How To Start A Model??

[jjgold]

We have two that I know of "Model Guys"

LT and KVB

 

[BobbyFK]

@elihu.feustel is probably the one to ask about a model JJ

 

[jjgold]

I forgot about him. He’s a big model guy.

 

[MoMoneyMoVaughn]

Start by building a decision tree that takes data through a set of binary options. That’s probably the most basic algorithm for training some sort of machine learning model.

I’m probably going to try to build a Python data scraper that spits out college hoops games using Ken Pom data.

hard part will be using past data to train the algorithm.

 

[JDS]

@elihu.feustel is probably the one to ask about a model JJ

Who the hell is that ? & where the bleep did he go …

 

[JDS]

I forgot about him. He’s a big model guy.

What forum do you know him from ?

 

[BobbyFK]

Who the hell is that ? & where the bleep did he go …

That's Justin7

 

[JDS]

That's Justin7

Ah ok yeah know that name quite well from SBR.

 

[JDS]

I must‘ve miss the day @elihu.feustel announced he was in-fact Justin7.

 

[flyingillini]

I get all my modeling done by Loshak.

 

[jjgold]

Illini might be the winner

 

[slickfazzer]

Modeling/originating too time consuming for most.

 

[jjgold]

Most of the guys it’s all automated as far as crunching the numbers

 

[MoMoneyMoVaughn]

Modeling/originating too time consuming for most.

It’s worth doing if you are in that field. One of the hardest things to do is finding datasets that are rich and detailed enough but that are also interesting to you.

if you aren’t into programming or math then it’s probably more trouble than it’s worth to learn about just for gambling cuz it takes a ton of background knowledge in data structures and algorithms to build a semi decent model and even then it’s likely to hit at the same rate that just being a good value gambler will win at.

there’s too much variance in a group of dudes trying to get a ball in a hoop to build anything that uses hard data as it’s North Star.

if you can build something that spits out good data AND use your sports knowledge then it’s something that can help you find angles you probably might have missed with either strategy in isolation.

 

[KVB]

...there’s too much variance in a group of dudes trying to get a ball in a hoop to build anything that uses hard data as it’s North Star.

if you can build something that spits out good data AND use your sports knowledge then it’s something that can help you find angles you probably might have missed with either strategy in isolation.

If you can build something that spits out data then build something that includes your sports knowledge too, and spit out better data.

Learning to quantify that knowledge, well, that's the trick.

Track your bets and why you made them...that's the answer.

 

[MoMoneyMoVaughn]

If you can build something that spits out data then build something that includes your sports knowledge too, and spit out better data.

Learning to quantify that knowledge, well, that's the trick.

Track your bets and why you made them...that's the answer.

I just mean that the sports knowledge is built into the algorithm you train the model with.

Sometimes it will spit you out a game that you know just ain’t so.

I’m trying to do some work in neural networks this year. Not entirely sure it’s the right Tech for this kind of application but I’m going to try and adapt it if I can figure out something that makes sense

 

[kidcudi92]

First find a good looking girl, then have her make picks

 

[MoMoneyMoVaughn]

My girl is a better caper than me but I’m too embarrassed to ask for her picks cuz it’s emasculating

 

[chilidog]

I run python scripts every morning that scrape a few sites for yesterday's stats, then connect to some APIs over at rapidapi to grab lines.

There's a few things left I'd like to automate, just haven't gotten around to them yet.