Player impact on teammate's offensive stats

Using the numbers from the new plus/minus tool at basketball-reference.com, I thought it would be interesting to explore how the Rockets players impact their teammate's efficiency when they are on the floor with them.

I'll describe what I did in 4 steps.

(1) What are the total team stats when each Rocket player is on the floor?

Using the new Plus/Minus Finder at basketball-reference.com, I found the following stats through 3/23:

Spoiler

<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: #B90D0D; color: #FFF; font-weight: bold;} </style>
<table class="tableizer-table"> <tr style="background-color: #B90D0D; color: #FFF; font-weight: bold;"><th>Player </th><th>Tm </th><th>Season </th><th>G </th><th>MP </th><th>Tm </th><th>Opp </th><th>Pace </th><th>FG </th><th>FGA </th><th>3P </th><th>3PA </th><th>FT </th><th>FTA </th><th>ORB </th><th>DRB </th><th>TRB </th><th>AST </th><th>STL </th><th>BLK </th><th>TOV </th><th>PF</th><th>PTS </th></tr> <tr><td>Kyle Lowry </td><td>HOU </td><td>2011-12 </td><td>38</td><td>1346.8</td><td>2606</td><td>2606</td><td>92.9</td><td>1045</td><td>2311</td><td>209</td><td>572</td><td>462</td><td>571</td><td>313</td><td>859</td><td>1172</td><td>590</td><td>211</td><td>140</td><td>427</td><td>580</td><td>2761</td></tr> <tr><td>Kevin Martin </td><td>HOU </td><td>2011-12 </td><td>40</td><td>1263.6</td><td>2446</td><td>2442</td><td>92.8</td><td>977</td><td>2183</td><td>186</td><td>541</td><td>437</td><td>527</td><td>293</td><td>793</td><td>1086</td><td>533</td><td>179</td><td>137</td><td>381</td><td>545</td><td>2577</td></tr> <tr><td>Chandler Parsons </td><td>HOU </td><td>2011-12 </td><td>46</td><td>1245.6</td><td>2398</td><td>2409</td><td>92.6</td><td>978</td><td>2125</td><td>182</td><td>493</td><td>448</td><td>574</td><td>320</td><td>788</td><td>1108</td><td>534</td><td>187</td><td>134</td><td>396</td><td>514</td><td>2586</td></tr> <tr><td>Luis Scola </td><td>HOU </td><td>2011-12 </td><td>48</td><td>1483.1</td><td>2851</td><td>2849</td><td>92.2</td><td>1156</td><td>2518</td><td>212</td><td>582</td><td>482</td><td>598</td><td>328</td><td>937</td><td>1265</td><td>683</td><td>225</td><td>153</td><td>478</td><td>643</td><td>3006</td></tr> <tr><td>Samuel Dalembert </td><td>HOU </td><td>2011-12 </td><td>48</td><td>1178.7</td><td>2239</td><td>2246</td><td>91.3</td><td>927</td><td>1999</td><td>163</td><td>448</td><td>381</td><td>469</td><td>305</td><td>740</td><td>1045</td><td>531</td><td>177</td><td>128</td><td>390</td><td>483</td><td>2398</td></tr> <tr><td>Goran Dragic </td><td>HOU </td><td>2011-12 </td><td>48</td><td>1092.1</td><td>2062</td><td>2059</td><td>90.6</td><td>857</td><td>1853</td><td>156</td><td>444</td><td>342</td><td>451</td><td>258</td><td>664</td><td>922</td><td>467</td><td>173</td><td>87</td><td>317</td><td>437</td><td>2212</td></tr> <tr><td>Courtney Lee </td><td>HOU </td><td>2011-12 </td><td>40</td><td>1088.3</td><td>2051</td><td>2055</td><td>90.5</td><td>835</td><td>1847</td><td>165</td><td>442</td><td>351</td><td>461</td><td>259</td><td>640</td><td>899</td><td>458</td><td>179</td><td>94</td><td>325</td><td>441</td><td>2186</td></tr> <tr><td>Patrick Patterson </td><td>HOU </td><td>2011-12 </td><td>46</td><td>1042.5</td><td>1981</td><td>1985</td><td>91.3</td><td>806</td><td>1804</td><td>149</td><td>416</td><td>340</td><td>436</td><td>254</td><td>632</td><td>886</td><td>415</td><td>156</td><td>92</td><td>299</td><td>414</td><td>2101</td></tr> <tr><td>Chase Budinger </td><td>HOU </td><td>2011-12 </td><td>40</td><td>832.8</td><td>1602</td><td>1601</td><td>92.3</td><td>663</td><td>1440</td><td>124</td><td>324</td><td>247</td><td>329</td><td>186</td><td>534</td><td>720</td><td>391</td><td>128</td><td>71</td><td>253</td><td>342</td><td>1697</td></tr> <tr><td>Jordan Hill </td><td>HOU </td><td>2011-12 </td><td>32</td><td>469.2</td><td>896</td><td>898</td><td>91.8</td><td>381</td><td>815</td><td>56</td><td>169</td><td>122</td><td>156</td><td>104</td><td>307</td><td>411</td><td>206</td><td>71</td><td>37</td><td>135</td><td>183</td><td>940</td></tr> <tr><td>Marcus Camby </td><td>HOU </td><td>2011-12 </td><td>3</td><td>54.4</td><td>101</td><td>104</td><td>90.4</td><td>35</td><td>97</td><td>6</td><td>20</td><td>17</td><td>25</td><td>19</td><td>36</td><td>55</td><td>17</td><td>7</td><td>7</td><td>18</td><td>19</td><td>93</td></tr> <tr><td>Marcus Morris </td><td>HOU </td><td>2011-12 </td><td>9</td><td>64.7</td><td>129</td><td>128</td><td>95.3</td><td>49</td><td>110</td><td>8</td><td>27</td><td>23</td><td>31</td><td>14</td><td>47</td><td>61</td><td>26</td><td>9</td><td>3</td><td>22</td><td>28</td><td>129</td></tr> <tr><td>Greg Smith </td><td>HOU </td><td>2011-12 </td><td>7</td><td>49.5</td><td>87</td><td>89</td><td>85.3</td><td>33</td><td>80</td><td>8</td><td>17</td><td>10</td><td>18</td><td>11</td><td>28</td><td>39</td><td>14</td><td>6</td><td>5</td><td>11</td><td>20</td><td>84</td></tr> <tr><td>Courtney Fortson </td><td>HOU </td><td>2011-12 </td><td>4</td><td>32</td><td>64</td><td>64</td><td>96.2</td><td>22</td><td>57</td><td>3</td><td>12</td><td>11</td><td>13</td><td>8</td><td>23</td><td>31</td><td>10</td><td>5</td><td>6</td><td>12</td><td>12</td><td>58</td></tr> <tr><td>Terrence Williams </td><td>HOU </td><td>2011-12 </td><td>12</td><td>181.5</td><td>336</td><td>336</td><td>88.9</td><td>120</td><td>295</td><td>23</td><td>81</td><td>62</td><td>80</td><td>44</td><td>115</td><td>159</td><td>64</td><td>18</td><td>17</td><td>54</td><td>77</td><td>325</td></tr> <tr><td>Jonny Flynn </td><td>HOU </td><td>2011-12 </td><td>11</td><td>134.9</td><td>251</td><td>258</td><td>90.6</td><td>96</td><td>221</td><td>16</td><td>37</td><td>38</td><td>56</td><td>32</td><td>82</td><td>114</td><td>55</td><td>13</td><td>15</td><td>39</td><td>53</td><td>246</td></tr> <tr><td>Hasheem Thabeet </td><td>HOU </td><td>2011-12 </td><td>5</td><td>22.8</td><td>43</td><td>46</td><td>93.8</td><td>9</td><td>31</td><td>1</td><td>4</td><td>13</td><td>16</td><td>6</td><td>18</td><td>24</td><td>4</td><td>2</td><td>2</td><td>11</td><td>10</td><td>32</td></tr> <tr><td>Jeff Adrien </td><td>HOU </td><td>2011-12 </td><td>8</td><td>62.8</td><td>127</td><td>125</td><td>96.4</td><td>36</td><td>99</td><td>8</td><td>26</td><td>29</td><td>44</td><td>11</td><td>52</td><td>63</td><td>22</td><td>9</td><td>7</td><td>22</td><td>34</td><td>109</td></tr></table>

In particular, here are some team efficiency statistics when each player is on the floor (PPP is "points per play" -- its sort of like TS% adjusted for turnovers).

Spoiler

<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: #B90D0D; color: #FFF; font-weight: bold;} </style> <table class="tableizer-table"> <tr style="background-color: #B90D0D; color: #FFF; font-weight: bold;"><th>Player </th><th>2fg%</th><th>3fg%</th><th>efg%</th><th>ts%</th><th>tov%</th><th>ppp</th><th>ORTG</th></tr> <tr><td>Kyle Lowry </td><td>48.1%</td><td>36.5%</td><td>49.7%</td><td>53.9%</td><td>16.4%</td><td>0.924</td><td>105.9</td></tr> <tr><td>Kevin Martin </td><td>48.2%</td><td>34.4%</td><td>49.0%</td><td>53.4%</td><td>15.6%</td><td>0.922</td><td>105.4</td></tr> <tr><td>Chandler Parsons </td><td>48.8%</td><td>36.9%</td><td>50.3%</td><td>54.4%</td><td>16.5%</td><td>0.932</td><td>107.8</td></tr> <tr><td>Luis Scola </td><td>48.8%</td><td>36.4%</td><td>50.1%</td><td>54.0%</td><td>16.8%</td><td>0.922</td><td>105.4</td></tr> <tr><td>Samuel Dalembert </td><td>49.3%</td><td>36.4%</td><td>50.5%</td><td>54.4%</td><td>17.4%</td><td>0.924</td><td>107.1</td></tr> <tr><td>Goran Dragic </td><td>49.8%</td><td>35.1%</td><td>50.5%</td><td>53.9%</td><td>15.4%</td><td>0.934</td><td>107.3</td></tr> <tr><td>Courtney Lee </td><td>47.7%</td><td>37.3%</td><td>49.7%</td><td>53.3%</td><td>15.8%</td><td>0.920</td><td>106.6</td></tr> <tr><td>Patrick Patterson </td><td>47.3%</td><td>35.8%</td><td>48.8%</td><td>52.6%</td><td>15.1%</td><td>0.916</td><td>106.1</td></tr> <tr><td>Chase Budinger </td><td>48.3%</td><td>38.3%</td><td>50.3%</td><td>53.5%</td><td>15.8%</td><td>0.923</td><td>105.9</td></tr> <tr><td>Jordan Hill </td><td>50.3%</td><td>33.1%</td><td>50.2%</td><td>53.2%</td><td>15.1%</td><td>0.923</td><td>104.9</td></tr> <tr><td>Marcus Camby </td><td>37.7%</td><td>30.0%</td><td>39.2%</td><td>43.1%</td><td>17.8%</td><td>0.738</td><td>92.1</td></tr> <tr><td>Marcus Morris </td><td>49.4%</td><td>29.6%</td><td>48.2%</td><td>52.2%</td><td>17.1%</td><td>0.886</td><td>100.0</td></tr> <tr><td>Greg Smith </td><td>39.7%</td><td>47.1%</td><td>46.3%</td><td>47.8%</td><td>12.6%</td><td>0.849</td><td>96.6</td></tr> <tr><td>Courtney Fortson </td><td>42.2%</td><td>25.0%</td><td>41.2%</td><td>46.2%</td><td>18.8%</td><td>0.776</td><td>90.6</td></tr> <tr><td>Terrence Williams </td><td>45.3%</td><td>28.4%</td><td>44.6%</td><td>49.2%</td><td>16.1%</td><td>0.846</td><td>96.7</td></tr> <tr><td>Jonny Flynn </td><td>43.5%</td><td>43.2%</td><td>47.1%</td><td>50.1%</td><td>15.5%</td><td>0.864</td><td>98.0</td></tr> <tr><td>Hasheem Thabeet </td><td>29.6%</td><td>25.0%</td><td>30.6%</td><td>42.1%</td><td>25.6%</td><td>0.653</td><td>74.4</td></tr> <tr><td>Jeff Adrien </td><td>38.4%</td><td>30.8%</td><td>40.4%</td><td>46.0%</td><td>17.3%</td><td>0.777</td><td>85.8</td></tr></table>

(2) Who is each player playing with?

Using all the lineup data (which can be found also using the lineup finder or at basketballvalue.com), I determined how frequently pairs of players play together. For example, if Lowry is on the floor, what percentage of those possessions is Martin, Luis, etc. also on the floor with him. You can see this from the table below:

Spoiler

The cell (r,c) in the table, indicates what percentage of player r's possessions he's on the floor with player c. For example looking at the first row, it says that 74% of Lowry's possessions he's on the floor with Martin, 60% of his possessions he's on the floor with Parsons, etc. By color-coding the table, we can see some groupings of players who tend to play with eachother. Lowry, Martin, Parsons, Scola, and Dalembert form one grouping, while Dragic, Lee, Budinger, and Patterson form another. Well, there's a third (Williams, Flynn), but I'm happy to forget about them. The recent injuries to Lowry and Martin have made these groupings less defined than they used to be.


(3) What sort of stats do we expect when a player is on the floor?

Clearly, who you are playing with will impact the overall team stats when you're on the floor. Some stats try to adjust for this using regression on the full lineup data (adjusted +/-), but I'll try something simpler which ignores opposing lineups.

Based on player's season stats and the lineup statistics found in (2), I determined "expected" team stats when each player is on the floor. I make a simplifying assumption that each player's per-possession statistics should stay about the same regardless of who they are playing with. This obviously isn't perfect, since certain stats are very usage-dependent. For example, of course players playing next to a guy who shoots a lot will shoot on average less than normal. However, I think it is reasonable to "expect" floor efficiency to be about the same.

The first table shows expected stats for just the teammates the player is playing with:

Spoiler

Expected Efficiency Stats for Teammates while Player is on the Floor
<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: #B90D0D; color: #FFF; font-weight: bold;} </style> <table class="tableizer-table"> <tr style="background-color: #B90D0D; color: #FFF; font-weight: bold;"><th></th><th>2fg%</th><th>3fg%</th><th>efg%</th><th>ts%</th><th>tov%</th><th>ppp</th></tr> <tr><td>Kyle Lowry </td><td>49.2%</td><td>35.5%</td><td>50.0%</td><td>53.6%</td><td>11.5%</td><td>0.938</td></tr> <tr><td>Kevin Martin </td><td>48.9%</td><td>37.0%</td><td>50.2%</td><td>53.8%</td><td>13.4%</td><td>0.918</td></tr> <tr><td>Chandler Parsons </td><td>48.0%</td><td>36.2%</td><td>49.4%</td><td>53.9%</td><td>13.6%</td><td>0.926</td></tr> <tr><td>Luis Scola </td><td>48.1%</td><td>36.3%</td><td>50.1%</td><td>54.4%</td><td>12.3%</td><td>0.939</td></tr> <tr><td>Samuel Dalembert </td><td>48.1%</td><td>36.1%</td><td>49.7%</td><td>53.9%</td><td>13.3%</td><td>0.929</td></tr> <tr><td>Goran Dragic </td><td>47.9%</td><td>37.4%</td><td>49.6%</td><td>52.9%</td><td>10.4%</td><td>0.931</td></tr> <tr><td>Courtney Lee </td><td>48.4%</td><td>35.2%</td><td>49.3%</td><td>52.9%</td><td>12.9%</td><td>0.909</td></tr> <tr><td>Patrick Patterson </td><td>48.2%</td><td>36.2%</td><td>50.0%</td><td>53.9%</td><td>12.8%</td><td>0.925</td></tr> <tr><td>Chase Budinger </td><td>48.2%</td><td>34.2%</td><td>48.8%</td><td>52.7%</td><td>12.4%</td><td>0.910</td></tr> <tr><td>Jordan Hill </td><td>46.9%</td><td>37.0%</td><td>49.3%</td><td>53.2%</td><td>12.2%</td><td>0.924</td></tr> <tr><td>Marcus Camby </td><td>47.4%</td><td>31.0%</td><td>47.1%</td><td>50.1%</td><td>11.6%</td><td>0.874</td></tr> <tr><td>Marcus Morris </td><td>47.2%</td><td>33.8%</td><td>48.0%</td><td>50.9%</td><td>11.4%</td><td>0.879</td></tr> <tr><td>Greg Smith </td><td>46.0%</td><td>34.4%</td><td>47.6%</td><td>51.6%</td><td>11.8%</td><td>0.896</td></tr> <tr><td>Courtney Fortson </td><td>46.7%</td><td>33.1%</td><td>47.3%</td><td>49.2%</td><td>7.4%</td><td>0.896</td></tr> <tr><td>Terrence Williams </td><td>47.8%</td><td>33.0%</td><td>48.2%</td><td>52.0%</td><td>11.7%</td><td>0.899</td></tr> <tr><td>Jonny Flynn </td><td>47.7%</td><td>35.3%</td><td>48.8%</td><td>51.5%</td><td>9.6%</td><td>0.908</td></tr> <tr><td>Hasheem Thabeet </td><td>39.7%</td><td>30.0%</td><td>41.0%</td><td>45.2%</td><td>11.3%</td><td>0.781</td></tr> <tr><td>Jeff Adrien </td><td>46.4%</td><td>35.0%</td><td>48.0%</td><td>51.3%</td><td>11.6%</td><td>0.886</td></tr></table>

Just as a comment, we should expect the teammates of "shot-creators" to have a higher PPP, since they will typically not be relied on the make as many plays and will hence have lower turnover-rates. OK, next I show expected team stats for both the player in question and his teammates when he's on the floor.

Spoiler

Expected Team Efficiency Stats while Player is on the Floor
<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: #B90D0D; color: #FFF; font-weight: bold;} </style> <table class="tableizer-table"> <tr style="background-color: #B90D0D; color: #FFF; font-weight: bold;"><th></th><th>2fg%</th><th>3fg%</th><th>efg%</th><th>ts%</th><th>tov%</th><th>ppp</th></tr> <tr><td>Kyle Lowry </td><td>48.3%</td><td>36.5%</td><td>49.9%</td><td>54.2%</td><td>16.2%</td><td>0.934</td></tr> <tr><td>Kevin Martin </td><td>48.4%</td><td>36.0%</td><td>49.8%</td><td>54.2%</td><td>16.2%</td><td>0.936</td></tr> <tr><td>Chandler Parsons </td><td>48.8%</td><td>35.4%</td><td>49.8%</td><td>53.6%</td><td>15.8%</td><td>0.924</td></tr> <tr><td>Luis Scola </td><td>48.7%</td><td>36.2%</td><td>49.9%</td><td>54.0%</td><td>16.5%</td><td>0.931</td></tr> <tr><td>Samuel Dalembert </td><td>48.9%</td><td>36.1%</td><td>50.0%</td><td>54.2%</td><td>16.3%</td><td>0.932</td></tr> <tr><td>Goran Dragic </td><td>48.5%</td><td>35.8%</td><td>49.7%</td><td>53.1%</td><td>15.3%</td><td>0.917</td></tr> <tr><td>Courtney Lee </td><td>48.1%</td><td>36.9%</td><td>49.8%</td><td>53.4%</td><td>14.8%</td><td>0.925</td></tr> <tr><td>Patrick Patterson </td><td>47.6%</td><td>36.1%</td><td>49.2%</td><td>52.7%</td><td>14.2%</td><td>0.918</td></tr> <tr><td>Chase Budinger </td><td>48.0%</td><td>36.7%</td><td>49.7%</td><td>53.3%</td><td>14.8%</td><td>0.926</td></tr> <tr><td>Jordan Hill </td><td>47.9%</td><td>36.4%</td><td>49.5%</td><td>53.2%</td><td>14.9%</td><td>0.922</td></tr> <tr><td>Marcus Camby </td><td>47.9%</td><td>31.0%</td><td>47.6%</td><td>49.8%</td><td>12.6%</td><td>0.879</td></tr> <tr><td>Marcus Morris </td><td>44.9%</td><td>24.9%</td><td>43.1%</td><td>46.8%</td><td>13.7%</td><td>0.818</td></tr> <tr><td>Greg Smith </td><td>47.4%</td><td>34.4%</td><td>48.5%</td><td>52.2%</td><td>13.0%</td><td>0.904</td></tr> <tr><td>Courtney Fortson </td><td>47.3%</td><td>26.4%</td><td>45.3%</td><td>48.0%</td><td>13.6%</td><td>0.844</td></tr> <tr><td>Terrence Williams </td><td>44.9%</td><td>35.6%</td><td>46.9%</td><td>50.3%</td><td>14.6%</td><td>0.869</td></tr> <tr><td>Jonny Flynn </td><td>44.4%</td><td>32.6%</td><td>45.4%</td><td>48.9%</td><td>13.2%</td><td>0.854</td></tr> <tr><td>Hasheem Thabeet </td><td>47.5%</td><td>30.0%</td><td>46.9%</td><td>50.1%</td><td>11.3%</td><td>0.876</td></tr> <tr><td>Jeff Adrien </td><td>45.9%</td><td>35.0%</td><td>47.3%</td><td>51.0%</td><td>13.2%</td><td>0.889</td></tr></table>

(4) How do actual team stats compared to what we should expect given who each player is playing with?

To answer this question, I compare the stats from (1) with the expected stats determined from (3). In particular, I'm going to focus on the shooting and turnover efficiency stats shown in the tables above, and also only on the top 10 players in minutes played for the team.

First, I'll look purely at the teammate's stats. How much better or worse is it than what was expected?

Actual vs Expected Net Stats of Teammates
<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: #B90D0D; color: #FFF; font-weight: bold;} </style>
<table class="tableizer-table"> <tr style="background-color: #B90D0D; color: #FFF; font-weight: bold;"><th></th><th>2fg%</th><th>3fg%</th><th>efg%</th><th>ts%</th><th>tov%</th><th>ppp</th></tr> <tr><td>Kyle Lowry </td><td>-0.3%</td><td>-0.0%</td><td>-0.2%</td><td>-0.4%</td><td>-0.5%</td><td>-0.006</td></tr> <tr><td>Kevin Martin </td><td>-0.3%</td><td>-2.8%</td><td>-1.0%</td><td>-1.1%</td><td>-1.3%</td><td>-0.013</td></tr> <tr><td>Chandler Parsons </td><td>-0.0%</td><td>+1.9%</td><td>+0.6%</td><td>+0.9%</td><td>+0.1%</td><td>+0.016</td></tr> <tr><td>Luis Scola </td><td>+0.1%</td><td>+0.2%</td><td>+0.2%</td><td>+0.1%</td><td>-0.4%</td><td>-0.004</td></tr> <tr><td>Samuel Dalembert </td><td>+0.5%</td><td>+0.3%</td><td>+0.5%</td><td>+0.2%</td><td>+0.5%</td><td>-0.003</td></tr> <tr><td>Goran Dragic </td><td>+1.5%</td><td>-1.1%</td><td>+0.9%</td><td>+1.0%</td><td>-0.5%</td><td>+0.028</td></tr> <tr><td>Courtney Lee </td><td>-0.5%</td><td>+0.6%</td><td>-0.2%</td><td>-0.0%</td><td>+0.4%</td><td>+0.002</td></tr> <tr><td>Patrick Patterson </td><td>-0.4%</td><td>-0.3%</td><td>-0.5%</td><td>-0.1%</td><td>+0.2%</td><td>+0.003</td></tr> <tr><td>Chase Budinger </td><td>+0.4%</td><td>+2.2%</td><td>+0.8%</td><td>+0.3%</td><td>+0.3%</td><td>+0.004</td></tr> <tr><td>Jordan Hill </td><td>+3.1%</td><td>-3.3%</td><td>+0.8%</td><td>+0.0%</td><td>-0.4%</td><td>+0.007</td></tr></table>


And here are the net team stats for teammates and player combined. This final table tells us how much better or worse the team is than expected with each player on the floor.

Actual vs Expected Net Team Stats with Player
<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: #B90D0D; color: #FFF; font-weight: bold;} </style> <table class="tableizer-table">
<tr style="background-color: #B90D0D; color: #FFF; font-weight: bold;"><th></th><th>2fg%</th><th>3fg%</th><th>efg%</th><th>ts%</th><th>tov%</th><th>ppp</th></tr> <tr><td>Kyle Lowry </td><td>-0.3%</td><td>-0.0%</td><td>-0.2%</td><td>-0.3%</td><td>-0.5%</td><td>-0.005</td></tr> <tr><td>Kevin Martin </td><td>-0.3%</td><td>-1.6%</td><td>-0.8%</td><td>-0.8%</td><td>-1.3%</td><td>-0.009</td></tr> <tr><td>Chandler Parsons </td><td>-0.0%</td><td>+1.5%</td><td>+0.5%</td><td>+0.8%</td><td>0.1%</td><td>+0.013</td></tr> <tr><td>Luis Scola </td><td>+0.1%</td><td>+0.2%</td><td>+0.2%</td><td>+0.0%</td><td>-0.4%</td><td>-0.003</td></tr> <tr><td>Samuel Dalembert </td><td>+0.4%</td><td>+0.3%</td><td>+0.4%</td><td>+0.2%</td><td>0.5%</td><td>-0.003</td></tr> <tr><td>Goran Dragic </td><td>+1.2%</td><td>-0.7%</td><td>+0.7%</td><td>+0.8%</td><td>-0.5%</td><td>+0.022</td></tr> <tr><td>Courtney Lee </td><td>-0.4%</td><td>+0.4%</td><td>-0.2%</td><td>-0.0%</td><td>0.4%</td><td>+0.001</td></tr> <tr><td>Patrick Patterson </td><td>-0.3%</td><td>-0.3%</td><td>-0.3%</td><td>-0.0%</td><td>0.2%</td><td>+0.003</td></tr> <tr><td>Chase Budinger </td><td>+0.3%</td><td>+1.6%</td><td>+0.6%</td><td>+0.3%</td><td>0.3%</td><td>+0.003</td></tr> <tr><td>Jordan Hill </td><td>+2.4%</td><td>-3.3%</td><td>+0.7%</td><td>+0.0%</td><td>-0.4%</td><td>+0.006</td></tr></table>

Some observations from these results:

• If the measure of a good point guard is "making his teammates better", Dragic appears to fare very well here. While his teammates don't seem to shoot the 3 as well as we would expect, they do score more efficiently overall based on eFG% and ts%. Overall, he leads the team in net PPP for teammates.
• A good, in-control point guard should lessen the turnover-rates of his teammates, and Dragic and Lowry have the same net rating for teammates on this.
• Next to Dragic, Parsons also appears to excel at "making his teammates better".
• The teammates of Lowry, by comparison, appear to be slightly less efficient overall with him on the floor controlling things.

One important caveat worth mentioning again: this analysis doesn't factor in opposing lineups. If McHale tends to keep Lowry and Martin on the floor against stingier defenses, which I think he does, then their teammates could appear to play worse with them. The opposite argument, of course, could be made for some of the reserves. Its probably not an accident that the teammates of starters seem worse off than the teammates of reserves (Parsons being the notable exception).

And I should make it clear, in case it isn't, that this is not the same thing as telling us who is helping or hurting the team overall. I'll use Lowry as an example. The results show that overall the team is performing less than what is expected (using my not-so-sophisticated method) -- but what's expected includes his production. If we replaced Lowry with a league average PG, I'm very comfortable with saying the team's stats would not be as good.

The question isn't who's the best, but rather for which player do his teammates play more/less efficiently with him than without him. From this analysis, Dragic and Parsons are the two standouts.

 

Hmmm ... that phrase doesn't make sense to me. Stats are just bits of information -- "facts". And opinions should be shaped by facts.

What you refer to as knowledge is actually merely a debatable opinion. Of course, its your choice whether you want to bring facts to the table to support that opinion and possibly convince others, or whether you just want to state your opinion and let others reach their own judgment.

 

You could be right. However, as there is a considerable sample overlap between the actual stats and what went into calculating the expected stats, I didn't anticipate there to be differences exceeding a couple %s in any case. That the difference is as much as it is for Dragic and Parsons, to me, suggests that its something beyond statistical noise at least in their case.

 

That's a difficult question to answer without factoring in what I see on the court. The stats, last I checked, confirm your observation that Scola is used to much better effect at the start of games and less so as the game wears on. I don't remember this pattern existing in previous years.

Scola's quarter-by-quarter splits this season:

Code:

       [B]shots/36min    tov/36min     points/36min    TS%       PPP[/B]
1st       20.3          2.5            22.6        55.7%     0.991
2nd       17.2          2.8            18.6        54.1%     0.930
3rd       16.2          3.9            16.8        51.9%     0.836
4th       14.8          2.5            12.3        41.6%     0.711
Clutch    14.1          2.5            12.7        45.0%     0.765

League average for TS% is around 53% and for PPP its around 0.92. Clearly, Scola is struggling in second halves. I don't think it can be attributed solely to him not getting enough touches, because his turnover rate actually goes up in the 2nd and 3rd quarters as his shot attempts are reducing.

I think the Rockets play more comfortably and run more set offenses at the beginning of the game. As the game wears on, defenses start to figure out what they're doing, and their offense runs less smoothly. Maybe it just impacts Scola's ability to get good, clean looks more than others.