SIM-ply Put: Duke Will Beat Purdue - Game Simulation -

SIM-ply Put: Duke Will Beat Purdue - Game Simulation

by WebMaster

Posted: 3/26/2010 4:50:39 AM

This is not meant to be an opinion piece, but a look at a simulated game between Duke and Purdue.  This article is the introduction of our new Game Simulator.  It is still only days old, so if you find oddities let us know.  The Game Simulator gets a game tempo from the two teams and plays out a game using the two teams' tendencies. 

Want more specifics?  If so, keep reading or otherwise skip this paragraph.  In seconds, we play out the game possession by possession.  First, we have to understand that a possession can end in a field goal attempt, turnover, or free throw attempt.  We then determine how often a given team does each by creating 3 buckets (of different proportional sizes).  After generating a random number we find out which bucket this possession will fall into.  We then continue analyzing the possession.  For example if a field goal attempt, we determine who would the shooter be?  Are they more likely to shoot a 2 or 3?  For that type of shot, how effective are they?  If a miss, was it an offensive or defensive rebound?  Etc. Etc.

After running a few simulations, I was quite pleased with the numbers I was seeing.  That all said, lets use our new feature to help forecast the game tonight between Duke and Purdue.

FINAL SCORE -- DU (2010): 70 - Purdue (2010): 61

Matchup Simulation Stats

  • This game has been simulated 20 times.
  • DU (2010) has won 19 times (95.0% of the time)
  • Purdue (2010) has won 1 times (5.0% of the time)
  • The average score is DU (2010): 70.6 - Purdue (2010): 54.8
  • The game has gone into overtime 0 times (0.0% of the time)

Recent Articles from WebMaster

Recommended Articles

mehmattski   (Not Registered)
Some Suggestions - 3/26/2010 2:11:11 PM - Login Verification: None

I assume that the simulations are based on the season-long efficiency statistics, and account for a player's %Min and %Poss ratings. If so, there should be ample data in the last ten games to calculate the simulation without needing to rely so heavily on the full-season data. Robbie Hummel and Olek Czyz are obvious examples, as they will clearly not play tonight. But, for example, Brian Zoubek has had his role increase over the last 15 games, and I would like to see that reflected in the simulation. Jon Scheyer's shooting slump could also be accounted for. I think the simulations could be improved if recent results were more highly weighted. Very interesting concept, though!

~Matt Johnson

Webmaster   (Not Registered)
re: sim suggestions - 3/26/2010 4:08:11 PM - Login Verification: None

Matt, thanks for the suggestions, I'll look into this and see what can be done.

ninjaflexx   (Not Registered)
awesome - 1/11/2014 3:46:52 PM - Login Verification: None

 dude this thing is AWESOME!

i use i all the time

Post a Comment

Posting As:   Remember Me?

Before submitting this form, please type the color of the second character:

SCACC Hoops has no affiliation to the NCAA or the ACC
Team logos are trademarks of their respective organizations (more/credits)