[Cracked Sidewalks] Responses to Sagarin questions; Consistent MU has played like a Top 20 team 11 of 17 games

[CrackedSidewalksSays]

Responses to Sagarin questions; Consistent MU has played like a Top 20 team 11 of 17 games

Written by: jpudner@concentricgrasstops.com (bamamarquettefan1)

There were three excellent questions in response to my post “Seton Hall’s improvement with Hazell really pushes MU projected finish to 9-9.”  I summarized the questions below, and in researching the answers, reached a basic conclusion:

MU plays like a Top 20 team two out of every three times they go on the court for a big game.  One in three times, they battle through a subpar performance but have played hard enough to sneak out half of those games (Bucknell, UWM and South Florida).

If this trend continues in MUs last four home games they could play like a Top 20 team and have easy wins in any three of the four games, BUT will likely find themselves in a battle in at least one of those four games during which they may need to pull out an ugly win a la Bucknell, UWM or South Florida.

Only once in 17 games (Notre Dame win) has MU played well enough to win AT UConn or Georgetown, so unless they can take it to another level for one of those games, MU needs to sweep the four home games or face a potential season-ending bracket buster at Seton Hall to get the 10th win and punch their ticket to the NCAA.

But to address the three questions I received on MU Scoop – paraphrasing:

A. Don’t the Sagarin “Game Rankings” vary a lot from game-to-game anyway?

No.  MU has put out virtually the same great effort in 10 of 17 games this year against Top 150 teams.  Basically, an 88 or higher means you are a Top 20 team, and in 10 of 17 games, MUs rating has been between an 88 and a 91.  Assuming MU’s opponents play above or below their average on a given night, the fact that MU is staying within a 3-point margin at this high level is unbelievably consistent.

The blowout of Notre Dame was one of the best games played by any team this season, resulting in a game rating of 107, meaning Sagarin says MU would have beaten No. 1 Ohio State by 13 points on a neutral court that night.

That leaves six subpar performances; Bucknell 84, USF 81, Wisconsin 81, Gonzaga 80, UWM 79 and by far the worst performance of a 75 against UConn (though UConn just looked hot and flawless to me that night).

B. “You say that Seton Hall’s rating has improved by 7 points since Hazell arrived, but for comparison what is Marquette’s rating over the past 8 games?”

Compared to the first 9 games against top 150 teams, Marquette has improved 2 points in the last 8 games.  MU averaged an 86 through their first 9 games against Top 150 teams, and has averaged an 88 through the last 8 – so nothing close to Seton Hall’s 7-point improvement, but still improving.  MU  averaged an 87 against the seven creampuffs.  If the Notre Dame win had been 10 days earlier though, then MUs average would have been the same in the first 9, second 8 and vs. the creampuffs.

C. How do you calculate these game ratings?

Note that due to rounding and perhaps some other nuisances of Sagarin’s overall system, this simple game-by-game calculation gives MU an overall average one point higher than the 86 Sagarin calculates through today.  Also note that if you calculate once and then do it again a week later, the game rating might fluctuate slightly due to games played by other teams.  However, the only game score that fluctuated by more than one point since my first column was Notre Dame, whose overall rating had jumped 3 points, thus increasing MUs game rating for the blowout win from 104 to 107.

1. Pull up the current Sagarin ratings here:

http://www.usatoday.com/sports/sagarin/bkt1011.htm


2. Find your opponent and go to the number on the far right.

The predictor number is in blue.  For last night’s game, you find South Florida as the 126th ranked team in the country and with a 75.72, which I always round, so a 76.

3. Adjust 4 points for home court.

Adjust that number based on where the game was played.  Sagarin calculates home advantage is 3.91 so far this year, but I always round to 4.  So if MU had been at home last night you would subtract 4 (-4), on a neutral site make no adjustment, but because it was at South Florida you add 4 to Marquette’s game score (76+4=80) to get to 80.

4. Final margin.

Finally your most important number is the actual difference in the final score.  MU won by 1 point, so you add one to make the total equation 76+4+1 and MU’s game rating is an 81.

Game Rating vs. S. Florida = 81

Feel free to use decimals to be more precise, but if you are sneaking this in between work projects or before going to bed like me, going with rounded numbers speeds things up incredibly and may save your marriage and job.

I choose to leave non-top 150 teams out because at that point you are judging a team based on whether or not the blow someone out by 20 points or 40 points – which is silly since obviously coaches use those games to try out subs, etc.  Overall, the creampuff game ratings average out to virtually the same performances as MUs other games, but I really don’t care that our 30-point win over DePaul is technically our second best performance of the season, or that our 29-point win over Centenary is technically our 2nd worst performance of the year.  When it comes to evaluating the games that matter, MU either plays like a Top 20 team (two-thirds of the time), or guts their way through a bad night to get back into single digits and take their best shot (one-third of the time).  They quit 0% of the time from what I can see.

http://www.crackedsidewalks.com/2011/02/responses-to-sagarin-questions.html

[babytownfrolics]

Curious, you say that MU has to play essentially their best game to beat either UConn or Georgetown, does this account for the possibility that UConn or Georgetown might happen to play poorly or below their standards in either of those games?

[rocky_warrior]

Thanks for the (more detailed) explanation of where the numbers come from.  Now I probably could do it myself if I had too.  Luckily, you're on top of it so far

[bamamarquettefan]

That is a great question.  

In most games, the result is within 6 points either way of the teams two ratings.  So my gut is that typically a team is about a +3 on a good night, and a -3 on a bad night.  So if you take the averages now:

UConn = 89 average
Georgetown = 90 average
MU in 10 of 17 games = between an 88 and a 91

Let's say that on a good night we are a 91, an average night is really an 88, and a bad night is worse and we get killed in these games.

Because UConn and Georgetown get a 4-point boost for being at home, even if we have a good night (91), we need both of them to have a poor night (86 and 87 respectively) just to make the game dead even.

So we need to play very well and they need to play poorly for this to come down to overtime or a last second shot.

But a few times each year a team team typically plays way above their range, so obviously if we exploded for one night like we did against Notre Dame, then we can win.

We have to either play our second best game of the year OR we have to play very well on the same night UConn/Gtown plays poorly and then get a bounce at the end of the game to pull it out.

[bamamarquettefan]

Thanks for the (more detailed) explanation of where the numbers come from.  Now I probably could do it myself if I had too.  Luckily, you're on top of it so far

thanks, but maybe you missed my line on saving your marriage and job by rounding!  Please feel free to jump in and calculate the game ratings anytime you want so I can get back to work :-)