I received a question yesterday about NCAAGameSim.com and NBAGameSim.com and thought others might find the response beneficial.

I honestly want to know why there are some teams that still lose the game even though the simulation favors them? What are the factors behind this? And how should I rely on simulations to determine the outcome of the game?

Here was my response.

Check out this graph from our site, this is something that our Subscribers have access to.  This is for tonight’s Lakers/T-Wolves game.

Think about any given simulation as a series of probabilities.  There are some results that could happen but are unlikely to happen, those are the ones on the outer edges of this graph.  Lakers might win by 50, Timberwolves might win by 40.   Both are unlikely, but both are technically possible.  More likely is what is in the middle of the graph and the peak often lines up with what Vegas predicts (the red dotted line).   As more simulations are run, this graph will turn into a perfect bell curve highlighting all the probabilities of the results.

The factors are –

  • Shooting percentages, one team or the other may make or miss everything
  • Rebounding, one team may grab all the rebounds
  • Turnovers, one team might be super sloppy with the ball
  • Etc
  • Etc

It’s common to look at the unlikely scenarios and say this doesn’t make sense.  What you want to find as you simulate more and more games is what that middle portion of the graph looks like.  That is what is most likely to happen. Hope that helps.

Hopefully this makes sense, and if you see a 16 seed win a simulation over #1 Gonzaga then this explains why.  It's unlikely, but people also thought it was unlikely UMBC would beat Virginia.