Analysis – Stocky vs Reality: Did your team outperform? (Pt II)

The Stocky is the main forecasting tool driving the analysis on this site. It’s a simulator of the season ahead, using the Monte Carlo method and based on Elo ratings, that gives insight into the future performance of each club. My main interest has been the number of wins, as it determines ladder positions which in turn have a big impact on the finals. The Stocky might not be able to tell you which games a team will win, but it is good at telling you how many wins are ahead.

But how does a computer simulation (in reality, a very large spreadsheet) compare to reality? To test it, I’ve put together a graph of each team’s performance against what the Stocky projected for them. Each graph shows:

  • The Stocky’s projection for total wins (blue)
  • Converting that projection to a “pace” for that point in the season (red)
  • Comparing that to the actual number of wins (yellow)

It will never be exactly right, particularly as you can only ever win whole numbers of games and the Stocky loves a decimal point, but as we’ll see, the Stocky is not too bad at tracking form and projecting that forward.

This week is Part II, from North Queensland to Wests Tigers. Part I, from Brisbane to Newcastle, was last week. Also see this week’s projections update for some errors in the Stocky.

Continue reading “Analysis – Stocky vs Reality: Did your team outperform? (Pt II)”

Advertisements

Analysis – Stocky vs Reality: Did your team outperform? (Pt I)

The Stocky is the main forecasting tool driving the analysis on this site. It’s a simulator of the season ahead, using the Monte Carlo method and based on Elo ratings, that gives insight into the future performance of each club. My main interest has been the number of wins, as it determines ladder positions which in turn have a big impact on the finals. The Stocky might not be able to tell you which games a team will win, but it is good at telling you how many wins are ahead.

But how does a computer simulation (in reality, a very large spreadsheet) compare to reality? To test it, I’ve put together a graph of each team’s performance against what the Stocky projected for them. Each graph shows:

  • The Stocky’s projection for total wins (blue)
  • Converting that projection to a “pace” for that point in the season (red)
  • Comparing that to the actual number of wins (yellow)

It will never be exactly right, particularly as you can only ever win whole numbers of games and the Stocky loves a decimal point, but as we’ll see, the Stocky is not too bad at tracking form and projecting that forward.

This week is Part I, from Brisbane to Newcastle. Part II, from North Queensland to Wests Tigers, will be next week.

Continue reading “Analysis – Stocky vs Reality: Did your team outperform? (Pt I)”

Analysis – Another bloody mid-season review (Part I)

With the conclusion of round 13, it’s half time in the 2017 NRL season. It’s the ideal time to do what everyone else is doing and look back at the season so far. This week we’re looking at the first eight clubs that come up in alphabetical order.

Part II to come next week.

Benchmarks

There are some important benchmarks to consider when looking ahead to the end of the season.

Firstly, let’s look at the regular season. I’ve tallied up the average number of wins for each position, the average for-and-against and the number of teams with a negative for-and-against for each spot on the ladder. The dataset covers 1998 to 2016, so there are some inconsistencies from seasons which had twenty or fourteen teams and where points penalties were applied to the 2002 Bulldogs, 2016 Eels and 2010 Storm.

The main takeaways are that twelve wins should get you into the finals and eighteen should get you the minor premiership. Six or seven wins will still only get you the bottom spots on the ladder (unless the 2016 Knights are playing).

Continue reading “Analysis – Another bloody mid-season review (Part I)”