My point is that those 3 games don't matter. Remove the 3 best games of any team, and everyone's margin goes down at least 1 pt (assuming +20 victory for 3 games over 60 games). The disconnect is that many people think margin equates to an absolute record. It doesn't. Margin and win percentage are all relative metrics compared to every other team in the league.
Vegas is usually a good indicator (before the betting starts). They had us at 33 wins pre-Harden, 38 wins post-Harden. Some online had us even lower. I don't recall anyone from media had us as playoff bound. They had us behind Jazz, Mavs, and Lakers, and that itself is an outstanding achievement! True! If you add all the metrics of win percentage and points differential of all teams, it will end up as 0 point differential and win percentage of .500.
Our point differential is skewed by the run and gun offensive philosophy. This graph illustrates this undisciplined system can't consistently win in close games or the playoffs.
Who's that team closest to the Rockets on the graph, right above? McHale's absence was a major disruption to the season. Some will say we did well in his absence, but coaching is not the kind of thing where you get a new coach and the results show on the same day.
So the thing I was curious about is does our style of play lend itself to an inflated point differential (i.e. we win by more points than we lose). Do up-tempo teams tend to "underachieve" by not winning as many games as their point-differential would suggest? Or maybe teams that shoot a lot of 3s or some other factors? Here is the correlation between what I'll call W-L-diff (difference between W-L% and the expected W-L% based on point differential) and various team statistics. Pace: # possession per 48 minutes ORtg: points scored per 100 possessions DRtg: points allowed per 100 possessions 3pa%: percentage of true shot attempts that were 3-pointers *-tm: four factors for team *-opp: four factors for opponent avgW-L%: average of W-L% and "expected" W-L% I split it up between two 5-year periods. I figured if correlation is somewhat consistent from one period to the next for a particular stat then its more likely that there is something real there instead of just randomness. It appears the tendency is that the better teams (according to avgW-L%) tend to exceed their point differential more than the bad teams. Opponent turnover% is negatively correlated to W-L-diff, suggesting that teams which don't force a lot of turnovers tend to exceed their point differential more than the teams that force a lot of turnovers. Neither of these factors seem very relevant for us this season. Next, I found this linear formula for predicting W-L-diff based on various team stats from 2002-2007: Code: Coefficients Intercept -0.56 Pace 0.00 3pa% -0.12 eFG%-tm -3.95 TOV%-tm 0.04 ORB%-tm -0.01 FT/FGA-tm -0.93 eFG%-opp 4.04 TOV%-opp -0.04 ORB%-opp 0.01 FT/FGA-opp 0.89 avgW-L% 0.90 The r^2 of this model for 2007-2013 was 42.4%, and it predicted whether the team would "underachieve" or "overachieve" correctly 77.8% of the time. For us this season, our W-L-diff% is -7.4%. The formula predicted it to be -2.1%.
You mean horizontally to the right side? That's Sixers of last season. Sixers, 35-31, .530, Points Differential of +4.2 Edit: immediate above is Knicks of last season, 36-30, win% .545, points differential of +3.2 (same as us).
I just ran the numbers for us and the Spurs. I took out the three best and worst games. Our differential went from +3.2 to +2.94. Theirs went from +8.1 to +8.31. So no, you are incorrect.
First off, Pythag Win/Loss is an incredibly good estimator of win probability and future playoff success. This is true not just in basketball but all sports. With that said, no tool is perfect in predicting future outcome. Pythag W/L has an error margin of about 3 games. According to their point differential, Pythag says the Roxs should be 5 games better than they are (the actual number is rounded up, it's closer to 4.5). This means the Roxs are at best 2 game below their actual records and at worst 8 games below their actual record. Secondly, let's dispel the misconception about blowouts. They have negligible effect on Pythag win/loss records. In fact, and this may seem counter-intuitive, but close games have a tremendous effect on Pythag win/loss records and can skew the analysis greatly. It's the ones that could've gone either way that define overachieving or underachieving. Roxs are: 15-23 in games by 10 points or less 12-18 in games by 8 points or less 8-11 in games by 6 points or less 3-5 in games by 3 points or less Keep in mind that there's a lot of garbage stuff that happens on the end of the game to dilute the final score. The argument that the Roxs have a lot of blow outs and that's why the Rockets seem like they're underachieving compared to Pythag W/L is not true. The Rockets are 31-11 in games decided by 10 points or more. And yes, they are 7-3 in games with 20+ margin of victory. But that just means that the Rockets are a really good team that could be winning by this margin on a more regular basis if they were more consistent. Furthermore, here's a secondary W/L analysis using Win Share as another estimator tool. According to B-Ref, the Roxs players have a combined Win Share of 38.1. See: <div class="sr_share_wrap"> <table class="sr_share" id="" style="border-collapse: collapse; font-size:.83em; border: 1px sold #aaa; overflow:auto;"> <colgroup><col><col><col><col><col><col><col><col><col><col><col><col><col><col><col><col><col><col><col><col><col><col><col><col><col><col><col><col></colgroup> <thead> <tr class="" onmouseover="" onmouseout="" id=""> <th data-stat="" align="CENTER" colspan="1" class="tooltip over_header" style="background-color: #ddd; border: 1px solid #aaa; padding: 2px;" onmouseover="" onmouseout="" onclick=""></th> <th style="background-color: #ddd; border: 1px solid #aaa; padding: 2px;" onmouseover="" onmouseout="" onclick=""></th> </tr> <tr class="" onmouseover="" onmouseout="" id=""> <th data-stat="player" align="left" class="tooltip sort_default_asc" style="background-color: #ddd; border: 1px solid #aaa; padding: 2px;" onmouseover="" onmouseout="" onclick="">Player</th> <th data-stat="ws" align="right" class="tooltip" tip="Win Shares; an estimate of the number of wins contributed by a player." style="background-color: #ddd; border: 1px solid #aaa; padding: 2px;" onmouseover="" onmouseout="" onclick="">WS</th> </tr> </thead> <tbody> <tr class="" data-row="0" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Harden,James" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/h/hardeja01.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">James Harden</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">10.6</td> </tr> <tr class="" data-row="1" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Parsons,Chandler" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/p/parsoch01.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">Chandler Parsons</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">5.6</td> </tr> <tr class="" data-row="2" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Asik,Omer" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/a/asikom01.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">Omer Asik</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">4.0</td> </tr> <tr class="" data-row="3" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Lin,Jeremy" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/l/linje01.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">Jeremy Lin</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">3.9</td> </tr> <tr class="" data-row="4" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Patterson,Patrick" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/p/pattepa01.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">Patrick Patterson</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">2.9</td> </tr> <tr class="" data-row="5" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Delfino,Carlos" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/d/delfica01.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">Carlos Delfino</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">2.6</td> </tr> <tr class="" data-row="6" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Smith,Greg" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/s/smithgr02.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">Greg Smith</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">2.6</td> </tr> <tr class="" data-row="7" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Morris,Marcus" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/m/morrima03.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">Marcus Morris</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">1.9</td> </tr> <tr class="" data-row="8" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Douglas,Toney" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/d/douglto01.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">Toney Douglas</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">1.4</td> </tr> <tr class="" data-row="9" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Beverley,Patrick" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/b/beverpa01.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">Patrick Beverley</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">1.2</td> </tr> <tr class="" data-row="10" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Motiejunas,Donatas" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/m/motiedo01.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">Donatas Motiejunas</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">0.5</td> </tr> <tr class="" data-row="11" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Anderson,James" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/a/anderja01.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">James Anderson</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">0.4</td> </tr> <tr class="" data-row="12" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Jones,Terrence" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/j/joneste01.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">Terrence Jones</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">0.2</td> </tr> <tr class="" data-row="13" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Aldrich,Cole" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/a/aldrico01.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">Cole Aldrich</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">0.1</td> </tr> <tr class="" data-row="14" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Cook,Daequan" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/c/cookda02.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">Daequan Cook</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">0.1</td> </tr> <tr class="" data-row="15" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Garcia,Francisco" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/g/garcifr01.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">Francisco Garcia</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">0.1</td> </tr> <tr class="" data-row="16" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Robinson,Thomas" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/r/robinth01.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">Thomas Robinson</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">0.1</td> </tr> <tr class="" data-row="17" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Ohlbrecht,Tim" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/o/ohlbrti01.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">Tim Ohlbrecht</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">0.0</td> </tr> <tr class="" data-row="18" onmouseover="" onmouseout="" id=""> <td align="left" class=" bold_text" csk="Machado,Scott" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick=""><a href="http://www.basketball-reference.com/players/m/machasc01.html?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">Scott Machado</a></td> <td align="right" class=" highlight_text" style="border: 1px solid #ccc; padding: 2px 3px 2px 2px; white-space: nowrap; " onmouseover="" onmouseout="" onclick="">-0.1</td> </tr> </tbody> <tfoot></tfoot></table><div id="" style="font-size: 0.83em;" class="sr_share">Provided by <a href="http://www.sports-reference.com/sharing.shtml?utm_source=direct&utm_medium=Share&utm_campaign=ShareTool">Basketball-Reference.com</a>: <a href="http://www.basketball-reference.com/play-index/psl_finder.cgi?request=1&match=single&type=totals&per_minute_base=36&lg_id=NBA&is_playoffs=N&year_min=2013&year_max=2013&franch_id=HOU&season_start=1&season_end=-1&age_min=0&age_max=99&height_min=0&height_max=99&birth_country_is=Y&birth_country=&is_active=&is_hof=&is_as=&as_comp=gt&as_val=&pos_is_g=Y&pos_is_gf=Y&pos_is_f=Y&pos_is_fg=Y&pos_is_fc=Y&pos_is_c=Y&pos_is_cf=Y&qual=&c1stat=&c1comp=gt&c1val=&c2stat=&c2comp=gt&c2val=&c3stat=&c3comp=gt&c3val=&c4stat=&c4comp=gt&c4val=&c5stat=&c5comp=gt&c6mult=1.0&c6stat=&order_by=ws&utm_source=direct&utm_medium=Share&utm_campaign=ShareTool#stats">View Original Table</a><br>Generated 3/13/2013.</div> </div> This means the Rockets should be at 38 wins, so Win Share says the Roxs are under performing by 4 wins. Note, Win Share also has an error margin of 3 wins. So at best, the Roxs are under performing by 1 win and at worst are under performing by 7 wins. However, you should treat error margins like standard deviations, meaning, extreme errors are extremely uncommon.
It's possible the teams we lose are the ones we are overconfident about and take easy. Also we have to account for back to back games. We definitely try harder against teams who are perceived as a really good team. The same mindset goes to teams who are under .500, they try hardest against us because we have a target on our back. psychology has to be accounted for
I'm not sure if I'm understanding you correctly, but if I am, this measure/graph isn't as sophisticated as you think. All it's doing is recording the point differential and W/L of every team in the past however many years to get some sort of expectation. When OKC blows out a team by 50, the only other team it effects is the team who lost. It doesn't have an effect on anyone else's point differential.
Not true. If you take all the games decided by 10+ points and change the final score so the margin is only 10, then our pythagorean expected wins drops from 38.7 to 34.2.
For the record, our 3 best games total +109 pts...and for the benefit of the previous user who misread my initial post, the Spurs' 3 best games total +108 pts.
In engineering, graphs without confidence intervals (usually 95%) or sometimes called error bars are not even looked at with any level of seriousness. R^2 values are neat in that they are embedded within Excel, but in reality are more related to the type of correlation you are trying to draw (change your regression line to polynomial instead of linear, for example). You chose to make a direct correlation, so it will skew the results. The only problem with this is that there is no serious level of analysis on how accurate or precise (two separate calculations) the model actually is in practice. Without it, there will just be conjecture on both ends by people who want to support their side of the argument.
I think this graph shows that we are a streaky team. That is usual for a talented by young shooting team. More consistency is needed.
Good work in putting all of that together. To be honest, i thought the Rockets would be a 35 win team. It's amazing that by this point differential graph they are playing like a 48-50 win team. Of course, youth and inexperience are probably the main reason why differential doesn't perfectly reflect actual results right now. These guys look like they are trying to win a championship in one game sometimes, they go all out to drive the score up and humiliate teams and then they are emotionally empty next night.
Not a chance in hell pre-Harden trade we are hovering around the 8-9 seed. The team we had prior to the Harden trade was not even close to a playoff team.
