[JayDee]

Hi, everyone!

I joined just a few days back and have been reading the archives and lurking around for a while. I've been a part-time sports bettor (depending on work load, current level of interest, and so on) for about five years and have been producing profits consistently. Not big profits, but still profits.

My dream is to be able to live from sports betting and investing sometime in the future. I'm 29 today, so I'm thinkning it won't be by the time I turn 35. It is a hobby for me but one I take seriously. I'm sincerely glad I found this forum and I hope to be able to participate and every once in a while contribute to the discussions here.

My betting approach is systematic / statistical and I have one main model I work with and I'm trying to come up with a few more (with a 106 % ROI, I need to make a lot more bets for the betting to produce money even remotely close to my current salary). Since I live in Sweden it's natural for me to focus my betting on soccer (English, Italian and Spanish league) and hockey (NHL). (PS. I use a flatbet 2 % of my bankroll and need to take a serious look into Kelly when I have time I guess. I've seen that there are lots of good posts about it so I'll make an evening out of Kelly soon).

Since I've started lurking around here I've put quite a few hours into Python and have now been able to write a program to backtest my model as well been able to actually produce a small new little system.

I'm having trouble evaluating my backtesting (I do work with different sample sizes to ensure I'm not over-optimizing) and I would like to ask you guys a few questions. If anyone took the time to answer, it would be greatly appreciated.

1) Do you believe in "robust" systems that perform good results in several markets (for example my soccer system produces decent results in tre big european leagues (England, Span, Italy, but not in all european leagues (Bundesliga, Eredevise, Ligue 1)) or do you work more with "fitted" systems for specific sports/leagues, say an over/under system specifically for the French Ligue 1?

2) When evaluating a system, what parameters do you use and what values are you looking for? A back testing of my soccer systems produces a Z-score of 1,46 (0,07% beat me ) over approximately 900 bets over 2,5 years with an ROI of 115 %. Is that a system I should continue with? Don't really know what I'm looking for here....

I'm glad I found your site and I'm looking forward to a lot of interesting discussions in the future.

/JayDee

[Justin7]

1. Yes. A good model will work well in different leagues, but needs to be customized to each league.
2. Anything over +2 is of interest.

Soccer, like MLB, is very player dependent. I'm not sure you can model soccer successfully without using player injury information.