This marks my twentieth season
on the forum. As always, I’m eager to be
back in the saddle after a long off-season.
For those not familiar with my handicapping brand, I’m the resident
geek; a programmer (number-cruncher).
While I welcome your visits, you will invariably find this forum locale low
on flamboyance (dull). This post will
demonstrate that point.
A slave to data/history, I have
utilized a self-created algorithm built entirely upon what has come before (a 1976-2018
database). While my algorithm had always
been a compilation of generic situation-trends, last year I took an exit ramp and
established a system based on Power Ratings (PR). A new paradigm for the Human Administrator…..
While the old method performed
well enough for me in an early timeframe (1994-2011), later seasons (2012-17)
were unsatisfactory. Although,
performance at the very top was viable. All of that aside, how did the new 2018 method
The system was volume prolific,
generating between 10 and 29 picks per week; a total of 220 best-bet selections. The lesser of those (the lower echelon) did
not win. However, the entirety of it was
shared/posted in order to observe any system Rank Value (R-V).
In 2018, the top twelve
selections had a success rate of .583 (88-63-2). So, the system produced strong volume along
with R-V in the broader context. The HA
savored that profitable top-outcome because it plays to his longtime tech philosophy
that – Any genuinely “worthwhile” system will reveal itself in strong volume,
while small-volume is inherently more volatile and may not always demonstrate
Switching to the PR model
was quite a transition for me in 2018. I
learned a great deal last year and then rolled that experience into some affirmative
research during the off-season. New but improving
systems commonly show their steepest gain in year #2 and I’m hoping that will
be the case.
So, what will the 2019 algorithm
look like? For starters, as an entrenched
techie I will still progressively input season data in the search for subsequent
successful output. My computer won’t
know injuries, team momentum, NCAA suspensions or coaching changes. It’s purely a rote vehicle.
The system as a whole will produce
category diversity, but the historical model shows a bit less confidence in the
home teams. As a result, you will observe
some reduced volume there. Each best-bet
selection will be assigned wager units having magnitude 4 thru 6. Those values may seem less than intuitive,
but they proportionally reflect my confidence.
You will also see a distinctive
tilt toward lesser-team match-ups. That
is, I will be more likely to embrace a game like Rice/NewMexSt, than a high-profile
contest like LSU/Georgia. That’s where my
profound dullness rears its ugly head, as I know most of the forum is
comprised of fundamental cappers attuned to those (marvelous) TV games. Square peg – round hole…….been there.
So, will this year’s model
bring value in 2019? Maybe. Win or lose, I am forever learning the craft
and the annual challenge never fails to deliver an education. Along the way, your input is most welcome.
I will be full-launch the
week of August 31, but may also be active on the earlier Arizona/Hawaii game. The system will start with low volume and build
as the season progresses. In the end I
hope to make a positive contribution to the forum, and likewise wish you all
well in the challenging new season ahead.