Surprisingly (or not), I kept betting this for about a week after the post and only had 1 losing day. Attached is a screenshot of my tracked PnL: https://imgur.com/a/gC8Pyik.
The overwhelming majority of bets were placed on BetRivers and I am still not limited up to today. The only drawback I've noticed so far is that as mentioned, it requires a bit of time. The +EV opportunities generally last for a while on an average weekday, but they disappear much faster on weekends and before the game starts -- still capturable, but you have to be actively watching the screen to bet it in time.
Just wanted to touch back on this, have you found that it is still working a month+ later? I was considering but it seems to be quite expensive, and it looks like Reddit isn't being kind to OddsJam either.
So, I kept it going for about 3 weeks total and generally had only 1 losing day per week making +EV bets almost exclusively on BetRivers with an average of ~$10 per bet.
It takes a good bit of time to continuously monitor the screener and place the bets, but the positive profit expectancy doesn't change. Considering the cost, an optimal strategy is to just go full-on for a week such that you make back the subscription cost but avoid the burnout that will come from trading this.
Looking forward to it QG. Have you seen some of the sports packages in R? I've been having fun with hoopR, fitzroy and more recently baseballr. Cleaning my data in R but still doing ML in python. Seems backwards seeing as we where always taught pandas was the choice for webscraping. But I'm finding the packages a lot easier to deal with rather than calling on API's.
And that's definitely interesting, in fact, last season when we built models, we actually used the package versions of MLB-SportsAPI and meteostat to get the historical stats and weather as opposed to scraping. So while I haven't seen the R versions, I definitely resonate with the packages being much better than figuring out the endpoints (although by learning the API, we can keep trucking on when a feature of the package inevitably gets deprecated, causes a confliction with other packages, or just stops working).
I think that the longer you go the API route, the easier it gets as you start to see that they all function about the same while each package varies in difficulty.
Would you be interested in building an automated version of this? We are looking for someone and you 100% fit the bill.
Is there a chance of being banned by the bookies? I heard stories about people getting banned after making a few smart bets.
How are things going with the testing of this concept? Is it still working as expected (or hoped for)?
Hi Eric,
Surprisingly (or not), I kept betting this for about a week after the post and only had 1 losing day. Attached is a screenshot of my tracked PnL: https://imgur.com/a/gC8Pyik.
The overwhelming majority of bets were placed on BetRivers and I am still not limited up to today. The only drawback I've noticed so far is that as mentioned, it requires a bit of time. The +EV opportunities generally last for a while on an average weekday, but they disappear much faster on weekends and before the game starts -- still capturable, but you have to be actively watching the screen to bet it in time.
Hi QG,
Just wanted to touch back on this, have you found that it is still working a month+ later? I was considering but it seems to be quite expensive, and it looks like Reddit isn't being kind to OddsJam either.
Hi Justin,
So, I kept it going for about 3 weeks total and generally had only 1 losing day per week making +EV bets almost exclusively on BetRivers with an average of ~$10 per bet.
It takes a good bit of time to continuously monitor the screener and place the bets, but the positive profit expectancy doesn't change. Considering the cost, an optimal strategy is to just go full-on for a week such that you make back the subscription cost but avoid the burnout that will come from trading this.
Looking forward to it QG. Have you seen some of the sports packages in R? I've been having fun with hoopR, fitzroy and more recently baseballr. Cleaning my data in R but still doing ML in python. Seems backwards seeing as we where always taught pandas was the choice for webscraping. But I'm finding the packages a lot easier to deal with rather than calling on API's.
Glad you enjoyed it!
And that's definitely interesting, in fact, last season when we built models, we actually used the package versions of MLB-SportsAPI and meteostat to get the historical stats and weather as opposed to scraping. So while I haven't seen the R versions, I definitely resonate with the packages being much better than figuring out the endpoints (although by learning the API, we can keep trucking on when a feature of the package inevitably gets deprecated, causes a confliction with other packages, or just stops working).
I think that the longer you go the API route, the easier it gets as you start to see that they all function about the same while each package varies in difficulty.