Does Chorus size matter?

It's time for another fun-filled geeky analysis.  This time we'll explore some statistics around winning choruses and the number of men they have on the risers.

For a while now, some folks have put out a dream of having 100 men on the risers.  Is that realistic?  Would it make a difference?  Is that the most important thing we can do to improve the chorus?

In our survey we also indicated that we wanted to be competitive in our District and Internationally.  We'll take a look at what kind of scores and #s on the risers we would need to be competitive Internationally.

Background
All this data comes from the Barbershop Harmony Website.  There is a great link on their trivia page  http://barbershop.org/competitions/competition-trivia.html

This downloads a spreadsheet with years and years of all sorts of information on it about every chorus that has gone to international.

For this analysis, I've taken all the choruses that have competed from 2003 to 2013 (11 years worth of data)

How does Chorus Size correlate with overall Score?

First let me introduce you to the concept of a scatter plot.  Scatter plots are used to track a whole bunch of data points and see if there is some relationship.  In this case we're looking to see if there is a trend of increasing score with increasing # of men.  If it's a really good correlation, a fat line of dots will appear that is very obvious.  In a bad correlation, you'll see what is called a shotgun plot - which is just basically a whole bunch of dots that are randomly placed.

Next you draw a line through the data and see two important things:
What is the slope of the line (in the case below - it's shown in red)
What is the r^2 value (that's a fancy way of saying does it look like a line or a shotgun).  An r^2 value of 1.0 is a perfect fit, and 0.0 is complete scatter.

In the case of men vs score, there is some small - though poor correlation.  An r^2 of .3294 is pretty low - still there is something resembling a linear trend.

We also see we have a slope of 0.11   What this means is that for every additional man, you get about .1 more points on your score.  20 more guys would give you 2 more points.

There are several points I draw from this graph:

1. Large 90+ men choruses routinely score in the 90's.   I suspect there is a "chicken and egg" thing going on.  Excellent choruses with high scores attract good singers.  Good singers get high scores.  In other words success breeds growth and growth breeds success.
2. Although large choruses seem to average higher scores, there are plenty of choruses with between 40 and 80 guys scoring in the 90's too.
3. At the size we are, choruses that compete at international score between 72 and 90.  With the average competing chorus for our size scoring just over 80.
4. As we performed at contest - we are below average in scoring for our size.  Remember though, this is only looking at choruses that go to International.  I expect for the society that we are above average for our size.

Overall - The conclusion I would draw from this graph is that size helps, but it's not the major factor in how you score.  It's one of many things you can do to improve score and probably not the most important thing.

 Let's look next at where choruses place in the International based on their size.

Each year around 30 choruses compete at International.  I arbitrarily decided that in order to be competitive internationally, we would need to be able to finish in the top 11 choruses for this analysis.

The graph below shows an average number (in blue) for each place finish.  A max number of men in the chorus is shown in light gray and a minimum number of men is shown in green.

These are the conclusions I draw from this chart

1. At our present size the best we could ever hope to achieve is 6th place.  No chorus in the last 11 years our size has placed higher than that.
2. The smallest chorus to come in first had 63 men.
3. The average number of men for choruses finishing 5th to 11th is 60 men.

Overall - The conclusion I make from this data is that 60 guys seems to be some kind of threshold.  It's possible to score 3rd at our size, but very rare (KVU did that in 2012 with 39 guys). 

What would we need to score to be competitive?

Below is a High-Low chart.  It's often used for things like temperature, where you show the average temperature on a day and then the high temp at one end of the line and the low temp at the other.  It tells you the range you can expect on any given day.  In this case we're looking at the range of scores possible for each place at International.

So, if we went to International contest and scored a 76.3, the best we could hope for would be to place 19th or 20th - we probably would place much lower than that.

If we want to come in at the top 10, we would need around 85, though some years choruses have come in 10th as low as 84 and as high as 88.

Overall  - I think we need to score over an 85 at a local convention before we consider going to International.  We've got some work to do.

Lastly, how often do really good choruses go to International?

Below is a chart showing chorus name and number of times they've gone to International in the last 11 years.  I only included choruses that have finished at least once in the top 10.  Remember that if you come in 1st, you can't compete for the next 2 years, so choruses like Ambassadors of Harmony, Vocal Majority, Masters of Harmony and Westminster will show up less than you would think.

Some of these choruses are pretty hard core about going to International.  But how does going to International a bunch affect your score?  Remember the first chart - where I described what a shotgun plot looked like?  Well, it turns out that there is no correlation between how often you go to International and the average score you get.  None.

I believe if I looked into how each individual choruses improved over time, we would see some sort of continuous improvement trend, but that's just a guess.

So - where does this leave us?

• Does the chorus need to grow to be competitive?
• In the scheme of things - is chorus size a major thing to focus on?  or a minor thing?

The data doesn't give us a completely clear answer on these questions.  I believe that we should focus on good singing and doing excellent shows and have targeted recruiting to good singers (hopefully ones with adolescent relatives that love to sing too).  If we sing and perform well - numbers will increase naturally.

Hope this analysis fosters communication as we prepare our goals for the chorus moving forward.

Great singing with you.

Brian Schuck

Chorus Contest analysis

Hey Chorus!

I really enjoy analyzing stuff, so below is some analysis of our latest contest score in relation to the past and other choruses in our district.  If you’re not into geeky analysis, don’t bother reading.  In short it says we should be very proud of how much we’ve improved.

It’s also hoped that this data can be useful as we decide where we want our chorus to go and what we want it to be.  I believe that with good information, good decisions can be made.

Summary conclusions

• Jumping from 67 to 76.3 in 2 years is phenomenal improvement
• There are choruses that score above 80 with the same number of guys we have.  It's possible, however it looks like it's easier to do with more on the risers.
• The main difference between us and the group that came in 3rd was in the Music category.

Over the past 5 years, 17 different chorus’s have gone up on Contest stage to compete in the Cardinal District.  Some have gone just one time, and others almost every time.  If you look at the average scores for each chorus (excluding Saturday’s), the Cavemen Chorus is the 6^th best chorus in the Cardinal District.   See graph below

In general, chorus’s score pretty close to how they’ve scored in the past at Contest.  That’s why it’s remarkable that we were able to move from 6^th place to 4^th place.

Also note:
·         All top 6 chorus’s competed.
·         We were very close to getting a better score than the Louisville Thoroughbreds and coming in 3^rd place.

It’s very illustrative to look at how much we improved over our average compared to the other top performing chorus’s in the District.

Also note that 5 out of 6 of the chorus’s are improved vs. their historical performance.

In fact, if the Thoroughbreds hadn’t improved by 3.2 points over their historical average, we would have come in 3^rd place easily.

It’s very rare to have that kind of jump in performance.  Our jump is double what the other top choruses have done in the recent past.

Other district comparions

Five other Districts have posted their Fall contest scores.

Cardinal District is an above average District.  We would have placed in the top 3 or better in another District:

Also of note is the location of which chorus’s scored better than us in this year's Fall Contest.

In general, larger metropolitan areas score better than us.

Having 50-60 guys on the risers makes a big difference too.

In the past, Louisville has had 38-40 on the risers, they increased that by 46% for this last Contest.

In contrast, we have also increased our chorus size by 47% since the last couple Contests.

Below is a graph showing a very loose correlation between Score and number on the risers.  In general, every person adds 0.2 to the score.  This correlation held up for the Thoroughbreds, they added 17 people and raised their score by 3.1 which is 0.18 points per person added (pretty close to 0.2)

By that logic, in order to score an 80.0, we would need to have 18 more people – 52 on the risers.  Now, I think we know it’s not just quantity of people, but a lot of other factors that influence the score.  Indeed, by that logic we should have only scored a 69 this time.  I credit our massive increase to more people and significantly better craft.  Nevertheless, more people on the risers would allow the chorus more flexibility and depth.

I looked through the 2012 Fall Contest results and there were 7 chorus’s that scored 79 or greater that had less than 40 members.  It’s possible to get high scores without  huge numbers on the risers, but the larger chorus’s invariably score higher.  Nevertheless, we ought to be able to score 80 or above just with the guys we've got!

Below are our 2013 results in detail:

What I note from this is the following:

• We were consistent across all 3 categories.  The spread from min to max was just 5 points 151 - 156
• We scored best in Singing Bright was the Night
• Our lowest scores were in Music for both songs
• I thought our presentation scores would be our opportunity (guess we looked pretty good).

Now, if you're like me, you might be confused about the difference between the categories.  Here is the quick definition straight from the judges handbook.

You can see that in Presentation and Singing we were pretty much tied with Louisville, they beat us in Music.

Brian's opinion on direction

I think it's important we decide what kind of chorus we want to be.  Getting super high scores may not be our goal.  I really enjoyed watching KVU and Circle City perform their sets, but the amount of work to coordinate that level of choreography and set/costume design must have been enormous.  Personally, I'd rather just sing well and leave the dancing and back flips to others.  If that means we don't win for the District, so be it.  I do believe that perfecting pieces for Contest helps make the other 15 pieces we sing throughout the year better, so going to District Contest is a good thing, it's just not the only thing.

Overall, we should be very proud of our accomplishment.  It’s great to be associated with all of you.

Brian Schuck