11-0 And Season Recap (2/8/11)

By Paul Bessire

To commemorate the Predictalator's astounding NFL postseason run, we have decided to offer the 2011 NFL and the 2011 NFL and college football combined packages at 2010 prices. For one week only - through next Sunday, February 13th - you can order the upcoming 2011 full season football packages at a significant discount. Click here to lock-in on 2011 football for 2010 prices.

Tuesday, February 8 at 5:12 PM ET
As I have throughout the season, I will recap the weekend in football each Monday afternoon (whoops it's Tuesday already - 11-0 celebration spilled over a day or so - and, yes, I used the same line last time). It is very important to us to be transparent and honest about our picks. We never remove picks from the site, just from the archive to keep it from becoming unruly. Every link follows the same pattern, if you need past links and can't figure it out, let us know... And right off the top, congratulations to the Green Bay Packers! The team for which I grew up cheering and in which my mother actually owns stock (she wants me to be clearer about the fact that the stock is in her name) just won the Super Bowl. Even without the professional satisfaction that gives me (and that I will ramble about for the rest of the blog), I am happy for team, the fans, the owners and pretty much the entire state of Wisconsin.

We did it. In our very first full year in existence, we successfully completed an undefeated NFL Playoff run (against-the-spread). It's a dream/milestone of any prognosticator to achieve such a feat and we are lucky - and good (but probably in that order) - enough to already be able to check that accomplishment off the bucket list. In fact, dating back to Super Bowl XLIV, PredictionMachine.com has yet lose an NFL playoff game ATS. And Paul is on a personal roll with his published Super Bowl picks reaching 8-0 ATS.

Before we get to what 11-0 means and the greater debate/conversation that it invokes, we want to provide the proof. We totally understand how unbelievable the comments above sound and get that people have a general right to be skeptical of anything, especially based on what others in this industry have done in the past. What follows are the links to the Predictalator picks and analysis for each NFL round - complete with projected boxscores and individual stats. As the 50+ radio stations that talked to us throughout the playoffs or members of the media such as CNBC's Darren Rovell and ESPN.com's Chad Millman could tell you, these picks have been unaltered.

NFL Playoffs Predictalated (original projections for each game):

Wild Card 1/8 Picks - NO 26 @ SEA 18 (+10.5) and IND 24 - NYJ 23 (+2.5)

Wild Card 1/9 Picks - GB 23 (+2.5) @ PHI 21 and BAL 25 (-2.5) @ KC 19

Division 1/15 Picks - GB 23 (+2.5) @ ATL 20 and PIT 21 (-3) vs. BAL 16

Division 1/16 Picks - NE 26 vs. NYJ 21 (+8.5) and CHI 25 (-10.5) vs. SEA 13

Conference 1/23 Picks - GB 22 (-3.5) @ CHI 17 and PIT 21 (-3.5) vs. NYJ 17

Super Bowl 2/6 Picks - Green Bay 22 (-2.5) vs. Pittsburgh 19

So now that you know that we did it, what does 11-0 ATS in the NFL Playoffs mean? For one, it means that if someone started with $50 and was able to double-up 11 times (wagering in between each game), he/she would have turned that $50 into $102,400. Obviously, that's not the most realistic approach to wagering and one we would actually argue against, but it helps to put into context just how valuable our pick information was/is/can be. Technically, that's a 204,700% return on investment. You don't see those very often. And, while we have not actually met anyone who turned $50 into more than $100k using our information, we have been fortunate to hear of many stories that included multiple successful four-team parlays at 10:1 odds going into each of the first two NFL Playoff rounds - followed by success in the final two weeks of the postseason.

It also means that we got lucky. While we do expect to get each pick individually right at the time, we don't expect to get every pick right. We saw it at the time with our over/under total and straight-up (we did not foresee the Seahawks defeating the Saints or the Jets defeating the Patriots as more likely than not) projections and have seen it since ATS with just one night of basketball. Consider a no-hitter in baseball. Even though an average pitcher has about a 74% chance of getting an average hitter out, that pitcher has about a 1-in-3,500 chance of throwing a no hitter. Even the best pitcher in history facing the worst lineup in history would only be about 1% likely to throw a no-hitter, meaning it probably would not even happen twice if we saw an entire season of just that matchup. The elite pitcher makes throwing a no-hitter about 35 times more likely than an average pitcher, but it's still not very likely to occur. In-fact, given our own projections, we would have only calculated our likelihood of going undefeated through the NFL Playoffs at about 1-in-115 (the math is not a straight-forward as most think there, contact me if you want to see it). So while we are extremely happy with and will promote heavily our performance, we easily could have made it through the entire existence of this site - generating equal confidence each NFL playoffs - without ever going undefeated against-the-spread in all 11 games.

But it also means that we are good. We pitched a no-hitter the way an elite pitcher would against a bad lineup. Just throwing darts (totally randomly) at NFL playoff teams to pick each game would give someone a 1-in-2,048 chance of going undefeated in the NFL Playoffs. With the confidence that we generate in our own picks from simulating the game 50,000 times, we provided 18 times better odds to do what we ultimately did than randomly choosing winners. We pitched a no-hitter - the easiest way possible. If the books (or the public, or just a random prognosticator) are Randy Wolf (a totally average pitcher), we are Nolan Ryan.

We have found a way to make manipulate information in a way that can be valuable, putting us in a decidedly advantageous position against our competition (which would be some combination of the books, public and other prognosticators). There is a generally held belief that markets efficiently represent all information available, making them mostly random, where anyone can benefit in the short-term, but where long-term success is dictated by the overall market itself. In sports, this would suggest that people merely use "information" to justify their desire to place wagers. Some will win and some will lose, but the books, since they take a percentage of each loss, will be the only long-term winners. To a much greater degree than you probably would have guessed, I agree with that - more so in other markets, but even as it applies to sports. If team X has been winning 70% of its games against-the-spread over a significant time period and if that is relevant to the current version of team X, the public will adjust, which will make the books/lines adjust and an efficient market is formed. This is why I cannot find value in a database of information from which I can determine that X% of all teams cover the spread when coming off bye weeks and playing at home against teams from the Pacific Time Zone, who are above .500 on the season. Even if the notion is to be believed that that is relevant information to this game, if it is, and anyone can just look it up, it should be reflected in the price. To me, that's true "information" and it can only lead to a more efficient market, not long-term gains.

However, what we do is very different. Having an 18 times better chance at going undefeated ATS in the NFL playoffs than most - and then backing that up by doing it - is not just luck. There is something we are clearly doing better than the market can. Sure, we have robust databases of information on teams, players, weather, coaches, stadiums, etc., but nothing we do with it can be looked up. I feel as though we have found happy medium between over-valuing ultimately irrelevant information and over-sophisticating simple problems. There are no magic stats that we can point to. I just need to know how likely anything is within the course of a simulated play and then, if it happens in that instance (which then sets up new situations and decisions to make). It's very straight- forward. We see games played in real-life all the time. But the methodology is still impossible for anyone to handle. It is the technology that manipulates the information that sets us apart. We strength-of-schedule-adjust every attribute to get an unbiased view of everything. We focus on the actual individual players that are going to be on the field and how often they will be utilized. We layer in everything else. Then, we let everything interact. Simulating one football game once usually generates thousands of decisions to make. Playing one game 50,000 times creates millions of decisions (made in the form of truly random numbers). Without the assistance of some form of simulation technology, it would take more than a lifetime to do that.

I would contend that this isn't information; it's the best possible way to project what's most likely to happen - and just how likely that (and everything else) is. And I feel that, for the most part, we have proven that. Thankfully, what is most likely to happen seems to be happening - and at or better than the rate we expect. We are proving, that, yes, luck is in the short-term, but detailed and rational analysis will be what can win/will win in the long-run.

I have heard many people say that they would not pay for "picks," citing that it's not worth the investment to have someone tell you who to play, when your intuition (and luck) could do about as well. I agree. That's not how I would describe our service (though I fully understand the word "picks" is everywhere - that's more for simple recognition on the potential customer side than it is emulation of others in the industry). We do the work. Our "Predictalator" assesses just how likely all outcomes are and suggests/recommends ways in which that analysis can be financially beneficial. Some people probably couple this with their own analysis. Many may even use us to provide piece of mind, justifying their own beliefs. That's totally acceptable. But, assuming we can do what we set out to do - provide accurate probabilities of success for each game - the site should be able to answer all of your questions. Education is important. Trust me, as the son of two teachers, it's very important to me to be thorough and helpful in aiding those that may not otherwise make wise decisions. Of course, I use my own findings for personal gain, but I find interacting with and helping people just as fulfilling.

What can't we do? Well, clearly are numbers have been stronger during one-and-done playoff formats (NFL Playoffs, NCAA Tournament) than at any other time (though still strong otherwise). As we have discussed at length throughout this blog, there are reasons for that that are grounded in our assumptions. We assume that every player is playing his hardest to win every game and that all players in the game are playing at the same levels we expect based on their past performances. Unfortunately, that's not always as valid as otherwise. As soon as players start telling us at exactly what percent of their normal abilities they are playing, we'll have just about everything figured out. Another assumption that we make is that everyone's career is projectable. In other words, it should be perfectly acceptable for new players to a league or veteran players joining new teams to just be plugged in for their new roles. As long as we have any information on them at all, we can come up with some projection of how they will fit in and ultimately how team and individual performance is impacted. While that is definitely true to a degree, the accuracy of those projections and assumptions improves the further we get into a full season of that player in that situation. And lastly, it's not as relevant with the NFL Playoffs, but removing concerns over homefield advantage helps with college basketball picks. We see verifiable proof that homefield advantage matters (in order of significance in: college basketball, college football, NFL and NBA), but it is hard to say exactly why, which consequently makes it very difficult to pinpoint exactly to what degree and how it impacts outcomes (especially venue-by-venue).

Essentially what I am getting at is part of an underlying debate and even a movement that we see happening across the world (literally) and in just about every industry (if it's not in your industry, you should lead that charge). The increased capabilities of technology allow us to handle and manipulate information (data) in a way that leads to optimal decision-making that can be very advantageous against a market that is not as advanced in general. If we have learned anything from this economic situation, it is that the long-term matters and that efficient choices are the right ones. It's not really about doing more with less or accumulating as much data as possible just to do it; it's about accounting for every possible factor that can go into a decision to determine likely outcomes and preparing for them accordingly. Simulation is the best approach to answering major questions because it can best incorporate elements of optimization, forecasting, regression, classification and characterization ("analytics") in a relatively quick and understandable manner. I choose to apply this to sports now because it's more fun and, since every fan thinks he/she is an expert for some reason, the market is and may always be exploitable. But I have, others have and it could be applied to almost all elements of business where any form of a market or a inefficient system could exist.

And what happens if everyone utilizes this approach and/or gains access to this information. First of all, we are very long way from that happening. Also, by definition, "optimality" is essentially impossible. And, no one will ever be able to figure everything out (in any industry). There will always be tweaks to make; we just happen to have a big head start on most in this field. But even if somehow we (or someone else) figure's everything out and everyone has 100% trust in our (or someone else's) analysis, the value in that information will only increase, not decrease - to think otherwise is to be short-sighted with greed and too focused on existing markets. I can think of worse things than being able to predict the future...

Sports Prognosticator Bucket-list:
After that way-too-deep rambling, I would love to gain some opinions on what has become an interesting topic around the office. If it is one of the top goals of anyone who predicts/picks sports to accurately pick every NFL Playoff game against-the-spread in the same season, what else is on the Sports Prognosticator's Bucket List? We would love to hear your thoughts. Contact us with your list so we can add some more goals to ours (though our main goal will always be to provide insight to you that will benefit you financially).

Below are our final numbers for the 2010-11 season. Keep in mind that anything over 52.38% (when/if getting -110 odds from a book and wagering the same amount on each pick) accurate is profitable.

The Football Numbers (the 2010-11 Football Season recap):

  • ATS Locks of the Week: 28-12-1 (70%)
  • Daily Top ATS Plays: 73-36-1 (67%)
  • Paul's Picks ATS: 98-63 (61%)
  • ATS - All Playable Games: 56%
  • O/U - All Playable Games: 55%
  • SU (NFL and FBS vs. FBS College): 73%
  • All-Time ATS NFL Playoffs: 12-0 (100%)


Super Bowl Props:
Since this will be my last weekly football blog (and probably last major blog entry until March), I want to make sure we don't ignore the Super Bowl props. In total, we hit 54.4% of our playable props. While that should be profitable in its own right, we did get more value with our better plays. Props that were evaluated at 60%+ went 9-6 ATS (60%). The top value play that we hit was that Ben Roethlisberger would throw an interception before a touchdown. We had that happening about 50% of the time, yet the payout was +160, which gave it tremendous value (almost 3X units - the second best play on our board). Other value was found with the Shaun Suisham missing a field goal (+300), the Packers finishing with more penalty yards than the Steelers (+130 paid off when Packers had 12 more penalty yards than Steelers) and with the UNDER on Packers' back-ups like John Kuhn and Brandon Jackson. We also went the opposite of many sharps, strongly recommending UNDER on Aaron Rodgers' rushing yards (20.5 was the line and he ran for -2), while going OVER on his completions (23 was line, we had 23.7 and he completed 24) and attempts (33.5 was line, we had 37.4 and he threw 39 times). The biggest losses occurred for us due to Emmanuel Sanders' injury for the Steelers. Forcing Antwaan Randle El on the field put our prop in play, but then he stayed on the field the rest of the game with Sanders out and caught two passes (the line was 1) for 50 yards (the line was 15.5 - and we had the UNDER strongly on both). Numbers for Hines Ward also benefited greatly from Sanders' injury. Brett Favre was never mentioned and Jerry Jones (himself) was shown only once on-air during the game. And lastly, Greg Jennings paid off by scoring a touchdown (+105), yet even his two TDs could not earn him the MVP. Had Jennings been thrown the almost sure-fire 70-yard TD down the seam that James Jones dropped, he likely would have taken in the distance and may have been MVP. Of course, Aaron Rodgers numbers would have been even more amazing then.

To commemorate the Predictalator's astounding NFL postseason run, we have decided to offer the 2011 NFL and the 2011 NFL and college football combined packages at 2010 prices. For one week only - through next Sunday, February 13th - you can order the upcoming 2011 full season football packages at a significant discount. Click here to lock-in on 2011 football for 2010 prices.

Beginning yesterday, PredictionMachine.com is also offering a TWO week complimentary trial of its daily NBA and College Regular Season basketball products. Check in every day to get the Predictalator's straight-up, against-the-spread and over/under picks with confidence for every game (college includes all games with rotation numbers) for free through Sunday, February 20th. You can also order the NBA and/or College Basketball regular season packages now. To know exactly when picks are posted, sign up for basketball pick availability alerts.

As usual, if you have any of your own 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.