I couldn't find a recent thread for this topic, so I started a new on- also, spring workouts start soon and I like to write and read about bsb. D1Baseball has been urging replacement of RPI with another data system that won't result in teams with higher seeding being underdogs to lower seeded teams. Their point is that RPI places too much and inaccurate emphasis on the strength of schedule, opponents SOS and SOS of opponents opponents and use arbitrary and invalid value on away W-L versus home W-L. I can't reasonably explain all of their complaints or the changes they recommend, but the presentation seemed appropriate. A few changes were announced at the American Baseball Coaches Association convention that jiggle the RPI info available to the selection committee that likely won't have dramatic results, but may affect the last few teams who get in or have bubbles popped. D1Bsb reports: The change most likely to have a direct impact on a team’s fortunes in the committee room is a tweak in how results will be grouped in quadrants. Until now, game results were grouped strictly by RPI, with the quadrants as follows: Quad 1 – RPI 1-25 Quad 2 – 26-50 Quad 3 – 51-100 Quad 4 – 101+Moving forward, the quadrants will take into account results being at home, at a neutral site and on the road, as follows: Quad 1 – Home games vs. RPI 1-25, neutral site games vs. 1-40, road games vs. 1-60 Quad 2 – Home games vs. RPI 26-50, neutral site games vs. 41-80, road games vs. 61-120 Quad 3 – Home games vs. RPI 51-100, neutral site games vs. 81-160, road games vs. 121-240 Quad 4 – Home games vs. RPI 101+, neutral site games vs. 161+, road games vs. 241+Note again that this does not signify any change in how the RPI is calculated. It simply changes the way the quadrant records appear on a team sheet in front of the selection committee. The committee also will have access to a data sytem used by the basketball committee, called "the Kevin Pauga Index," which is intended to assess quality of wins and losses, but is not a formal selection criterion, just a resource. That doesn't seem to be a promising resource as much as another way to shift responsibility and muddy selection explanations. Please pardon me for being skeptical of the NCAA.
Home/away is already overrated (detail next paragraph) -- this seems to make that worse, though I'm not sure "quadrants" are that big of a thing to the committee. (From the last 2-3 years, the priorities seem to be in rapidly declining order: RPI, being from a good conference, conference record, and not being NC State.) In the (baseball) RPI, home losses and away wins count 1, while home wins and away losses count 0.7, based on the fraction of times the home team wins throughout D1, which I think was 62%. But 1/0.7 imply that *equally matched* teams would win 62% at home and 38% on the road, ignoring the fact that good teams and teams who play a lot at home are correlated with "rich teams". (UF almost always plays FAMU at home and almost always wins... because there are rich/SEC teams, then there are poor-but-talented-teams (FAU, maybe), and then there's fifty feet of crap... and then there's FAMU.) What do you all think about the adjustment from D1 hockey -- that no win can ever hurt you? Last year, the Gators swept Siena (ended up 271/305 in RPI) by a combined 26-4 (Siena pitched well Friday, followed by two 7-inning run mercy rule games).
That was a point made by the D1Bsb writer that I left out of my post. The digital structure he pushed would include that concept- a team doesn't mathmatically lower their ranking score by winning games, no matter the opponent. It does away with the incentive to cancel games during games against low RPI opponents the season. I'm spoiled because the level to which Sully has raised the program means that Gators don't have to worry about those ranking idiocyncrasies, but it's a real issue for programs that scramble to barely grab an at large bid. In an earlier thread about expanding the number of teams in the postseason, some posts brushed off the idea of more teams in. Gators, LSU, Vandy are fine with athe existing systems, but coaches on a hot seat and fans of their programs REALLY wish they could quit drooling at the window and come in to join the party. I remember the days when Gators would kill for a bowl bid- and hired crooked Charley Pell- and never made bkb postseason. It was painful to be left out, joyous to get in.
One huge problem with schemes to evaluate individual games in a metric for evaluating baseball versus football and basketball is that not every baseball game is the same due to starting pitching. Beating a team with one of their weekday starters is not the same as beating them with their Friday night starter.
And as Sully likes to say, it all depends when you happen to catch them on the schedule. Some weekends a great team might have to sit a Friday starter to rest the arm or they might be shuffling the lineup to shake things up and they go flat. I don't think I saw any metric for those comparisons.
That very real fact is a reason I like the idea of expanding tournaments in each of those sports. Baseball and, especially, football are harder to expand than basketball- I don't see FB going beyond 16 or 32 which the lower levels already do successfully. Baseball easily could add a play-in round to get 32 more teams in by adding a bit more than one extra weekend. If basketball added two more full rounds they could go to 256 teams and a play-in round would put every single D1 team into the Big Dance. That would be pretty nutty for the TV networks, but so what? In every scenario, there would be illustrious programs knocked out early, but so what? The best baseball proposal I've seen would have the final 64 play three-game weekend series, which is the normal format for all teams in the regular season. It would eliminate loser bracket teams playing five games in a weekend. Best of all, it would reduce the sometimes inaccurate and unfair reliance on arbitrary algorithms or arbitrary subjective committee judgments. The issue is much more complex than I've presented it, but my emotional perspective is built on distant memories the Gators being among the teams left out.
I think looking at it as "a win hurts you" is the wrong approach. A win hurts you less than a loss does. We didn't lose value by beating Siena; we lost value by scheduling Siena. Given that we scheduled them, a win was still better for us than a loss was. It needs to be reflected that a win vs. Siena should necessarily be looked on less favorably than a win against, say, Wake Forest. So that relative loss of value has to be incorporated in some capacity. I'd like to see something that shows if some conferences (or even teams) consistently over or under perform at a given seeding; use that data for error corrective measures. *If* West Coast teams typically win consistently more on the road in the postseason at their given seed levels than East Coast teams, then that should be factored in to seeding and/or possibly even hosting modifications. *If* they don't, then publishing that information should lead to less second-guessing. Go GATORS! ,WESGATORS
Way back in ancient by Internet standards time in Usenet rec.sport.basketball.college there was a Duke geek that tracked these things in the NCAA tournament every year. It was very valuable information when making office pool selections given that most people back then in office pools didn't know who Sagarin was and later Ken Pomeroy. Back then traditional powerhouse teams were routinely overseeded.
I miss Usenet. Several good Star Trek boards. Wasn’t the hoops guy from Illinois? Of course Usenet was developed at UNC and Duke, so most of the traffic about bb related to those two schools
That was Milt Epstein. He ran a pick the spread contest for all of the NCAA tournament games. Sagarin Predictor was also a contestant.
