One thing I touched on in the last post concerning the results was the performance of the DNB bets. To recap for anyone who's not been reading this summer, I've spent a lot of time building a 3rd rating algorithm over the summer that only looks at whether backing a team is value for DNB purposes or not. This was an idea I had after watching so many of the systems struggle in the second half of the season last season due to an above average number of draws on all of the systems.
These new bets will be trialled on the blog and in the SBC forum this season. This will be systems 31-33. Due to the fact these are unproven and untested in a live environment, my advice has always been to ignore these new systems and the new rating algorithm as we should all concentrate on the established systems from last season.
However, after tracking the system results so far, I think I should point out something regarding DNB betting on the established systems. Of course, it's only based on a very small sample and I don't have the time to invest any more time in understanding the system results or building new systems as I'm gearing up for the first system bets next weekend. I'd feel like I'm not doing my job if I don't point this out though as some of you may be following DNB next season.
What I've noticed so far this season is that the results for DNB are better on the established systems when the DNB rating algorithm agrees that there is DNB value apparent in the games.
If we take Saturday for example, there were 21 value bets thrown up by my ratings algorithms. However, only 12 of these games were classed as DNB value. The 9 games where there was no DNB value resulted in 6 losses and 3 draws.
If we look over the games so far, for each set of games, backing the DNB value games only for the established systems has beaten the DNB returns for all the games. Is this too surprising?
Well, when you think about it logically, it's probably common sense. Backing DNB on games where the DNB rating algorithm thinks there is value in the DNB bet has to be a better idea than backing every game DNB thrown up by any rating algorithm.
Below are the results during the trial to allow us to make a comparison:
As you can see, there are 27 bets that have been thrown up so far on the systems that wouldn't have been value on the DNB algorithm. These must have amounted to a loss of 9pts on DNB betting (42pts total if you exclude these against 33pts total if you include these).
Importantly though, as well as increasing the profit, it also increases the ROI. So far, backing DNB on these DNB value games has actually ensured a higher profit on the systems than backing the teams to win. I don't expect this to remain during the season (or I hope not!) but it is interesting I think.
As I've said above, I don't want to go into great levels of detail or analysis based on 3 sets of games and less than 150 system bets so far this season. However, I actually have a piece of analysis already done to help us out here.
One of the new systems for next season is system 6-21-31. This is all the bets on system 6-21 where the DNB rating algorithm (31) thinks there is value too.
Last season, System 6-21 had a profit of 29pts (9.8%) from backing DNB. System 6-21-31 had a profit of 23pts (18%). I think this shows what I'm getting at here. Clearly, backing DNB on system 21 isn't a bad system in itself but cross referring it with the DNB rating algorithm nearly doubles your ROI. If I go back another couple of years, it increases the ROI by around a third overall.
Now, I don't have the time or inclination to look at every possible system for H/A returns and DNB returns as there would be thousands of combinations but I think the above is something to consider if seriously thinking about following DNB this season. Of course, you could break this down and look at homes v aways and price ranges and so on but that will have to wait until next summer!
Good luck for the season ahead. 10 days to the first system bets this season....
I'll be back with any weekend bets later this week.
No comments:
Post a Comment