Hollinger's 2010 Playoff Odds

[davemonsterband]

C's are old and if KG, Truth, Ray or Rondo (who gets incessantly hammered by everyone) get hurt, even individually, I think it's an automatic no-go. I agree with those odds. I wish I didn't.

[nickagneta]

To be honest....it's been 20 years since I was in a statistics class, so I might be rusty.  Also, our last discussion was based on what certain stats meant regarding their interpretation to real time basketball. Put 20 coaches with their personal stats experts in a room and you'll probably get 20 interpretations of the "sabermetric" basketball stats and what they mean.

Oh I understand trust me, if you don't use it you lose it. After two years away from school I forgot so much... As for NBA GMs and Coaches, we both know how dumb they are! Look at Jared Jeffries, Diop's, and Gilbert Arenas's contracts!

I say we table this discussion until we see some more "playoff odds". If my theory about the regression to the mean factor is right it should be fading out of significance soon. TP for the debate.

TP4U2 Faf! Good intelligent discussion as always.

[Fafnir]

To be honest....it's been 20 years since I was in a statistics class, so I might be rusty.  Also, our last discussion was based on what certain stats meant regarding their interpretation to real time basketball. Put 20 coaches with their personal stats experts in a room and you'll probably get 20 interpretations of the "sabermetric" basketball stats and what they mean.

Oh I understand trust me, if you don't use it you lose it. After two years away from school I forgot so much... As for NBA GMs and Coaches, we both know how dumb they are! Look at Jared Jeffries, Diop's, and Gilbert Arenas's contracts!

I say we table this discussion until we see some more "playoff odds". If my theory about the regression to the mean factor is right it should be fading out of significance soon. TP for the debate.

TP4U2 Faf! Good intelligent discussion as always.

I say we just go make a thread and complain about Suh getting robbed, then we can argue the same side for once!

[nickagneta]

To be honest....it's been 20 years since I was in a statistics class, so I might be rusty.  Also, our last discussion was based on what certain stats meant regarding their interpretation to real time basketball. Put 20 coaches with their personal stats experts in a room and you'll probably get 20 interpretations of the "sabermetric" basketball stats and what they mean.

Oh I understand trust me, if you don't use it you lose it. After two years away from school I forgot so much... As for NBA GMs and Coaches, we both know how dumb they are! Look at Jared Jeffries, Diop's, and Gilbert Arenas's contracts!

I say we table this discussion until we see some more "playoff odds". If my theory about the regression to the mean factor is right it should be fading out of significance soon. TP for the debate.

TP4U2 Faf! Good intelligent discussion as always.

I say we just go make a thread and complain about Suh getting robbed, then we can argue the same side for once!

We were on the same side in a lot of this summer's current events threads. [dang] those Republicans!!!

[fairweatherfan]

Update from Hollinger's Twitter (posted on ESPN.com):

"Quick update -- looks like we have issues with the hamster that generates odds for making and winning the Finals. We're working on it."

Sounds like I wasn't the only one to email.  So it is some error in the model - it didn't make much sense why it kept turning out that way. 

[nickagneta]

Update from Hollinger's Twitter (posted on ESPN.com):

"Quick update -- looks like we have issues with the hamster that generates odds for making and winning the Finals. We're working on it."

Sounds like I wasn't the only one to email.  So it is some error in the model - it didn't make much sense why it kept turning out that way. 

Ahhh. Vindication. TP for the update.

[nickagneta]

Update:

The Celtics won on the road last night and had their chances of winning the championship fall to 4.9% or lower than the Spurs and only slightly better than Utah and Oklahoma City that both have a 4.5% chance of winning it all.

There has to be something wrong with the model if, given the input, the output continues to given such results.

[Kwhit10]

Update:

The Celtics won on the road last night and had their chances of winning the championship fall to 4.9% or lower than the Spurs and only slightly better than Utah and Oklahoma City that both have a 4.5% chance of winning it all.

There has to be something wrong with the model if, given the input, the output continues to given such results.

The Lakers odds went up too I think even after that last loss. I would really like to hear the 'official' explanation from Hollinger.

*edit*  I see from that post he had problems with the computations.

[RAcker]

Hollinger could not have passed calculus.  Why doesn't he just stop already?

[4THQTR]

TP to faf and nick for a good discussion!

TP for fairweatherfan for the investigation as well!

now i'm really curious if the odds change and to what extent...

(and i generally don't care much about such things...)

[colincb]

- Hollinger’s power rankings that are the playoff-odds model's primary input are flawed in that they use total strength of schedule (“SOS”) rather than adjust it for the record of one's opponents on the road vs when they are home. As a result, with such an unbalanced home schedule as the Lakers have, the Lakers' SOS is misleading. If one adjusts their SOS accordingly, then LA's SOS is slightly higher than the C's and the Lakers' power ranking would a couple of points lower. The Lakers have played better teams, but they have played them at home.

Obviously, the Lakers schedule stands out, but other teams have unusual schedules too (e.g., the Suns). The model is biased even without knowing how adjustments are made to "regress to the mean". The Power rankings that are the primary input are flawed for a small data set that includes unusual outliers due to schedule imbalances. By season's end, this particular fault should be substantially diminished.

- One possible adjustment to "regress to the mean" that would hurt the Cs might be made to their road record. 12-1 puts the Cs on a record-smashing pace of 38 road wins. The current record is 33 wins piled up by the Bulls 72 win team. It would not be unreasonable to expect that the Celtics cannot win 38 road games as a simple projection of current road win% multiplied by 41 road games in a season. No one else has come close to 38 road wins. OTOH our home record is good, but nothing special and our SOS is low. Thus it's likely that if you adjust our road record downwards, then the Cs don't perform as well under the model as one might expect. How ESPN does this or even if they do is anyone's guess.

[lostjumper]

I'm not a huge proponent of odds at this time of year, and especially this year where the top 4 in the Eastern Conference seem al but set barring a major collapse/slew of injuries. The top 4 will be, not necessarily in this order:
Celtics
Cavs
Orlando
Atlanta

The Eastern conferences semis are going to have some great matchups, and I can't even pick right now who would be the best matchup for the C's. The road to Banner #18 is going to be tough if it requires say Atlanta, then the Cavs, then the Lakers right in a row.

[nickagneta]

I'm not a huge proponent of odds at this time of year, and especially this year where the top 4 in the Eastern Conference seem al but set barring a major collapse/slew of injuries. The top 4 will be, not necessarily in this order:
Celtics
Cavs
Orlando
Atlanta

The Eastern conferences semis are going to have some great matchups, and I can't even pick right now who would be the best matchup for the C's. The road to Banner #18 is going to be tough if it requires say Atlanta, then the Cavs, then the Lakers right in a row.

On the other hand the C's had to go through Atlanta, Cleveland, Detroit and LA in 2008 and that wasn't exactly a cakewalk. but the were successful. So that gives me confidence.

[fairweatherfan]

- Hollinger’s power rankings that are the playoff-odds model's primary input are flawed in that they use total strength of schedule (“SOS”) rather than adjust it for the record of one's opponents on the road vs when they are home. As a result, with such an unbalanced home schedule as the Lakers have, the Lakers' SOS is misleading. If one adjusts their SOS accordingly, then LA's SOS is slightly higher than the C's and the Lakers' power ranking would a couple of points lower. The Lakers have played better teams, but they have played them at home.

While SOS is blind to home-away, the model does adjust for home vs away when considering margin of victory - a 3.5 pt "home advantage" is added to home games, and vice versa for road.  So, all else being equal, if the Lakers win a bunch of home games by an average of 10 pts, and we win the same # of road games by an average of 3, the model considers those performances to be equally impressive.

Explanation of model

[fairweatherfan]

Holy cow, check the playoff odds now! Looks like they definitely got the kinks out.