Monday, 25 January 2016

Yet More UK Data – BHGS Doubles

In the last week or so I’ve posted both my annual review of the Northern League and then added the trends from the South West DBM League to create a supra-regional picture of 15 mm ancient & medieval wargames leagues here in the UK.

The extension of my first review was made possible by John Graham-Leigh who got in touch and provided the necessary data.  In the resulting email discussion he suggested broadening the analysis to include competitions in the South Western League’s catchment area.

I looked in to this but in the end I decided that adding data from annual events would only muddy the picture.  Not least because the factors governing participation in leagues are different from those of governing annual events.  For example, the level of commitment required to attend four or five rounds of a league is different from that required to attend a weekend competition.

Simultaneously Tim Porter also got in touch suggesting I look at the BHGS Doubles League.  At first this appeared to be a great idea: a national league to sit alongside the regional leagues.  It turned out not to be such a good idea and a lesson in “data quality management”.

Before I get to the details I’d like to thank Tim for providing links to various sources of data and his patience as I struggled with the twists and turns of this competition: not least its switch from a national league to an annual event in 2010.

In the end I was able to produce data in the same format as the previous reviews:



Do not be deceived by the simplicity of the above. Although it’s arithmetically straightforward to calculate the average number of players per round for the league phase (2004-09) and compare it with those of the annual event phase (2010-15) it is really “comparing apples and oranges”.

The above graph has the effect of normalising the league and annual event phases but one could argue that participation in a multi-round league should count for more than attendance at a one-off event.

This of course begs the question of why the Doubles stopped being a league?  I am tempted to assume this was a significant drop in player demand because it fits the overall pattern.  It could however have been “organiser fatigue”, venue issues or any number of other factors

With this in mind I am very cautious of drawing any strong conclusions.  From the data what I can say is:
  • The number of ancient & medieval players at individual “BHGS Doubles” events (either League rounds or annual events) has halved since 2004.
  • Nothing has matched the popularity of DBM in 2004 nor it’s rate of decline in 2005-08.
  • Overall numbers playing ancient & medieval wargames in the events phase declined steadily from 2010 to 2014. 2015 saw a small revival but this may have been a blip.
  • DBMM numbers are essentially stable whilst FOG AM has declined 59% from its peak in 2010.
Please note none of the above draw direct comparisons between the league and annual event phases; the first point is quite specific.

Given these reservations about the data (see below for even more) these observations are remarkably consistent with those I drew from the analysis of the Northern & Southern Leagues: ancient & medieval wargaming has been in decline for a number of years.

In some ways not a lot more needs to be said, however a deeper analysis does illustrate just how difficult it is to interpret such data. What follows next is mainly for those interested in what lies behind the analysis.  You have been warned!

As I’ve hinted the preceding analysis doesn’t tell the whole story. Unlike the two regional leagues, the BHGS Doubles has undergone a number of significant changes between 2004 and 2015:
  • As mentioned above, it ran as a multi-venue league until 2009 when it switched to an annual weekend event.
  • The number of league rounds varied from 2 to 6 during 2004 to 2009 presumably mirroring player demand.
  • The ancient & medieval period shared a venue with the Flames of War Pairs (2010-11) and FOGR (2001-12) competitions and doubtless competed for table space.
The variable number of league rounds makes the player average per round difficult to interpret with any certainty.  With only two rounds the average could have been boosted as the rounds were where player demand was highest.  With more rounds, player attendance could have been spread over more rounds lowering the average.  It’s impossible to prove or disprove either.

As far as the data is concerned the average player numbers for 2004-06 and 2009 have an inbuilt error.  The compiler stuck zeros everywhere in the tables so it’s impossible to tell the difference between a zero score and a team that didn't turn up for a particular round. 15 of the 25 rounds were affected by this and the player averages for these years may have been artificially lowered by a player or two.

The spike in numbers playing FOG AM in 2010 is clearly related to the switch from national a league to an annual event but the reason behind this is not obvious.  Perhaps it was novelty or withdrawal following the demise of the other rounds.  Who knows?

Taking the above into account there’s no doubt that the data from the Northern & Southern leagues is of better quality than the BHGS Doubles data.  The latter is beset by too many shifting circumstances and data quality issues to make it a reliable time series.

I am firmly of the view that mixing the two data sets would not produce a stronger data set; rather the opposite.  It’s far safer to keep them apart and draw separate conclusions which hopefully, as in this case, agree with one another.

For those interested, here’s the base data for the BHGS Doubles:

Year Rounds DBM DBMM FOG AM FOG R FOW Total AM Total
2004 5 74.0 74.0 74.0
2005 5 58.8 58.8 58.8
2006 6 31.3 31.3 31.3
2007 2 20.5 20.5 20.5
2008 5 26.2 26.2 26.2
2009 4 29.8 29.8 29.8
2010 1 44.0 20.0 44.0 64.0
2011 1 13.0 28.0 19.0 30.0 41.0 90.0
2012 1 15.0 21.0 19.0 36.0 55.0
2013 1 16.0 16.0 32.0 32.0
2014 1 12.0 17.0 29.0 29.0
2015 1 17.0 18.0 35.0 35.0

The values are average players per league round or annual event. Finally, here’s an alternative graph showing all the data:

No comments :

Post a Comment

Comments are always welcome but this blog no longer accept anonymous comments.