On Wednesday night, the Milwaukee Bucks dropped 50 points in the first quarter against the Wizards. A few hours later, the Warriors piled up 49 in the third quarter against the Spurs. Could someone actually hit 200 in a game?
Warriors GM Bob Myers is an overzealous fourth-grader balling-up newspaper against the side of his already too-steep model volcano. He’s splashing the dish soap onto a pile of baking soda and pouring a gallon of vinegar down the cardboard barrel. The chemical reaction is out of control, now, and he’s giggling as the foamy orange magma sputters and spills over the paper-mache precipice.
Steph Curry is the squeaky-clean soap, Kevin Durant the acrid vinegar, Klay Thompson the baking-soda powder keg. They’re volatile, get it? Very explosive. What I mean is, the Warriors are a freaking volcano. And your defense is the kitchen table.
Well, last night it was the Spurs turn to be covered in hot lava. After holding the Warriors to a manageable first-half total of 66 points, San Antonio was buried by a searing deluge of points in the third quarter. The Warriors caught fire, scoring 49 points on 17-of-22 shooting in the period (including 26 points on 11 perfect field goals from Curry, Durant, and Thompson). During one astounding stretch they hit 24-of-25 shots. And, yes, the Warriors came back to earth a bit in the fourth and ended up with “just” 141 points.
But what if they didn’t?
I mean, what if they hadn’t cooled off? What if they could play an entire game like they played the third quarter last night? Is it possible that they could score 200 points in a single game? I know it sounds ludicrous, but that was their SIXTH game with 140-plus points already this season. It’s ludicrous, but is it actually impossible?
To find out, I set up a probability-based simulation of the Warriors scoring and tested it one million times to see if I could find any plausible scenario in which they would score 200 points.
In the one million hypothetical games, the Warriors averaged 120 points per game (just about what we would expect based on their 119.1 point-per-game average in real life). Half of the games fell between about 110 and 130 points (this is the so-called interquartile range or ‘IQR’); just about 10 percent of the hypothetical games were above 140 points (compared to 6 of 52 real games this season); and 10 percent were below 100 points (compared to 5 of 52 games this year).
And, lo and behold, in each of Games No. 644536, 702741, 717943, and 902399 the pretend Warriors tallied 200 points or more! In other words, there is something like a 4 in 1,000,000 chance that the Warriors will score 200 points in their next game (0.0004 percent).
Finding it hard to believe? Using the dashboard below, you can thumb through the first 10,000 simulations to get a sense for how the points add up. Just move the slider or tap the arrows to move from game to game.
Here’s the way the simulation works. You start by looking at how many plays (possessions + offensive rebounds) the Warriors usually have in a game. On average they have about 114 plays per game; but it varies from night to night, with one fast-paced contest against the Kings featuring 131 plays for the Warriors. So — using all the individual games this year — you can approximate a normal distribution for the number of plays per game (114 ± 7). Then, to start the simulation, you randomly select a value for the number of plays out of that distribution. Chances are you’ll get something like 114 or 115 plays, but if you run enough simulations you’ll eventually get something like 130 plays or, very rarely, even a bit more than that.
Next, it’s time to divy up the plays into turnovers, free-throw trips, and field-goal attempts. From there, you can further divide the field-goal attempts into 2-pointers and 3-pointers. Finally, you can parse out each of the scoring opportunities by player.
You can build out the whole simulation using real data from the team this year. What percentage of Golden State’s 3-pointers does Curry take on a given night? What fraction of the team’s free throws does Durant attempt? What is the range of possibilities for Thompson’s 3-point percentage from game to game? Come up with a distribution to describe each contributing factor, do a bunch of random sampling, and you can sketch out a hypothetical box score.
You’re probably already recognizing the fatal flaw here. The big limitation of this simulation is my assumption that the Warriors shot distribution in any hypothetical game will be a reflection of their typical shot distribution from any other game this season. However, this assumption is likely to break down in an extreme scenario like Game No. 702741. For example, if the Warriors score 150 points thru three quarters in Phoenix on Friday — would Curry really have an opportunity to jack 19 3-pointers before getting pulled out of the game by Steve Kerr? Probably not.
Now, it would be possible for me to tweak the inputs of the simulation to reflect more realistic shot distributions based exclusively on blowout games where the stars played fewer minutes; but that would probably keep us from getting to the 200-level and that was kinda the point of this whole exercise, so…
Instead, let’s just posit that the Warriors would need an exceptional opponent to push them to an extreme outlier score over 200 points. Indeed, the two highest games on record were both triple-overtime games. And maybe that’s the type of shootout that would be needed to trigger the Warriors to erupt. Let’s wait and see. By my calculations, we should be due for some real fireworks sometime in the next 3,000 seasons or so.