Monte carlo simulation from start of round 12

The model has been run a couple of hundred times and has a number of spots locked in: notably the top four - although the order of the top three depends critically on a couple of games (more of this in the next post).  Fremantle currently have a mortgage on fourth spot.  The table below shows the predicted ladder position percentage for each team (columns 1-18), the left hand side summarises GF Wins, appearances and top 8 finish percentage predictions.  The table on the right ranks the overall ladder position probability.  (the first past the post simplistic model - not shown here as it is less accurate past four rounds ranks Hawks, Geelong then Sydney Fremantle)

The bottom third is almost locked in: North, Gold Coast, Brisbane, Bulldogs, Saints, Melbourne and GWS.

The interest comes in positions 5 to 11,  The weighted percentages shown in the table on the right: Collingwood sneaks in as do Carlton (surprise at present)

The next blog will discuss the detail a bit more of these teams - after a bit more analysis particularly on the "swing" games and percentages.

The post after that will focus on the top four positions and the role of home ground advantage.  But the range of predictions has the Hawks playing in 60% of the grand finals with Sydney closely following on 57% - my gut was telling me that Geelong would be above Sydney but the flow of the fixture and the model falls in the Swans favour - perhaps worth checking the odds for Sydney getting into the GF?

The predicted number of wins to make the 8 is 12.2 (with a range of 10.5 to 14)

 

Found an error in the calcualtions - a draw between Hawks and Swans awards Swans 2.5 wins; this resulted in a top position having 22 wins despite that not being possible this year.  Seems to be the only issue but still trackign down error; the rest of the graph I am comfortable with.