Small Figures and Large Collections?

The recent GWS post examining four wargaming traits and their relationships (see Tinker, Tailor...) prompted some discussion on social media.  While I am not present on Bluesky, I was asked a follow-up question based upon a discussion on that platform.

From my secondhand understanding, conversations centered around the hypothesis that gamers using smaller figures tend to have larger collection sizes.  I reckon the insinuation was that the analysis was being distorted by a missing Figure Scale variable.  The question posed to me was, it this true that there is a negative correlation between figure size and collection size?  That is, as the size of the figure decreases, collection size tends to increase.  Can the survey actually test this hypothesis?

The short answer is No, not directly.

While Figure Size (Scale) is included as a variable in the survey, the variable is a ranked choice field where a respondent can rank up to the Top 3 choices.  If many of the respondents are like me, they will have figures in more than one scale.  Perhaps even many scales.  I certainly do!

To answer this question definitively, a separate question would be needed to ask each respondent the distribution of their figure collection by number of figures in each scale.  Not something many would be able to provide without a lengthy (or even impossible) counting exercise.  If those data were captured, then, yes, a correlation analysis could be produced.  Could the existing survey offer any clues or provide a proxy to help answer the question on a possible inverse relationship between figure size and collection size indirectly?

Perhaps.

For the sake of exercise, let's claim that a respondent's top choice of figure size directly reflects the largest holding in a collection.  If so, top figure size might be utilized as a reasonable proxy to compute a correlation between figure size and collection size.  We could do something similar by aggregating across all Top 3 choices too.  Will this result be meaningful to answering the question posed? Perhaps not directly but possibly a useful insight or two will emerge.

Let's begin with respondent's first ranked figure size.  Graphically, the distribution of figure size stratified by collection size is illustrated in Figure 1.

Figure 1

First off, notice that Figure 1 demonstrates 25/28mm dominates all smaller figures sizes across all collection sizes.  When 28-32mm "Heroic" size is added into the 25-28mm mix, especially at the smaller collection sizes, this market concentration is impressive.  While 28-32mm "Heroic" makes up a sizable first choice in the smaller collection sizes, 25-28mm garners nearly 40% of the first choice in collections at least 501 figures in size.  

28-32mm "Heroic" first choice percentages of total fall markedly as collection size increases.  Conversely, 15-18mm figures sizes see an increase in first choice as collection size increases.  So far, this may support the original hypothesis if all else is equal but how to account for the large and fairly stable position of the 25-28mm category?  The near monotonic tendencies seen in 28-32mm "Heroics" and 15-18mm figures sizes are not seen in the smaller figure sizes as collection size increases.  06mm, 10-12mm "Epic", and 20mm show no first-choice pattern of either increasing or decreasing as collection size increases.  From the results in Figure 1, these three figure sizes cannot reinforce the original hypothesis that smaller figures necessarily lead to larger collections.  However, it is worth noting that these figures sizes make up a smaller representation of totals in comparison to other, larger figure sizes.

What if we tackle this hypothesis from a different angle?  Past analyses have demonstrated that a gamer's primary interest drives a number of decisions and tendencies.  Primarily Fantasy/Sci-Fi gamers (values of '5' and '6' in the survey) are typically younger, spend less, have fewer years of accumulation, prefer skirmish-type games, and require fewer figures to game.  Primarily Historical gamers (values of '0' and '1' in the survey) tend to be older, have more disposable income, have many years of accumulation, tend to fight entire battles, and often field larger armies.  

Turning to a gamer's primary interest and collection size, Figure 2 shows that the percentage of those respondents having a primary interest in Fantasy/Sci-Fi decreases monotonically as collection size increases.  Again, conversely, as collection size increases, the percentage of gamers having a primary interest in Historicals tends to increase.     

Figure 2

In this case, do results suggest figure size or primary interest drives collection size?  Of course, correlation does not mean causation.

What about examining these data by bringing figure size, collection size, and primary interest together into one analysis?  To accomplish this task, multiple correspondence analysis (MCA) is called up for action to plot the relative relationships between these three variables.  Figure 3 details the initial plot.

Figure 3

As seen in other analyses using MCA, the four quadrants are marked out to see if any meaningful inferences and labeling can be made for the underlying associations.  Figure 4 shows Figure 3 with quadrants delineated to help in visualizing the results.

Figure 4

See the clustering that falls out from MCA?  Figure 5 colors codes the quadrants to ease explanation.

Figure 5

Once color coded, three clusters emerge from the quadrants.  Each labeled cluster corresponds to one of the primary interests of Fantasy/Sci-Fi, Historical, and Mixed.  The Mixed cluster spans two quadrants.  With only three primary interests defined, having three clusters emerge makes sense.  Given that figure size was based only upon a respondent's first choice, these are interesting results.  The three clusters are:

• Fantasy/Sci-Fi (blue) grouping tends to favor collection sizes of 0-100 figures and prefers 28-32mm "Heroic" figures more than the other groupings.
• Historical (yellow) grouping tends to prefer figure sizes of 20mm and under and collection sizes greater than 5,000 figures.  That is, compared to Fantasy/Sci-Fi and Mixed. 
• Mixed (green) grouping tends to hold collections from 501 to 5,000 figures and prefers 25-28mm figures when compared against the other two clusters. 

Collections of between 101-500 tend to be seen in Fantasy/Sci-Fi and Mixed more so than in Historical gamers.

Once again, the survey results produce some interesting and (I think) remarkable results.  Based upon this study, can we reject the hypothesis that smaller figures lead to larger collections?  Even using a proxy for precise allocation of the number of painted figures stratified by figure size, data and results suggest, this answer is "no" to the hypothesis question.  There is a caveat, though.  A gamer's primary interest may have a hand in driving both figure size chosen and collection size amassed.

  

Where do I fit into these clusters?  Well, I do track painting output so the question can be answered authoritatively with no need for any proxy.  I am solidly in the Historical camp and my painted collection size is in the top category at 25,001+.  Only a small percentage are 6mm.  A larger (but still small percentage) is 10mm.  The mass of 15/18mm figures makes up the majority of my collection coming in at 56%.  25-28mm takes the second spot at 23% of total painted figures.  Perhaps my entry point into the hobby helped define my preference too.  15/18mm figures were my entry point into tabletop wargaming.  25/28mm figures came relatively later. 

Figure 6

As a wargamer, do you find yourself situated within one of the corresponding quadrants present in the MCA graph?  Does the hypothesis that smaller figures lead to larger collections hold with you?

If you have yet to do so, please take the time to complete the 2025 survey.  Survey closes on 31 August.