

Hi<
I see your point but trading is different to betting and laying which will take place over seperate events.
Trading tends to take place in the one event.
As for elon musk , hmmm ,he is a parable for the modern world where people of limited abilities shout and scream about how wonderful they are and suddenly they have a following .The problem with Musk is that his success is based on the greed of the money men .The Chines are light years ahead in the AI race with most of thier models created by non-profit making groups . One thing that has always annoyed me about Musk is that he dared to associate himself with Tesla . Talking about standing on the shoulders of giants!! .The guy is not fit to polish Tesla's shoes.It's no coincidence that Musk has teamed up with Trump...two of a kind.

[tico]

While on a short break recently, I read a load of academic papers on the use of AI on sports betting and / or Betfair and was surprised at how far off the mark they were.

Trying to solve problems that made no sense, a lack of understanding of the way the markets work, etc.

Some were quite extensive, but seemed to completely miss the mark. It seemed like they were trying to apply ML to things for the sake of it, rather than seeing what the problem was and trying to solve it.

I see this a lot, and I think Betfair trading has long suffered from 90% of advice coming from people who don't actually do it, and it now appears to be baked into LLM's as well. It's curious.
Top

Hi<
I think you are missing the point here .AI is not simply the sum of the imported information .It functions on Algorythmic structures and multiple arrays thereof. What (some) models do is create allegories of meaning and this is the element that will make or break humanity . When you read about AI getting things wrong it still doen't discount what it does ,it's vastness and potential.The future is AI we just need to ensure that it is developed in safe hands and not by morons as is the case in the western hemisphere .

[Euler]

I know how it works, I was just surprised at how badly used it was in the papers I was reading. It was like they were trying to light a fire with water.

[ShaunWhite]

AI at present isn't sensitive to the context, it will tell you how to wire your house so it'll catch on fire because 'double check & validate isn't a default, and it doesn't understand the seriousness of the topic. Similarly when it come to gambling it doesn't recognise that most content is crap.

It's not artificial intelligence it's augmented intelligence. It's not going to find money for you but it will teach you how to use the tools that will because it's good at generalisations such as training.

But on the general point of edges/secrecy etc, it's not about knowing edges (they evolve or fade over time anyway) it's about knowing how to find an edge. It's the classic case of having a net rather than a fish.

[tico]

AI at present isn't sensitive to the context,

https://www.youtube.com/watch?v=ARJ8cAGm6JE

[Big Bad Barney]

The Chines are light years ahead in the AI race with most of thier models created by non-profit making groups .

Yeah agree on musk...guys a knob...as far as I can tell from watching TV anyway ...sorry to bring him into the conversation.. onward and upward though..... I am not so sure about Chinese doing well in the AI race. Anecdotally, None of the models I use come from there, they are all American (could just be I am unco).... I recall when Deepseek had a blip on the radar and NVIDIA went down lots, (presume because they didn't actually go and try deepseek, just saw the youtube video that they did it cheaper).

But yeah, to me, context is the biggest problem with any AI...feeding it that is the pain point... I would think that is a manual trader's edge..