Sunday, July 26, 2015

MLB Baseball


With the All-Star game behind us and the second half of the MLB season underway, I will begin posting my MLB selections at this blog starting around Saturday, August 1st.

What the heck have I been doing all year, you ask? Very extensive reconstruction of the previous baseball math model. Basically, I tore it all down and started over. Took me way longer than I thought. Over the past three seasons, I have achieved modest success in this sport, made a reasonable profit all three years, but my work continuously evolves each and every year as I hone my skills in the art of statistical prediction models. Been doing this in earnest for around 13 years now.

There's little doubt the level of work I'm doing now is leaps and bounds better than anything I've done in the past. The simple reason being there is so much to learn and understand about statistical prediction models and so many different ways to do it....the #1 reason I love what I do. There's no right or wrong way, only different levels of quality. And now I've reached the "next" level.

This level started this past winter in NBA, where my newly developed basketball model was regression-fitted to show a 56.8% ATS expectancy if my predicted line was 5 points better than the actual line (51.9% for a single point of an edge over the Vegas line and a 1.223% increase for every added point of value, so just a 2-point edge would be profitable at 53.1%).

With baseball, I'm using the same approach although the differences are obvious. Money lines instead of point spreads. Different starting pitchers every day. Bullpens. Park effects, weather, etc. I have taken just about everything imaginable into account for the analysis. Being a brand new model, I will unavoidably be starting slow to feel things out, only playing the largest overlays that I can find until things settle in. By starting around August 1st, I will have two months of data (approx. 800 games) to work with and analyze in order to make it a more productive prediction model in 2016 (more selections). But this next two months will tell me a lot as I learn how much of an overlay is profitable and by how much. I've also set up my spreadsheet to be able to cross-analyze the results for the purpose of improving accuracy going forward.

I am confident of success even this year in the two-month trial period. Simply put, there are nuances in each sport that offer the diligent handicapper the opportunity to make a more accurate money line in baseball than Vegas does because Vegas' lines are biased or influenced by the betting action, either perceived or actual. My line is a true reflection of the two teams in question with no outside influences. So I simply compare my "truth" to their "offering". I'm looking for a 5-10% ROI range or better.

In other sports, I will be entering the LVH Westgate NFL SuperContest for the second time this year. I finished 78th out of 1,403 last year in my inaugural attempt, hitting 60% (51-34-0 ats), normally a good enough performance to cash many of the previous contests but I didn't win a dime in the largest field in contest history and I even had some negative luck going 1-5 in legitimate good/bad beats. It may not be this year, but I will make a top 5 run in that contest at some point. Wish I played it the two years prior when I hit 62% and 70% ats respectively in 2012 and 2013.

Anyway, those contest picks will be posted here as well as all of my NFL selections as I'll be opening up my repertoire to deliver significantly more plays than I usually do (70-90 vs 45-55 normally). Less than 7 weeks till kickoff.