[Cracked Sidewalks]Value Add projects MU 93rd; Burton 36% chance at NBA

[CrackedSidewalksSays]

Value Add projects MU 93rd; Burton 36% chance at NBA

I have put the entire database of Value Add projections into a Google Doc that you can open and even edit by clicking here. The database includes 4250 players listed alphabetically by team and then last name, and you can enter edits for missed players, etc.

Focusing purely on the Value Add's in the Big East, the picture would appear pretty bleak. Marquette had the 101st best Value Add in the country last year and 7th best in the conference. The projections give Marquette the 93rd best Value Add this season (ahead of only DePaul) and Creighton) and then 94th next season (ahead of St. John's, DePaul and Creighton). One note is that Creighton always ends up being projected too low for some reason.

Before I throw up a white flag I should note that the Value Add projections are just an estimate based on the average improvement each player has between seasons, and in fact individual improvement varies wildly when you are talking about college kids. From a purely statistical perspective, things looked every bit as bad heading into 2010, and suddenly Marquette was in the tournament.

The other good news is the players considering Marquette - for example, a signing of Diamond Stone alone would propel Marquette to 5th in the Big East and 60th in the country for 2016.

The other nice news is that the database now includes a percent chance of each player making the NBA, and Deonte Burton is now up to a 36% chance, so hopefully MU can keep the pipeline going. Here is the team table for Big East members:
<div>
</div><div><style type="text/css"> table.tableizer-table {  border: 1px solid #CCC; font-family: Arial, Helvetica, sans-serif  font-size: 12px; }  .tableizer-table td {  padding: 4px;  margin: 3px;  border: 1px solid #ccc; } .tableizer-table th {  background-color: #104E8B;   color: #FFF;  font-weight: bold; } </style>
<table class="tableizer-table"><tbody><tr class="tableizer-firstrow"><th>Team</th><th>Conf</th><th>2014</th><th>Rnk</th><th>2015</th><th>Rnk</th><th>2016</th><th>Rnk</th></tr><tr><td>Villanova</td><td>BE</td><td>25.98</td><td>22</td><td>31.27</td><td>16</td><td>24.17</td><td>25</td></tr><tr><td>Georgetown</td><td>BE</td><td>15.82</td><td>69</td><td>25.82</td><td>28</td><td>25.66</td><td>20</td></tr><tr><td>Providence</td><td>BE</td><td>17.50</td><td>54</td><td>22.68</td><td>40</td><td>11.48</td><td>71</td></tr><tr><td>Xavier</td><td>BE</td><td>14.63</td><td>81</td><td>21.78</td><td>44</td><td>24.83</td><td>23</td></tr><tr><td>St. John's</td><td>BE</td><td>15.15</td><td>76</td><td>19.14</td><td>51</td><td>4.39</td><td></td></tr><tr><td>Butler</td><td>BE</td><td>9.28</td><td>147</td><td>18.68</td><td>55</td><td>9.12</td><td>86</td></tr><tr><td>Seton Hall</td><td>BE</td><td>11.40</td><td>110</td><td>14.74</td><td>78</td><td>19.80</td><td>38</td></tr><tr><td>Marquette</td><td>BE</td><td>12.36</td><td>101</td><td>12.66</td><td>93</td><td>8.54</td><td>94</td></tr><tr><td>Creighton</td><td>BE</td><td>24.47</td><td>27</td><td>11.18</td><td>108</td><td>5.31</td><td>143</td></tr><tr><td>DePaul</td><td>BE</td><td>5.59</td><td></td><td>6.08</td><td></td><td>5.82</td><td>133</td></tr></tbody></table>
</div><div>

</div>

Source: Value Add projects MU 93rd; Burton 36% chance at NBA

[MU82]

Before I throw up a white flag I should note that the Value Add projections are just an estimate based on the average improvement each player has between seasons, and in fact individual improvement varies wildly when you are talking about college kids. ... The other good news is the players considering Marquette ...

The best news is that games aren't played in a computer program but on a basketball court!

[murara1994]

The other good news is the players considering Marquette - for example, a signing of Diamond Stone alone would propel Marquette to 5th in the Big East and 60th in the country for 2016

^^Except we all know that ain't happening.

[Silkk the Shaka]

The other good news is the players considering Marquette - for example, a signing of Diamond Stone alone would propel Marquette to 5th in the Big East and 60th in the country for 2016

^^Except we all know that ain't happening.

Replace the name Stone with Ellenson and the same probably holds.

[GGGG]

Wow for last year, this was remarkably consistent with how the standings ended up.

Team (VA Projection) (Place)

Villanova (1) (1)
Creighton (2) (2)
Providence (3) (3)
Georgetown (4) (7)
St. Johns (5) (5)
Xavier (6) (4)
Marquette (7) (6)
Seton Hall (8) (8)
Butler (9) (9)
DePaul (10) (10)

So really the only team that did vastly different than your VA prediction was Georgetown, whose poor performance kicked some teams up a spot.  Xavier, SJU and Providence tied for third so the differences there are not terribly significant.

So yeah, this doesn't portend very well for Marquette this year does it.  

[tower912]

The Value Add projected MU to finish 6th last year?    Huh.   

[mu03eng]

The Value Add projected MU to finish 6th last year?    Huh.   

Wait it did?  That is very perplexing.

[GGGG]

The Value Add projected MU to finish 6th last year?    Huh.   

Actually they were predicted to finish 7th.  Georgetown's underperformance is what pushed MU up to 6th.

So *if I understand the data correctly,* Marquette performed about as expected relative to the other BE schools last year.

[Lennys Tap]

Actually they were predicted to finish 7th.  Georgetown's underperformance is what pushed MU up to 6th.

So *if I understand the data correctly,* Marquette performed about as expected relative to the other BE schools last year.

You've said it, I've said it, others have said it. Marquette's "expectations" last year (as set by the conference coaches, AP, USA Today and some of our flat earth posters) were ridiculous - especially without Blue, Du Wilson and McKay. All the hand wringing over the coach "changing", not trying to win, etc., was nonsense. Buzz the coach was who he always was, Buzz the GM was the guy caught short on talent. Can't wait to see Ners and his posse weigh in and attack Pudner's accurate and prescient analysis of last year's team - had John published them earlier here maybe it could have prevented some of the silliness of the past 8 months.

[GGGG]

Yeah, well I am not sure I am understanding the data correctly.  But I think last year's predictions were based on people "stepping up" to levels they couldn't quite reach.  It most certainly is a cautionary tale for this year given what some people's expectations might be.

The more I think about it, the more I believe that the NIT would be a satisfactory outcome for this year.

[mu03eng]

You've said it, I've said it, others have said it. Marquette's "expectations" last year (as set by the conference coaches, AP, USA Today and some of our flat earth posters) were ridiculous - especially without Blue, Du Wilson and McKay. All the hand wringing over the coach "changing", not trying to win, etc., was nonsense. Buzz the coach was who he always was, Buzz the GM was the guy caught short on talent. Can't wait to see Ners and his posse weigh in and attack Pudner's accurate and prescient analysis of last year's team - had John published them earlier here maybe it could have prevented some of the silliness of the past 8 months.

I think you overestimate the willingness of some to accept a somewhat complex statistical premise.

I can't seem to find the link to the older data, but it would be interesting to see what the forecast for the individuals were last season and then what they turned out to be.  Especially given that MU performed to the value add "expectations".

My sense is that Derrick and Jamil vastly underperformed what was anticipated(previous performance plus growth), Davante and Chris were close to forecast and the only overachiever was Burton.  But from a math standpoint, if Jamil and Derrick were as bad as we think, multiple players had to overperform expectations for the team to just meet expectations.  I'd like to know who those folks are.

Also, something that is interesting in the analysis, or should be....you could point to Burton outperforming his expectations so he should have gotten more minutes.  That could have been true but there is nothing to suggest increasing his usage would have increased the team performance.  Increased usage could lead to decreased efficiency.

[GGGG]

I think you overestimate the willingness of some to accept a somewhat complex statistical premise.

"NNNNNNNNNNNNNEEEEEEEEEEEEEEERRRRRRRRRRRRRRDDDDDDDDDDDDDDSSSSSSSSSSSS!!!!!!!!!!!!!!!!"

[Canned Goods n Ammo]

Listen up nerds,

There is no stat to account for Buzz's insane substitution pattern with Dawson. Anybody who has played the sport at a high level knows that Buzz ruined JJJ and Dawson with his crazy substitution patterns.

Apparently you guys have never played high level ball, and can't recognize these facts. Yes, these are FACTS that I have presented.

Have fun on your computers.

[Lennys Tap]

Eye testers vs science - "Screw those meteorologists, I'm staying on the beach. I've been around a bit, and it doesn't even look like rain to me".

[mu03eng]

"NNNNNNNNNNNNNEEEEEEEEEEEEEEERRRRRRRRRRRRRRDDDDDDDDDDDDDDSSSSSSSSSSSS!!!!!!!!!!!!!!!!"

I resemble that remark

[TAMU, Knower of Ball]

Listen up nerds,

There is no stat to account for Buzz's insane substitution pattern with Dawson. Anybody who has played the sport at a high level knows that Buzz ruined JJJ and Dawson with his crazy substitution patterns.

Apparently you guys have never played high level ball, and can't recognize these facts. Yes, these are FACTS that I have presented.

Have fun on your computers.

It's still early but I think you win scoop for the day.

[Dr. Blackheart]

And Georgetown underperformed because Buzz "discovered" Todd Mayo and Steve Taylor in MU's two victories

[GGGG]

And Georgetown underperformed because Buzz "discovered" Todd Mayo and Steve Taylor in MU's two victories

And you raise a good point here.  I don't have the time (nor the desire) to do this, but how did Marquette's players perform to their Value Add statistics last year?  Hypothetically, some could have performed less than the model anticipated, and some more.  Could they have had better results playing more of the latter?  I don't know enough about the model.

I really don't want to this to become another debate about who should have played more.  Really my only point was that the expectations for last year's team were inflated when compared to the level at which they reasonably could perform.   And that we should use that as a cautionary tale for this year as well.

[TAMU, Knower of Ball]

And you raise a good point here.  I don't have the time (nor the desire) to do this, but how did Marquette's players perform to their Value Add statistics last year?  Hypothetically, some could have performed less than the model anticipated, and some more.  Could they have had better results playing more of the latter?  I don't know enough about the model.

I really don't want to this to become another debate about who should have played more.  Really my only point was that the expectations for last year's team were inflated when compared to the level at which they reasonably could perform.   And that we should use that as a cautionary tale for this every year as well.

FIFY

[tower912]

Second question.   Is there a way to factor  the VA statistics with Blue, McKay and DuWilson added in?    In other words, according to value added, how much better would MU have been with those 3?

[Lennys Tap]

And Georgetown underperformed because Buzz "discovered" Todd Mayo and Steve Taylor in MU's two victories

And Magic Dawesome in one.

[mu03eng]

And you raise a good point here.  I don't have the time (nor the desire) to do this, but how did Marquette's players perform to their Value Add statistics last year?  Hypothetically, some could have performed less than the model anticipated, and some more.  Could they have had better results playing more of the latter?  I don't know enough about the model.

I really don't want to this to become another debate about who should have played more.  Really my only point was that the expectations for last year's team were inflated when compared to the level at which they reasonably could perform.   And that we should use that as a cautionary tale for this year as well.

I like this even more than when I said it 

It would add credibility if John compared beginning of the year prediction to end of the year actuals for each year.  Like I said before, if Derrick and Jamil vastly underperformed like we all think, than several players had to wildly overperform for the team to finish at the predicted level.

[Canned Goods n Ammo]

I like this even more than when I said it 

It would add credibility if John compared beginning of the year prediction to end of the year actuals for each year.  Like I said before, if Derrick and Jamil vastly underperformed like we all think, than several players had to wildly overperform for the team to finish at the predicted level.

Also, because basketball is a team sport, if Jamil had a great year, arguably, the PG & SG spots might not have been as painful as they were.

You can have a defensive/pass first/no shoot PG, but you better have weapons at the other 4 spots. On the flip side, having a non-scoring PG can put a lot of pressure on the other 4 to produce, especially to keep the floor spread.

Jake was a nice role player, but MU was asking him to do a lot last year, and I think it led to some inefficiency and over-exposure.

[The Equalizer]

I like this even more than when I said it 

It would add credibility if John compared beginning of the year prediction to end of the year actuals for each year.  Like I said before, if Derrick and Jamil vastly underperformed like we all think, than several players had to wildly overperform for the team to finish at the predicted level.

Its easy enough to calculate.

He includes a historical database, so we have access to 2013 actuals. 
He's identified the average year-over-year improvement (102% for rising frosh, 31% for sophs, 18% for juniors).

Therefore, we can calculate what each player would have been expected to be for 2014 if they were average.  I thought it would be interesting to include the 2014 actual value-add, per the database.

Player	2013 Value Add	Expected Improvement	Expected 14 VA	Actual 2014 VA
Juan Anderson	1.12	31%	1.47	1.1
Davante Gardner	6.13	18%	7.23	4.75
Todd Mayo	0.66	18%	0.78	1.82
Chris Otule	2.06	18%	2.43	0.41
Steve Taylor	1.47	102%	2.97	0.08
Jake Thomas	0.44	18%	0.51	0.77
Derrick Wilson	0.9	31%	1.17	0
Jamil Wilson	4.2	18%	4.96	1.47
TOTALS	16.98		21.52	10.40

Even accounting for the likely error in data (I don't think for all the criticism directed at Derrick Wilson that he truly is a zero Value Add), there was a bigger problem last year than merely playing to expectations.

[mu03eng]

Its easy enough to calculate.

He includes a historical database, so we have access to 2013 actuals. 
He's identified the average year-over-year improvement (102% for rising frosh, 31% for sophs, 18% for juniors).

Therefore, we can calculate what each player would have been expected to be for 2014 if they were average.  I thought it would be interesting to include the 2014 actual value-add, per the database.

Player	2013 Value Add	Expected Improvement	Expected 14 VA	Actual 2014 VA
Juan Anderson	1.12	31%	1.47	1.1
Davante Gardner	6.13	18%	7.23	4.75
Todd Mayo	0.66	18%	0.78	1.82
Chris Otule	2.06	18%	2.43	0.41
Steve Taylor	1.47	102%	2.97	0.08
Jake Thomas	0.44	18%	0.51	0.77
Derrick Wilson	0.9	31%	1.17	0
Jamil Wilson	4.2	18%	4.96	1.47
TOTALS	16.98		21.52	10.40

Even accounting for the likely error in data (I don't think for all the criticism directed at Derrick Wilson that he truly is a zero Value Add), there was a bigger problem last year than merely playing to expectations.

What are the actual value adds for the freshman and what is the "average" value add that John assumes for his model?