UPDATE NOTICE: SPRS (Sport Pundit Rating System) Name Change to YPPSYS (tm) viz. YPP-SYS (tm)
We have discovered that confusion might be remotely possible between our SPRS college football prediction system and something called Thompson SPRS which we noted for the first time today (October 12, 2008) at The Prediction Tracker summaries but which we have been unable to find online. When we first named our SPRS system, there was no Google conflict, otherwise we would have chosen a different name.
In order to identify our system as a separate system, and since we have no reason to be bound to our original name if confusion of any kind is possible, we are changing the SPRS name today from SPRS System to YPPSYS (or YPP-SYS ), which means "Yards Per Play System", a change which we will implement as an update paragraph on the relevant older pages explaining our original SPRS. There is no online conflict for this name and we hereby trademark the names YPPSYS viz. YPP-SYS as acronyms used by us for our college football prediction system, a rating, ranking and prediction system which was previously called SPRS by us, based on net average yards per play advantage (NAYPPA).
We do this particularly since our system this year is beating nearly all the systems listed at The Prediction Tracker and we want no confusion.
the SportPundit Football Ranking System (SPRS System) - see the final college football rating for the 2007/2008 season
The SportPundit Football Ranking System (SPRS System) is based on a power value calculated to rate the relative DOMINANCE of a team as determined primarily by
the net average yards per play advantage (NAYPPA) of offense over defense.
The SPRS System dominance rating of a team is calculated using 4 parameters:
1. The net average yards per play advantage (NAYPPA) of a team's total offense vs. that same team's own total defense:
E.g. West Virginia averaged 6.5 yards per play on total offense and 4.4 yards per play total defense in the 2007/2008 college football season, giving the Mountaineers a +2.1 net average yards per play advantage (NAYPPA). We obtain the total offense and total defense figures from the individual team websites, from the conference websites, from the NCAA and NAIA websites and from cfbstats.com.
2. A prorated (proportional) adjustment for schedule difficulty
Schedule difficulty affects total offense and total defense stats (we use the schedule rank from the Massey Ratings rather than from Sagarin as our independent source of schedule difficulty since Massey includes all college football divisions whereas Sagarin includes only FBS and FCS teams, which skews his ratings from reality.
But Massey's ratings are also not optimal.
Schedule difficulty is not linear. The adjustment for schedule difficulty made by our SPRS System is one-tenth of one point (.1) for each 10 places of schedule rank by Massey up to 50 - always rounded upwards - as follows:
For a schedule difficulty rank from 1 to 10, a value of .1 is deducted from the NAYPPA.
For a schedule difficulty rank from 11 to 20, a value of .2 is deducted from the NAYPPA.
For a schedule difficulty rank from 21 to 30, a value of .3 is deducted from the NAYPPA.
For a schedule difficulty rank from 31 to 40, a value of .4 is deducted from the NAYPPA.
For a schedule difficulty rank from 41 to 50, a value of .5 is deducted from the NAYPPA.
For example, West Virginia's schedule difficulty is ranked by Massey at 42, which is 5 x 10 places of schedule rank (4 x 10 +2 rounded upward, i.e. 5 x 10) so that .5 points are deducted from the NAYPPA in the case of West Virginia (+2.1 minus 0.5 = +1.6).
For a schedule difficulty rank from 51 to 96, the multiple is increased from 1 to 1.5. Again, this is because schedule rank is not constantly linear.
For example, Massey gives Kansas a schedule rating of 59, which in our system would ordinarily mean a deduction of .6 from the NAYPPA. Here, however, 59 is multiplied by 1.5 giving a result of 88.5 which is rounded up to 90.0 and then divided by 100, so that the deduction is .9 rather than .6. This is a simply our algorithmic calibration for the non-linearity of the data.
We make another adjustment when we hit the 97th schedule rank, where we increase the multiple to 2. We could have done this at the 100th position, but 97 seems to be the right cut.
For example, for a schedule rank of 140 , 140 is multiplied by 2 = 280 and then divided by 100, which gives a deduction of 2.8 points from the NAYPPA.
Again, all we are saying here is that as schedule difficulty decreases, it is easier for teams to put up more yards per play on offense and to reduce yards per play on defense, so that an adjustment has to be made for this. However, after the 97th-ranked schedule, no additional adjustment seems to be necessary, except that the ranking of Division II and Division III schedules might be improved.
3. An adjustment for won/loss records
.2 is deducted from the NAYPPA for each game lost by a team.
This third parameter recognizes that there are other team variables that are not accounted for simply by looking at the per play total offense and total defense stats or by adjusting for schedule difficulty. These are elements such as takeaways (turnovers), penalties, special teams, kicking game, coaching in general (a quite large variable), fan support and other factors, which are difficult to integrate as variables in ranking. Wins and losses are a strong reflection of these important non-quantifiable variables.
We have found that there is very good correlation between the results of our SPRS system utilizing NAYPPA and the general ranking of teams by other rating systems. Indeed, we think that this system provides a new improvement for football rating systems used elsewhere. It is a virtually objective tool for team rating that seems to work extremely well as a predictor and as an analytic tool to explain game results.
One great advantage to this system is that anyone can easily use it to rate a football team, either in a ranking system, or in head-to-head competiton.
These are all hand calculations - please report errors if you find them. In the case of ties, the team from a higher NCAA Division is ranked above one from a lower division and the NAIA is placed at the end. We had to have some kind of a rule there, but it may not be right in all instances, especially for the top teams of the NAIA, which are very highly ranked. In the case of ties between teams from the same NCAA Division, the team with the better total defense in yards per play is ranked higher than the other team(s) because defense is more important than offense for winning football.
4. An adjustment for weakness of the defense
As any cursory examination of the total offense and total defense statistics of NCAA teams reveals, total defense is a much stronger indicator of the strength of a team than is total offense, all other things equal. Hence, we make the following adjustment:
.1 is deducted from the NAYPPA of any team whose total defense averages 5.0 yards per play or worse and .2 is deducted from the NAYPPA of any team whose total defense averages 6.0 yards per play or worse .
SAMPLE RATING CALCULATION Nr. 1
Number 1. West Virginia 4.4 6.5 +2.1 42 +1.6 +1.2
West Virginia averaged 6.5 yards per play on offense and 4.4 yards per play on defense, giving the Mountaineers a per play advantage of +2.1 yards per play. The Mountaineers had a schedule ranked 42nd in the country by Massey (40th by Sagarin) so that .5 (the .42 is rounded up to .50) is deducted from the 2.1 yards per play average to compensate for the strength of schedule, leaving +1.6. West Virginia lost 2 games for each of which .2 is further deducted, leaving a final value of +1.2, and that rating result turns out to be the top ranking in the country.
SAMPLE RATING CALCULATION Nr. 2
Number 43. FCS North Dakota State 4.9 7.1 +2.2 160** -1.0 -1.2
FCS North Dakota State averaged 7.1 yards per play on offense and 4.9 yards per play on defense, giving it a per play advantage of +2.2 yards per play. The Bison had a schedule ranked 168th in the country by Massey (160th by Sagariin), so that 3.2 (2 x 1.6) is deducted from the 2.2 yards per play average to compensate for the strength of schedule, leaving -1.0. North Dakota State lost 1 game for which .2 is further deducted, leaving -1.2, and that result turns out to be the 43rd top ranking in the country for all teams and all divisions. Since the Bison are an FCS team, how does this match up with performance? In fact, the Bison beat two Division I-A (FBS) teams this year, one of them the bowl team and Mid-American conference champion Central Michigan 44-14 (rated 82nd in the SPRS System rankings) and Big 10 Minnesota 27-21 (rated 129th in the SPRS System rankings).
On average, 1 point of ranking difference equals about 9 points on the scoreboard, but of course this is difficult to apply to any one game, where chance and circumstance prevail and where things such as home field advantage, etc. must be taken into account in predicting games.
Wednesday, January 16, 2008
Most Popular Posts All Time
- Samsung Digital Picture Frame 2006 is Clear Designer Prior Art to the Later "Design" of the iPhone and iPad
- How to Remove Disabled Greyed-Out Add-on Extensions in Firefox 4: e.g. the Microsoft .NET Framework Assistant (ClickOnce) Extension
- Week 1 2013 College Football FBS Game Predictions by SportPundit
- Legal Graphologists : A Query to You : What About the Christopher Marlowe - William Shakespeare - Controversy and The Signatures of These Two Authors? Are They By the Same Hand?
- PONG ! About that Ridiculously Granted Bounce-Back Scrolling Patent US 7469381 Which Is Anticipated by the Prior Art and Obviousness of the Old Atari Game of PONG
- Speaking of Mouse Trap Economics, What About that Yarn that Patents Stimulate Invention: Les Earnest Testifies Before the USPTO
- 2011 Preseason College Football Rankings
- "The Adobe Flash plugin has crashed" in Mozilla Firefox - A Solution for this Vexing Problem
- Oil & Gas Cartels and American Inertia on Energy Issues : Major Culprits for the Current Economic Situation in the World?
The ISandIS Network
Our Websites and Blogs: 3D Printing and More 99 is not 100 Aabecis AK Photo Blog Ancient Egypt Weblog Ancient Signs (the book) Ancient World Blog AndisKaulins.com Anthropomorphic Design Archaeology Travel Photos (blog) Archaeology Travel Photos (Flickr) Archaeo Pundit Arts Pundit Astrology and Birth Baltic Coachman Bible Pundit Biotechnology Pundit Book Pundit Chronology of the Ancient World Computer Pundit DVD Pundit Easter Island Script Echolat edu.edu Einstein’s Voice Energy Environment and Climate Blog Etruscan Bronze Liver of Piacenza EU Laws EU Legal EU Pundit FaceBook Pundit Gadget Pundit Garden Pundit Golf Pundit Google Pundit Gourmet Pundit Hand Proof HousePundit Human Migrations Idea Pundit Illyrian Language Indus Valley Script Infinity One : The Secret of the First Disk (the game) Jostandis Journal Pundit Kaulins Genealogy Blog Kaulinsium Kiel & Kieler Latvian Blog LawPundit.com Law Pundit Blog LexiLine.com LexiLine Group Lexiline Journal Library Pundit Lingwhizt LinkedIn Literary Pundit Magnifichess Make it Music Maps and Cartography Megalithic World Megaliths Blog) Megaliths.net Minoan Culture Mutatis Mutandis Nanotech Pundit Nostratic Languages Official Pundit Phaistos Disc Pharaonic Hieroglyphs Photo Blog of the World Pinterest Prehistoric Art Pundit Private Wealth Blog PunditMania Quanticalian Quick to Travel Quill Pundit Road Pundit Shelfari SlideShare (akaulins) Sport Pundit Star Pundit Stars Stones and Scholars (blog) Stars Stones and Scholars (book) Stonehenge Pundit The Enchanted Glass Twitter Pundit UbiquitousPundit Vision of Change VoicePundit WatchPundit Wine Pundit Word Pundit xistmz YahooPundit zistmz