“Now it is such a bizarrely improbable coincidence that anything so mind-bog-gglingly useful could have evolved purely by chance that some thinkers have chosen to see it as the final and clinching proof of the non-existence of God.
The argument goes something like this: `I refuse to prove that I exist,’ says God, `for proof denies faith, and without faith I am nothing.’
`But,’ says Man, `The Babel fish is a dead giveaway, isn’t it? It could not have evolved by chance. It proves you exist, and so therefore, by your own arguments, you don’t. QED.’
`Oh dear,’ says God, `I hadn’t thought of that,’ and promptly vanished in a puff of logic.
`Oh, that was easy,’ says Man, and for an encore goes on to prove that black is white and gets
himself killed on the next zebra crossing.”
-Douglas Adams The Hitchhiker’s Guide to the Galaxy
Let’s recap.
We’ve spent a lot of time together on this blog looking at what makes a team successful in the playoffs. We’ve seen some very interesting patterns. I’ve done some complicated math. An at the start of the playoffs, I decided to make some picks using what we know (or believe we know). Because to quote Feynman( The quote is from “Simulating Physics with Computers”):
I want to talk about the possibility that there is to be an exact simulation, that the computer will do exactly the same as nature. If this is to be proved and the type of computer is as I’ve already explained, then it’s going to be necessary that everything that happens in a finite volume of space and time would have to be exactly analyzable with a finite number of logical operations. The present theory of physics is not that way, apparently. It allows space to go down into infinitesimal distances, wavelengths to get infinitely great, terms to be summed in infinite order, and so forth; and therefore, if this proposition is right, physical law is wrong.
So an accurate simulation is possible just not very easy.
Let’s Review the relevant facts first:
It’s about to get math intensive here so if it’s your first visit (or if you need to bone up) that’s what the Basics are for.
The NBA Playoff Preview and the Picks (Part 1: Just the Facts) Breaks down all the Maths and theories in detail
with the visual summary here:
The Playoff Checklist for all the Factors:
And The Team Vitals for everyone:
The Team Vitals (Top 6 Approximation):
The NBA Playoff Preview and the Picks (Part 2: Skin in the Game) Breaks down all the picks.
How did I do?
Let’s see:
Eastern Conference Round 1
Series 1: Chicago vs Indiana: Win % @ Neutral Site: >70% for Bulls
The Numbers said:Bulls in 5
Home Team Win Margin | 70% ++ |
W4 | 23% |
W5 | 34% |
W6 | 18% |
W7 | 15% |
L4 | 1% |
L5 | 1% |
L6 | 4% |
L7 | 4% |
My Prediction: Bulls in 4.
Result: Bulls in 5 (most likely predicted scenario at 34%)
The Numbers and I both had the Bulls right. I had the Lenght of the series wrong (the numbers did not).
Series 2: Miami vs Philadelphia Win % @ Neutral Site: just about 70% for Miami
The Numbers said:Heat in 5
Home Team Win Margin | 70% |
W4 | 23% |
W5 | 34% |
W6 | 18% |
W7 | 15% |
L4 | 1% |
L5 | 1% |
L6 | 4% |
L7 | 4% |
My Prediction: Heat in 5.
Result: Heat in 5 (most likely predicted scenario at 34%)
Everyone got this right.
Series 3: Boston vs New York Win % @ Neutral Site: I took 65%
The Numbers said: Celts in 5
Home Team Win Margin | 65% |
W4 | 17% |
W5 | 30% |
W6 | 18% |
W7 | 18% |
L4 | 1% |
L5 | 3% |
L6 | 7% |
L7 | 6% |
My Prediction: Celts in 6.
Result: Celts in 5 (4th most likely predicted scenario at 17%)
Got the series right but missed the length. The numbers came close to getting this one perfect too (and iffy call in game 1).
Series 4: Orlando vs Atlanta: Win % @ Neutral Site:About 68% for Orlando if I’m really,really nice to the Hawks and ignore their post All-Star collapse.
The Numbers said:Magic in 5
Home Team Win Margin | 68% |
W4 | 20% |
W5 | 32% |
W6 | 18% |
W7 | 16% |
L4 | 1% |
L5 | 2% |
L6 | 5% |
L7 | 4% |
My Prediction: Magic in 5.
Result: Hawks in 6 (5th most likely predicted scenario at 5%)
How do you outscore your opponent by almost two points a game with the best player on the court and still lose a series? Ask the Magic, they just did it.
Western Conference 1st Round:
Series 1: San Antonio vs Memphis Win % @ Neutral Site: 54% for Memphis.
The Numbers said: Grizz in 6
Home Team Win Margin | 46% |
W4 | 4% |
W5 | 12% |
W6 | 11% |
W7 | 18% |
L4 | 8% |
L5 | 13% |
L6 | 21% |
L7 | 14% |
My Prediction: Grizz in 7.
Result: Grizz in 6 (Most likely predicted scenario at 21%)
Again, me and the numbers both nail the result but the numbers get it spot on (I’m guessing you’re sensing a pattern here 🙂 )
Series 2: LA vs New Orleans Win % @ Neutral Site: 65% Lakers.
The Numbers said: LA in 5
Home Team Win Margin | 65% |
W4 | 17% |
W5 | 30% |
W6 | 18% |
W7 | 18% |
L4 | 1% |
L5 | 3% |
L6 | 7% |
L7 | 6% |
My Prediction: LA in 5.
Result: LA in 6 (2nd Most likely predicted scenario at 18%)
Never underestimate Chris F#$%ing Paul. I thought (and so did the numbers) he’d get one but he got two.
Series 3: Dallas vs Portland Win % @ Neutral Site: 60% Dallas.
The Numbers said: Dallas in 5
Home Team Win Margin | 60% |
W4 | 12% |
W5 | 25% |
W6 | 18% |
W7 | 20% |
L4 | 2% |
L5 | 4% |
L6 | 11% |
L7 | 8% |
My Prediction: Dallas in 7.
Result: Dallas in 6 (3rd Most likely predicted scenario at 18%)
Brandon Roy and the freaky comeback made the numbers wrong. I’ll take great basketball over any model any day of the week.
Series 4: OKC vs Denver. A coin flip my guess was 55% Denver based on the Altitude.
The Numbers said:Denver in 6
Home Team Win Margin | 45% |
W4 | 4% |
W5 | 11% |
W6 | 10% |
W7 | 17% |
L4 | 8% |
L5 | 13% |
L6 | 22% |
L7 | 14% |
My Prediction: Denver in 6.
Result: OKC in 5 (5th Most likely predicted scenario at 11%)
I spent hours on this series. I ran every scenario in my head. I thought Homecourt advantage would be the decider (in Denver’s favor). I forgot about the star system (i.how stars get calls differently in playoffs see here). That and this little ditty:
What exactly was George Karl Thinking?
To recap:
6-2 for Round #1 for Picking the winners (Arturo and the Numbers)
Numbers got the length right as well for 3 of the series (I only got one). The lessons as always kids, trust the numbers.
Got the coin flip wrong . Missed Orlando not showing up.
Let’s talk round 2 now.
I had:
- Chicago over Orlando in 7.
- Miami over Boston in 7.
- Denver over Memphis in 7.
- LA over Dallas in 7.
And for the rest of playoffs I had Miami over Chicago in 6. Denver over LA in 6 and Miami over Denver in 5.
I obviously need to review these in detail.
First the Numbers for the remaining teams (Note: I had an error on this table initially. Fixed now. Did not affect anything else.):
The really nice part is that now I have the actual minute allocation each team has used for the playoffs. I combine that with the raw productivity (ADJP48 not Wins Produced because having a big as your sixth man? In general better than having a small).
With that data, I can do this:
Which is how each team projects based on playoff minute allocation and their ADJP48 numbers.
If I use that to project the matchups I get:
My logic is as follows:
- Chicago is clearly better than Atlanta but they play down to the level of their competition and are banged up.
- Miami skews better than Boston when we look at overall but that doesn’t show head to head. This is a coin flip series. If Shaq plays more than 40 minutes and is effective? Throw out that number and prepare for Banner 18.
- Dallas matches up well with LA. Dallas was also the better team down the stretch. I also don’t trust Bynum’s knees. I go with the average number here.
- I always felt the winner of OKC-Denver would win Round 2 against Memphis and beat the winner of the battle of the oldsters (LAL-DAL) to go to the Finals. Love Memphis but OKC is clearly better.
Using the numbers I ran the math and this time I’m sticking to it. The math looks like so:
And the 2nd Round Picks are:
Chicago in 4 over Atlanta (Say goodbye to the Hawks!)
Miami in 7 over Boston (I hope to god I’m wrong here. Or better yet, pray for Shaq over Baby!)
Dallas over LA in 6 (I could totally see the Mavs getting screwed by the Refs though)
OKC over Memphis in 7 (Vancouver-Seattle is going to rock!)
For the rest, again just guessing, OKC over Dallas in 6 (Kidd gets abused by Westbrook) , Miami over Chi in 6 (Superstars get the calls, Bosh goes off on Boozer) then Miami over OKC in 7 in what could be an epic first part in a series.
I’m soooo looking forward to the rest of the playoffs.
Kyle
05/01/2011
Arturo- why are zbo’s numbers showing he was awful in the reg season and post all-star???
arturogalletti
05/01/2011
Kyle,
Good catch. It was an error on the table that I had fixed prior to doing the other calculations. Somehow I forgot to update the table. Fixed now.
Any requests?
Devin
05/01/2011
Oh god – not only is that Big Country in that picture, but I think that the other player is Greg Foster. Brings back some terrible memories of him in Toronto.
ilikeflowers
05/01/2011
So just going by the numbers that looks like 1 overwhelming favorite, 1 mild favorite, and two tossups.
arturogalletti
05/01/2011
Ilf,
Yeah. Extremely good playoffs this year. I do think we may be seeing Marc Gasol make the leap in the playoffs and if that’s the case? Memphis may just go all the way.
EntityAbyss
05/01/2011
Honestly, I’m rooting for the lakers so Miami dismantles them in the finals. I say Bynum stays healthy, and LA’s homecourt takes them to the finals. 🙂 The Shaq prediction ignores the fact that Haslem has a higher chance of playing before Shaq. If Miami gets Haslem and he’s productive… Shaq won’t mean much. Then Miami beats up Chicago and… my dream match happens.
For the good of all basketball, Lebron James has to destroy Kobe in the finals, so it’s best to hope for the best with Bynum’s knee.
ilikeflowers
05/02/2011
I doubt that the data supports this, but I think of Dallas as chokers. I don’t know if the players that they’ve added will change this. So given that they’re the road team, and that I think the Lakers will be better than usual since Kobe is injured (therefore deferring to the bigs more out of necessity), I gotta go with LA in 6.
BPS
05/02/2011
I suppose it’s well and good to say that you went 6-2 on picking winners, but that isn’t really what you’re doing here; you have a model for predicting outcomes, and to put probabilities on teams winning. That’s substantially different, especially when you want to evaluate your model.
Just as a crude oversimplification, you picked 5 heavy favorites in round 1: Chicago, Miami, Boston, Orlando, and Los Angeles. However instead of just picking them, you put odds on all of them; and the odds you gave implied a slightly better than 50% chance that at least one of these matchups would end in an upset (~49.1% chance that all 5 favorites won).
So while you can say that you picked 4 of the 5 winners in these matchups (and be correct), you actually made a more provocative claim, that a major upset in the first round was even money – and seeing that upset (small data sets aside) supports the model.
I’d be more worried if your model is showing significant chances of upsets, and you never see them; if the 9:1 favorites always win, your model isn’t calibrated properly.
arturogalletti
05/02/2011
BPS,
Yep. What’s even more interesting is that there’s a 85% chance that Mia,LA or OKC will lose.