Tuesday we had two 1 unit plays and went 2-0. Both were in the NBA with Orlando +12.5 winning the game outright 95-83 and Detroit +3.5 also winning outright 112-89.
I want to apologize before getting into today’s musings. I know everyone is looking for last second shots for the wins. Where everything is in the balance the last 30 seconds. So it is bound to be disappointing when a 12.5 dog wins by 12 and a 3.5 dog wins by 24. I will try to pick less solid plays to get things dicier:)
So far today I have sent out no plays.
An interesting email discussion to share:
“Concerning models, do you create your own models, or do you look at models that are publicly available? I know of a few publicly available models that I believe to be quite good like kenpom for college hoops, football outsiders for football, and fangraphs for baseball. Are these the types of models that you might consider looking at? Any info on the subject
would be quite helpful. Thanks!”
My response:
“Those are fine if you do not have your own.
I used to use kenpom back when I was betting totals on pinnacle. But those lines were off by 20 pts on occasion so I didn’t need anything more accurate than kenpom.
Football outsiders is also ok as long as you know the formula for converting to a spread and us the right home field adjustment.
For baseball, any site like Fangraphs will do.
The key is that models should be used only as an exclusionary filter. Not for predicting outcomes. Even then you can make an argument that models like this are useless.
It’s a very complex topic.
I use my own models, however.”
Followup:
“Thank you for the response. I would like to start creating my own models, but I am not really sure where to start. I do understand that the models are just one factor in the big picture and should be used only as a filter. I was hoping that just using some of the public ones would be sufficient as a filter, but I guess this sports betting thing is a lot more difficult than I had hoped. When you go about creating your own models, is the process similar to, for example, kenpom? Do your models attempt to come up with a predicted score for each game based on some advanced team stats and regressions, or are your models doing something else entirely? ”
My Response:
“If I were starting out and know what I know now I would not spend a lot of time and energy on models. Models are the most time consuming and have the least rewards.
There was a time when I felt models were a complete waste of time and discarded them completely. But my results suffered so I went back to using them.
Go to Thepredictiontracker.com
They have models along with data going back 10 years or more. Put them in a database or spreadsheet.
Then run an analysis on which are more predictive .Do not take the web sites numbers for how they did as a good result. The website does not differentiate between a prediction that is .5 pts different compared to 1 that is 10 pts different than the line.
So you need to determine at what point each model becomes predictive over 52% winners. It might be a 5 pt difference or maybe a 10 pt difference from the spread or maybe not at all. But I am sure you will find one or two models that work out.
If you want to get more accurate on this you need to break it down between home, away, neutral. And then again favorite or dog.
Depending on the sport you will find that Home dogs in many instances are not as rewarding as road dogs. It’s rare to find it the other way around but again there are certain circumstances where an exception exists.
Now if you do all of that and run the proper analysis you will be able to circumvent the very difficult task of starting a model from scratch.
If your intent on modeling get back to me and I can give you a few more tips. But remember modeling while many think it’s the holy grail for handicapping its far from that. Only a filter if you know how to use them.
I have thought from time to time do that myself however I have my routine in place that I use and I spend my time on other areas of handicapping that tend to bring better results if understood and applied correctly.
”
If anyone has any questions on this topic feel free to send me an email. I enjoy discussing the various handicapping techniques that bettors use. And most discussions can be a great learning experience for all.
RickJ
RickJ’s Handicapping Picks
