Weekend Football Review (10/31/11)

By Paul Bessire

Monday, October 31 at 9:15 PM ET
It is very important to us to be transparent and honest about our picks. I always try to recap each football weekend. At this point, through Sunday, Paul's Picks, which include the top three ATS plays Saturday and Sunday as well as the top weekday college against-the-spread (ATS) plays and the Monday Night play, are 37-28-2 ATS (57%) to start the season. The ATS Top Plays of the Day, the strongest opinion ATS each day overall in football are 22-20 ATS (52%) and the "Locks of the Week" are 8-8 ATS (50%). Including last season, this brings our all-time record during football to 134-93 ATS (59%) for Paul's Picks, 95-56 ATS (63%) in ATS Top Plays of the Day and 36-20 ATS (64%) for Locks of the week.

As a reminder, at midnight ET each day, we make all of our previous day's subscriber content available for free for registered users. Performance is also tracked in the TrendFinder Database (updated every morning from the previous day). We're never going to hide anything. So even though we have to swap out articles in the archive to focus on new ones, articles never go away. Just make sure to use the correct week and date in the URL - or ask us for the link...

The Football Numbers/TrendFinder Database:
All football numbers from the 2011 season are now available in the TrendFinder Database. We encourage you to review this new application (which now includes Home and Away Conferences). Since I/we have spent a significant portion of the last two days looking through the TrendFinder, researching actual outcomes for trends with significant sample sizes and deciphering tweaks that may be warranted, this blog will focus on that exercise.

While those numbers above do not suggest a strong week - and I would struggle to say that it was a great week - had I run either Saturday or Sunday one more time, the impression may be completely different. The "Locks of the Week" for College Football and the NFL both lost, while picks behind them that were separated by one or two tenths of a percent, won. Meanwhile, the Paul's Picks are currently above .500 (4-3) for the week. And, most importantly, we maintained some great (and not so great) streaks. We have opened the season with nine consecutive weeks in college football in which the "normal" or better Over/Under picks have performed at .500 or better, highlighted by a "green" 2X pick on the OVER (61.5) in this week's Missouri @ Texas A&M game (that Mizzou won 38-31). On the season, normal+ O/U in college are 45-28 (58%). Similarly, we have opened the season with nine consecutive weeks in the NFL in which all playable ATS picks have performed at .500 or better. On the NFL season, playable ATS picks are 62-37 (63%).

Externally, the most important performance attribute in this business, at least it relates to perception and user experience, is consistency. The worst thing that we could be is 47%-53% accurate on everything with wild swings in performance from week-to-week. Whether picks consistently perform well above .500 or below it, users can draw conclusions and enact strategies based on that consistency. Fortunately, we have exhibited, over a season and a half in existence, that picks are consistently strong in NFL ATS, and, within the scope of the last year, we have shown similar success with college football O/U, MLB plays and college basketball plays (especially at levels of stronger confidence). Unfortunately, for this particularly unique season at least, we have consistently performed poorly with college football ATS picks and NFL O/U (not to the same degree as with the good things, but it is close).

To some of you, that may sound like a great fade opportunity (and I know that I essentially said as much to open the last paragraph), but, that's NOT how we ever intend to approach this business. The most important attribute to our site is that our confidence reflects our performance. We have had success in the past with college football ATS (four consecutive winning seasons - out of four attempts - before this one) and with NFL O/U (it was a good week with the "normal" picks this week, but NFL O/U rarely presents much value like that either way) and our intent is to return to those levels of success, to do so with consistency and to do so without directly (negatively) impacting areas in which we perform strongly.

As we delve deeper into the TrendFinder Database and look closer at our losses, there are some actionable takeaways that should help us improve, especially with college football ATS performance. Whether I like to admit it or not, and, while it could mean that some level of (tempered) subjectivity enters the picks each week, particularly with 18-22 year olds, it appears as though there are some factors that we can value stronger and some that we may not be factoring that we can/should consider. For instance, if I put myself in Kirk Cousins' shoes, going through three consecutive big games, throwing a hail mary to beat an undefeated team and making the media rounds throughout the week, I would probably have a hard time playing at the same levels I have played at before during a road game against a quality opponent. I can't say that I gave that specific analogy much thought until about 3:00 pm ET on Saturday, but it is honest. There are not many purely objective ways to factor that in, yet there are some additional factors that we can input and, there are ways to layer in some elements of subjectivity in unique circumstances - not to change the pick (Nebraska allowed Keith Price to have 274 yards passing and four TDs, while Chris Polk had 130 yards rushing), but to minimize the damage of a loss from a game where external factors beyond the Predictalator may come into play.

Let's look at some numbers from the TrendFinder... We have talked in the past about how we can trust teams from certain conferences more than others. Here is our (playable) ATS and O/U performance in conference games this season (most conferences have around 30 conference games included):

  • ACC: 23% ATS, 40% O/U
  • Big 12: 33% ATS, 71% O/U
  • Big East: 29% ATS, 50% O/U
  • Big Ten: 44% ATS, 60% O/U
  • C-USA: 50% ATS, 50% O/U
  • MAC: 59% ATS, 33% O/U
  • Mountain West: 50% ATS, 63% O/U
  • PAC 12: 32% ATS, 63% O/U
  • SEC: 70% ATS, 67% O/U
  • Sun Belt: 33% ATS, 44% O/U
  • WAC: 63% ATS, 88% O/U

It should not be surprising to those that follow both sports that SEC football ATS performance looks more like the NFL than the rest of college football. Furthermore, we may have this "consistency" theory a little skewed. While the sample sizes above are not all overly significant, the ACC and Big East have historically been the conferences for which this technology has struggled the most. As best as I can rationalize it, conferences for which there is the greatest talent discrepancy between the best and worst player in the starting lineups tend to have the weakest performance. Theoretically, this should should still result in greater than 50% success, yet that theory supports the trends that we have seen over multiple years of testing this approach where SEC and WAC picks have consistently performed better than other picks. Needless to say, we have reevaluated our modifiers for picks by team/conference to account for this phenomenon.

Looking at the conference analysis, there appears to be a direct correlation between ATS performance and O/U performance. Looking at this in greater depth, in the games where we have had at least 57% confidence in the Over/Under pick (and the spread has been less than our typical college cutoff of 20.5 points), we are 35-33 ATs. That's not necessarily off the charts, but far better than our ATS performance in other college games. In other words, when we have a strong opinion O/U and ATS, we have performed well (or at least better than otherwise) in both. This reflects what we see relative to conferences as well. This is very valuable for us (and you) to know as we evaluate the games for which we are most confident in general.

See anything else that stands out in the TrendFinder? Let us know by contacting us.

College Football FREE Week:
With the enormity of this upcoming college football weekend - LSU @ Alabama, South Carolina @ Arkansas, Oklahoma State @ Kansas State, Texas A&M @ Oklahoma, Oregon @ Washington, Texas Tech @ Texas, etc. - with the new features we have launched recently - TrendFinder, Play Analyzer, Parlay Calculator in Customizable Predictalator - with the strength of our Over/Under college picks and with the efforts to improve recent college football ATS performance, everyone will receive this weeks college football information - ATS, SU, O/U picks with Play Analyzer and Customizable Predictalator access - for every game entirely for FREE. Check the site after 8:00 pm ET on Wednesday to get everything for the week.

College Basketball Trial
As this is PredictionMachine.com's first full college basketball season providing straight-up, against-the-spread and over/under picks for every DI game (with a rotation number), we will be offering a three week free trial. Picks will be available to all from November 7 - November 27. The full season picks package is available for the rest of the regular season (including conference tournaments) from November 28, 2011 - March 11, 2012. Picks will also be sold on a monthly basis. Daily picks are posted at 4:00 pm ET each day.

The (Final) Baseball Numbers:
The MLB Playoffs went very well. As seen previously with other postseason tournaments, there is something to be said for clear motivation, a full season of data and trustworthy lineup and pitching staff information. Below are final win/loss stats from MLB Playoff picks using the published Play Value Calculator recommendations for a $50 bettor.

  • All Playable Games: +$653 (45-35, 56%)
  • Normal+ Plays: +$549 (12-2, 86%)
  • "Half-Bet" Plays (when PVC recommendation is greater than $25): +$696 (28-17, 62%)

And since May, even including September, all playable MLB plays are +$2973 for a normal +$50 player.

As usual, if you have any of your own comments about this article or suggestions about how to improve the site, please do not hesitate to contact us at any time. We respond to every support contact as quickly as we can (usually within a few hours) and are very amenable to suggestions. I firmly believe that open communication with our customers and user feedback is the best way for us to grow and provide the types of products that will maximize the experience for all. Thank you in advance for your suggestions, comments and questions.