Mucking about with Park Factors – ABL style

I’ve always wanted to know just how much Blue Sox Stadium kills home runs compared to Holloway Field (Brisbane) or Norwood Oval (Adelaide).

Newly armed with an R script shamelessly filched from the excellent SABR101 edX course I’ve recently finished, a free evening and an ununsual lack-of-loathing of data entry, I now have an answer.

It’s a pretty rubbish answer, because even under the best circumstances (ie, nothing on TV, way too cold to go out, smashed legs from cycling home into a headwind and a pack of excellent corn chips at my side) my lack-of-loathing for data entry is slightly shorter than one season of summary box score ABL data.

So, I’ve grabbed only the information you can see when you click onto the results of any previous games

Basic box score

Note, that when I went through and data-entried on Tuesday, this was what the summary information for each game of the day consisted of. On Wednesday, that summary information now looks like the below. I can 100% guarantee that I will not be data-entering any other previous season now.

Summary of the day

But the total 2014-15 ABL season?  That’s only 46/47 (depending on rain) games for each team.  And then you’re really comparing the home half of those games to the away half of those games – so it’s 23v23 approximately.  For reference, Fangraphs uses 5-years of MLB games (weighted towards recency)[i], so you’re looking at just over 400 v 400 in your comparison of home vs away runs/HR/whatevers.

So, as to be expected, my park factors for last year are hilariously exaggerated and in no way likely to accurately model reality – consider it more like looking at Shinjuku through the wrong end of a telescope.  But since I though other people may also enjoy them, and because I understand what I’m doing in R much better as a result of this exercise, these are what the crudely calculated and hilariously useless ABL park factors look like for the 2014/15 season.

Ballpark Home Team Park factor – Runs Park Factor – Homeruns Park factor – Hits Park factor – Errors
Barbagallo Perth Heat 118.05 63.89 113.61 93.50
Norwood Adelaide Bite 112.41 253.25 96.88 69.26
Melbourne Melbourne Aces 93.63 111.61 98.77 108.17
Narrabundah Canberra Cavalry 82.83 63.27 93.47 127.27
BISP (BlueSox Stadium) Sydney BlueSox 78.48 23.86 96.60 79.72
Holloway Field Brisbane Bandits 124.50 176.10 101.60 136.32

So these numbers are set (much like IQ) so that the average ballpark on any particular measure would score 100.  Which means, to answer my original question, it was around 10 times harder to hit a home run at BlueSox stadium than it was at Norwood (2.53/.24).  You know, with all those caveats about the sample size in play.  Regardless, congrats to the hitters of the 14 homeruns scored there during the regular season. You know who you are.

Other fun stuff to look at / think about:

  • Holloway Field leads the way as being the park where it’s easiest to make errors. I wonder why? The null hypothesis would be that playing during daylight hours in the stinking tropical heat of Brisbane has absolutely no affect on people’s ability to concentrate.  Or maybe the Brisbane scorers have very high standards and see more errors?
  • I liked that all the grounds came out roughly around average in terms of hits. I would have found it really disturbing and something that probably needed serious looking at if any of them were particularly far away from average on that.

Overall, Holloway still looks like the most entertaining for watching a game, with runs, home runs and errors park factor numbers that all promise plenty of spectator fun.

[i] And a much more complicated formula than 100* home-ratio / away-ratio.  Also, for reference, I spent about an hour trying to figure out how to get R to check that I had correctly associated ballpark and home team, pretty much the only typo/error that would be ‘easy’ to spot and correct.