Never Apologize, Never Explain 2017-18

Never2

This page describes the methodology for this year's TBRW? Predictions.  As always if you catch an error please contact the humble author.

RS

RS is each team's points in the prior ECAC regular season:

RS  
Brown 7  
Clarkson 23  
Colgate 15  
Cornell 31  
Dartmouth 16  
Harvard 34  
Princeton 19  
Quinnipiac 27  
RPI 12  
SLU 28  
Union 34  
Yale 18  

PS

PS is the number of upset (lower seed) advances (+1) or eliminations (-1) in the previous season's ECAC tournament.  The consolation game is ignored.

The upsets in the last season's ECAC tournament were:

PS  
Brown 0  
Clarkson 0  
Colgate 0  
Cornell +1  
Dartmouth 0  
Harvard 0  
Princeton 0  
Quinnipiac +1  
RPI 0  
SLU -1  
Union -1  
Yale 0  

 

Imp

Imp is the difference in points gained in the second half (games 12-22) minus the first half (games 1-11) of the prior ECAC regular season, divided by 2, rounded down.  A positive Imp indicates the team had a better second half:

Pts 2nd ½ 1st ½ Diff Imp  
Brown 7 3 4 -1 0  
Clarkson 23 10 13 -3 -1  
Colgate 15 7 8 -1 0  
Cornell 31 15 16 -1 0  
Dartmouth 16 6 10 -4 -2  
Harvard 34 19 15 +4 +2  
Princeton 19 12 7 +5 +2  
Quinnipiac 27 14 13 +1 0  
RPI 12 10 2 +8 +4  
SLU 28 10 18 -8 -4  
Union 34 15 19 -4 -2  
Yale 18 8 10 -2 -1  

 

RP%: Returning Points Percentage

This year we are replacing the old methodology for estimating returning strength that combined forwards, defense and goaltending.  Instead we use a simpler statistic: the percentage of returning point scoring (RP%).  The old metric didn't seem to be doing a great job of predicting the fall-off due to departing talent.  The goaltending figure overwhelmed the other components.  The old measure had an arbitrary cutoff with relatively weak theoretical justification.  We intoduce another measure, RA (described later), to capture what is lost in this simplification.

The following players with eligibility remaining are not returning:

 RP% = Points Returning / Total Points from Prior Season

  Total Sr Early Return RP%  
Brown 179 16 0 163 .91  
Clarkson 318 107 36 175 .55  
Colgate 200 80 0 120 .60  
Cornell 268 100 0 168 .63  
Dartmouth 222 59 0 163 .73  
Harvard 395 193 0 202 .51  
Princeton 281 34 0 247 .88  
Quinnipiac 336 82 0 254 .76  
RPI 226 48 11 167 .74  
SLU 275 84 0 191 .69  
Union 386 109 81 196 .51  
Yale 270 84 0 186 .69  

 

RA: Returning Awardees

Since returning scoring only captures the offense, we're also introducing as new metric, Returning Awardees (RA), to account for the return (or loss) of highly talented players.  RA is the net return (or loss) of ECAC First Team, POTY, and ROTY awards.  The potential repetition of players (e.g., First Team and POTY) is intentional to allow for the impact of unusually talented players,

The 2017 ECAC awardees were:

Award Name Team Status
1st team Foo Union Depart
1st team Kerfoot Harvard Depart
1st team Vecchione Union Depart
1st team Bayreuther SLU Depart
1st team Fox Harvard Return
1st team Hayton SLU Depart
POTY Vecchione Union Depart
ROTY Fox Harvard Return

The calculation: add one point for each returning awardee while subtracting one for each departure:

  Return Depart RA  
Brown 0 0 0  
Clarkson 0 0 0  
Colgate 0 0 0  
Cornell 0 0 0  
Dartmouth 0 0 0  
Harvard 2 1 1  
Princeton 0 0 0  
Quinnipiac 0 0 0  
RPI 0 0 0  
SLU 0 2 -2  
Union 0 3 -3  
Yale 0 0 0  

 

Avg10

Avg10 measures traditional program strength, by taking the team's mean number of Points over the prior 10 seasons.  This will be used to compute the expected Incoming factor.

Brn Clk Cgt Cor Drt Hvd Prn Qpc RPI SLU Uni Yal  
2008 15 33 21 25 15 27 28 22 15 16 25 22  
2009 10 20 17 29 24 24 28 21 13 26 20 32  
2010 16 11 26 31 17 17 18 22 23 23 28 32  
2011 18 19 11 24 26 15 24 19 24 13 36 35  
2012 14 22 23 30 19 25 16 23 17 21 32 22  
2013 20 19 15 19 22 14 20 37 27 22 24 25  
2014 17 24 29 24 16 16 8 28 21 18 37 24  
2015 13 19 26 22 26 25 6 35 18 29 17 28  
2016 12 23 14 22 22 28 9 37 23 25 18 31  
2017 7 23 15 31 16 34 19 27 12 28 34 18  
   
Pts 142 213 197 257 203 225 176 271 193 221 271 269  
Avg10 14.2 21.3 19.7 25.7 20.3 22.5 17.6 27.1 19.3 22.1 27.1 26.9  

Summary:

Avg10  
Brown 14.20  
Clarkson 21.30  
Colgate 19.70  
Cornell 25.70  
Dartmouth 20.30  
Harvard 22.50  
Princeton 17.60  
Quinnipiac 27.10  
RPI 19.30  
SLU 22.10  
Union 27.10  
Yale 26.90  

 

Putting it All Together

Now we put everything together to predict the final standings.  This is a function of both the strength of returning players and an estimate of the strength of the new players.

The Past

Firsty, we want a measure of total returning strength, both the quality of the prior roster and its quantity -- the percentage of that roster returning.

Prior, the quality is the sum of RS, PS, Imp, and RA.

RP%,  the quantity, is simply carried down from above.

Past, the relative measure of returning strength, is the product of Prior and RP%.

RS PS Imp RA Prior RP% Past  
Brown 7 0 0 0 7 .91 6.38  
Clarkson 23 0 -1 0 22 .55 12.10  
Colgate 15 0 0 0 15 .60 9.00  
Cornell 31 +1 0 0 32 .63 20.06  
Dartmouth 16 0 -2 0 14 .73 10.28  
Harvard 34 0 +2 1 36 .51 18.91  
Princeton 19 0 +2 0 21 .88 18.46  
Quinnipiac 27 +1 0 0 28 .76 21.17  
RPI 12 0 +4 0 16 .74 11.82  
SLU 28 -1 -4 -2 21 .69 14.57  
Union 34 -1 -2 -3 28 .51 14.28  
Yale 18 0 -1 0 17 .69 11.71  

 

The Future

Secondly,  we want to makew the same two estimates for the incoming player: their quality and quantity.

Avg10 is a proxy for quality we take the mean RS for the prior 10 seasons.

Inc%, the estimate of quantity, is simply 1 - RP%.

Fut, the relative estimate of incoming strength, is the product of Avg10 and Inc%.

 

Past Avg10 Inc% Fut  
Brown 6.38 14.20 .09 1.26  
Clarkson 12.10 21.30 .45 9.59  
Colgate 9.00 19.70 .40 7.88  
Cornell 20.06 25.70 .37 9.59  
Dartmouth 10.28 20.30 .27 5.40  
Harvard 18.91 22.50 .49 11.00  
Princeton 18.46 17.60 .12 2.13  
Quinnipiac 21.17 27.10 .24 6.61  
RPI 11.82 19.30 .26 5.04  
SLU 15.96 22.10 .31 6.76  
Union 14.28 27.10 .49 13.28  
Yale 11.71 26.90 .31 8.37  

 

Conclusion

All that's left to do is add Past and Fut together (Net), calculate a pro-rated normalization so the final prediction is against a mean of 22 points (Norm = 22 - mean of Net = 22-21.42 = 0.58), and then apply that normalizsation to produce our old friends, the predicted RS (Nieu) and standing (Pred).

Total:

Past Fut Net Norm Nieu Pred  
Brown 6.38 1.26 7.64 .70 8.34 12  
Clarkson 12.10 9.59 21.69 .70 22.39 5  
Colgate 9.00 7.88 16.88 .70 17.58 9  
Cornell 20.06 9.59 29.65 .70 30.35 2  
Dartmouth 10.28 5.40 15.68 .70 16.38 11  
Harvard 18.91 11.00 29.91 .70 30.61 1  
Princeton 18.46 2.13 20.59 .70 21.29 7  
Quinnipiac 21.17 6.61 27.78 .70 28.48 3  
RPI 11.82 5.04 16.86 .70 17.56 10  
SLU 14.57 6.76 22.72 .70 22.04 6  
Union 14.28 13.28 27.56 .70 28.26 4  
Yale 11.71 8.37 20.08 .70 20.78 8