Opening night, and the model struggled with the picks, but went 3/3 on the Over/Under. The model is using the 3rd party projection data, so hard to know if its the model or the projections that were off.
Let us walk through the wreckage honestly, starting with Celtics at Pistons. The sims ran it 1,080 times and came out with Boston 113.2, Detroit 112.5, a coin flip leaning green. The actual: Detroit 121, Boston 113. The sims nailed the Celtics score to the decimal point and still got the pick wrong, because Detroit decided to win the opener by eight. That is the cruelest kind of miss, and it still counts.
Then 76ers at Knicks, the worst of the three. The sims had New York 114.0, Philadelphia 104.6 across 1,240 runs, a comfortable home win. The actual: Sixers 106, Knicks 100. Philly won at MSG, and New York did not even crack 100. The process read was Knicks home-opener firepower. The actual read was that the Knicks left their shooting touch in the locker room.
And Thunder at Spurs, where the sims predicted a 227-point track meet, Spurs 114.8, Thunder 112.4 over 1,320 simulations. The actual: Thunder 93, Spurs 80. A 173-point rock fight. OKC won going away, and the sims missed the winner, the margin, and the entire character of the game. One game in, the young Thunder defense looks nothing like the spreadsheet version.
The honest ledger: 0-for-3 on NBA winners last night, 3/3 on the Over/Under. Elsewhere the model did fine: 6-5 on the NHL slate, 3-0 on a quiet CFB Tuesday.
The headliner: the sims have the Lakers by nearly 20 over the Warriors
10:00 PM ET in Los Angeles, and MyGameSim is not hedging. Across 380 simulations, Warriors at Lakers comes out Lakers 121.9, Warriors 102.0. That is a 19.9-point margin in a rivalry game, and it is the boldest call on the board.
My take: the sims are betting that depth and home-court math bury the Warriors. Golden State at a projected 102 points is the model saying the Lakers defense makes every possession miserable. Golden State fans will tell you this is where Curry has his say, and they are right to say it. But 380 runs is 380 runs. When the gap is this big, the model is not predicting a game. It is predicting a tier.
Two more blowouts, and the model is not blinking
At 7:30 PM ET, Bulls at Raptors: Toronto 119.2, Chicago 102.2, a 17-point gap. Then at 8:00 PM, Jazz at Grizzlies: Memphis 124.3, Utah 106.9, a 17.4-point gap. Out west at 10:30, Kings at Clippers wraps the night with Sacramento 118.9, Los Angeles 107.0, Kings by 12.
I have followed this model long enough to know that double-digit home calls like these are its way of saying the roster talent is not close. Two 17-point calls on back-to-back tip times, plus a 12-point call to close the night. One of them will be the bounce-back statement. The other will be the reminder that October basketball is chaos. I am not naming which.
The coin flips, all four of them
Here is where the night gets interesting. Suns at Trail Blazers at 10:00 PM is the slimmest margin on the slate: Phoenix 109.0, Portland 107.5, one and a half points across 380 sims. That is a coin that lands on its edge.
Pacers at Pelicans at 8:00 PM is barely wider: Indiana 114.3, New Orleans 112.3, Pacers by two. At 7:00 PM, Bucks at Wizards is Milwaukee 111.5, Washington 108.9, Bucks by 2.6, and Hawks at Magic tips at the same hour with Atlanta 115.1, Orlando 111.8, Hawks by 3.2.
Four games, four margins of 3.2 points or less. Last night the sims went 0-for-3, and two of those were coin flips of this exact type. Tonight the math gets four more chances, and the honesty box says the model has not earned our blind trust in close games yet. It has earned our attention, which is cheaper and more fun anyway.
The undercard with real edge
Timberwolves at Heat at 7:30 PM is my favorite game of the slate: Miami 116.2, Minnesota 112.3, Heat by 3.9. A Western Conference contender walking into South Beach against a Heat team the sims clearly respect at home. This is the one I am watching.
At 8:30 PM, the Texas rivalry: Mavericks at Rockets, Houston 112.5, Dallas 106.2, Rockets by 6.3. And at 7:30, Hornets at Nets rounds out the 11-game slate at Charlotte 106.7, Brooklyn 102.8, Hornets by 3.9.
Perry line: Lakers win going away in Los Angeles, the Raptors and Grizzlies both cover their 17-point sims, and the model goes 8-for-11 on the night. The 0-for-3 opener becomes a footnote by Thursday.
All scores above are MyGameSim projections from hundreds of simulations, and none of these games have been played yet. The sims were humbled on opening night. Run your own and tell me where the model is wrong. I dare you.
