I've built my own model, but it failed. I mean, the data I used, gave no clear answer of what I should bet. Getting the right data is a serious work and the most important step to profit.
We were in the same spot as you when I published this video two years ago. Since then, we've built backtesting engines and defined "Probability Thresholds" to allow our software to pick the bets from all the value bets identified. This project has expanded a lot now, and it's the foundation for my company, Systematic Sports, which I run with my business partner, Estephe. www.systematicsports.co.uk
even if the odds are mathematically worse they need to provide better odds to the other team in order to cover the bets. Lets say statistically team A has 70% and team B has 30% chance of winning. If 90% of people bet team B you can't provide the 30% odds because even if that outcome is less likely you would lose too much money in that unlikely event, so they even up the odds depending on the volume of bets.
Yessir! I've been working on it ever since this video! I built it out into a company with my co-founder Estephe and you can find out more about what this project became on our website. www.systematicsports.co.uk
I didn't make this video to gloat. The entire output of this code that I've written is available on my website for everyone's benefit. There's even snippets of it on my Medium page
@@LiamHartley Hi liam thanks for providing this! But I am unable to locate this on your site unless I am being blind. Reason I ask is I want to replicate this but for my own understanding and for me the best way is to learn through already written code. Many thanks
Its fun making models ,there use to be a brilliant website years ago some math professor posted his models for free ,excel related ,if you want to make money make it easier for yourself than footie.
This is the best thing I have ever watched. Over the past few weeks I have flirted with attempting to make my own algorithm, not quite realising how much needs to be done even with my own limited knowledge of coding. I'll learn as I go I suppose. Great work.
The data is the secret sauce (source)! It's not something that I can share publicly unfortunately, but you can find out more about what this project evolved into on the Systematic Sports website: www.systematicsports.co.uk
I do (@liamjhartley) but I don't really use it. @liamhartley.co.uk on Instagram is the best place to keep up with me and the bets. I've now bet £1,300 and down about 3% at the moment so it swings in roundabouts! All the bets I've made are on my website :)
Absolutely. I put a video out on my tiktok about this the other day. We’re in the middle of getting ready for the next football season right now! vm.tiktok.com/ZMNDQrUVv/?k=1
I am working on my own algorithm and want to try it on my own, so I don't watch the video yet. I will come back to it after successfully running a code or after giving up. 😂
@@LiamHartley It would be a great idea to improve the current articles with more info. Including, if possible, further details (even if we have to pay for it, that would be great).
Would you have any resources on scraping the bookies odds and then importing the data with python? I know you wont give away the secret sauce of the processing but I'm just looking to get started by getting in the odds for horse racing, have limited java knowledge.
Here’s an article I wrote about scraping lineups from an old data source, you should be able to apply the same principles to bookmakers python.plainenglish.io/how-to-scrape-predicted-lineups-from-fantasy-football-scout-with-python-b11e92ea57f2
Whoa man one of the gratest videos I have ever seen! I learning pyhton right now and this is the type of thing I want to do; Amazing job, you will doing a video explaining how u have done it in python or something?? or if it possible could you show the scrpit or something?? Impressive job man Congrats!!!
You can find some more code snippets on my Medium page: medium.com/@liamjhartley I won't be showing all of my code but I will be explaining how you can learn to build your own in the future!
Great video.The only thing i dont understand is in your site in the category with suggestions for value bets how do i calculate how mutch i have to bet? Lest say my bank is 100$ and in the value bet the Kelly Criterion is 0.50. How do i calculate the amount i have to bet?
Kelly Criterion: F = BP - Q / B , F = percentage of bankroll you should allocate B = payout in terms of odds P = probability of winning Q = probability of losing
We do all of these calculations for you in our new project, Systematic Sports. We use the probability of the outcome that is calculated by our algorithm to generate the value for the Kelly Criterion. Happy to answer any other questions on bankroll management and you can learn more about how this project has evolved here: www.systematicsports.co.uk
nice, i was thinking a sport like basketball might have less variability because number of points is higher vs football where goals scored can be low and have high variability
Really liked the vid! But just one thing, bookies don't balance their books, it's common for an outcome to result in a large loss. Most errors come from autotraders/model errors!
@@LiamHartley yeah you're right I should have been a bit more explicit, I think most people believe bookies setup odds so that no matter the result of a game they come out with a profit
It’s the % difference between our calculations and the bookies odds. E.g. bookies offer 4.0 and we calculate 2.0 then 4/2 = 100% alpha. Higher alpha indicates more value in the opportunity but not that it’s more or less likely to happen
@@LiamHartley also I wondered if it would be possible for you to put the full game on the betting calculation on your site so that people know they’re looking at the right match and not data from a previous game
We've got 3,200 players across multiple leagues and seasons to allow for weekly predictions and backtesting. We input new players every transfer window/season. We've also calculated 50,000 outcomes to test our model's accuracy which I'll do a full write up on this summer. There are also 285,000 rows of bookies data and 18,000 value bets identified across all leagues this season.
@@LiamHartley Not at all complaining here, cause this tool has helped me confirm the value i find, but it doesnt seem to up update very often? Also, have you considered adding some calculations for CL?
@@LiamHartley Yeah, im refreshing. But as you can see now on the bundesliga page, last odds update was on 02/09, with games played on the 3rd and 4th. Not too important, just thought i'd let u know. :)