mean ol' dressage (puns + plots)

Another favorite thing about reading so many excellent blogs is that I frequently find inspiration there. Sometimes it's about a specific riding exercise. Other times maybe a solution to a problem that's been nagging at me. Today? Today it's inspiration for a post topic.

bc without that inspiration, probably all you would have gotten is this pic. well, you're still getting it anyway, but now there's other stuff too!!! (also, isn't it cheating that those poles are fixed to the standards?!?)

You see, Megan wrote yesterday about applying statistical concepts to dressage. Specifically, she wrote about regression to the mean. If you haven't yet, go read her post.

This is already a familiar concept to me (I'm technically somewhat of a statistician by profession) so subconsciously I've always viewed my scores through that lens. As I commented on that post, this perspective helps me understand that just because we got an 8 on a canter depart in our last test doesn't mean that my canter departs are actually 8s now.

Rather, it might be indicative that I'm moving the needle in the right direction - that the mean score might be closer to 6 now as opposed to 5.

I also keep this concept in mind when looking back to my personal best dressage test ever at Loch Moy, where we scored a 26.8, and then followed it up two weeks later with my personal worst score of a 45.3 at Waredaca (ouch). Somewhere in between those scores is a perfectly respectable 35 36.

But the post got me thinking. Because of my insane compulsion to document literally every single thing, I have the data on hand from each dressage test I've ever ridden. Meaning every score from every movement. And because I do eventing dressage rather than pure dressage, there isn't a huge difference in movements from one test to another in the lower levels.

Some of the scores are combined (like the final center line is scored separately in BN A, but combined with the walk-trot transition at K in BN B). But really, it's mostly the same movements.

So, what if I looked at each individual movement to see how I've scored over time since riding my first ever dressage test in July 2014?

Well. Haha. I'm glad you asked:

Data considerations: 

• 2 USDF Intro B tests (7/2014 and 9/2014)
• 5 USEF BN A tests (from 9/2014 to 8/2015)
• 3 USEF BN B tests (from 5/2015 to 6/2015)
• excludes recent novice test (n = 1)

What story does this tell? First up, my average scores are:

• Intro B - 41.7%
• BN A - 38.2%
• BN B - 32.2%

But I didn't need to break down the scores to calculate that. So what can I learn about each individual movement? (and I'm not going to talk about Intro, but it's included because those are the first two tests I ever rode so it seems like a good starting point.)

Right off the bat we can see that our strength lies in Isabel's trot work. We enter the ring well and get the ball rolling with a nice trot circle. Test BN B involves a lot more trotting than BN A (including those two diagonals), which may be the reason our collectives are so much better. Isabel's got a fancy trot and BN B gives ample opportunity to show it off.

we need moar dressage pics! this is from Fair Hill

And if you're curious about those collectives, that peak visible in all four spark lines is Loch Moy, and the immediate dip is Waredaca, both BN B.

The most interesting numbers are the canter work. It's no surprise that's our biggest weakness. What I did *not* expect was that, according to the data, we do better tracking right. Which is exactly the opposite of what I would have hypothesized.

Waredaca

I've got some theories here tho. Our first time at Jenny Camp last year saw us cross cantering into the left lead depart in BN A. I was too slow in correcting the lead and we scored 3s on both the depart AND circle. Oops. (This would be the dip on the left side of the spark lines). 

What about BN B averages, tho? Well, both departs are in a corner - friendly enough - but the right depart comes up quickly after trotting down the diagonal, meaning it's a sharp turn immediately preceding the depart. This is a more technical challenge, and it's therefore something dressage trainer C particularly focused on in our lessons. Therefore I think we are just better schooled in that specific movement.

Ultimately I still think our right lead depart is weaker in reality than the left lead, it's just shaken out differently in the show ring. Of course the biggest joke is that the strongest part of our canter work is when we stop cantering. The worst is when we start. Lol #figures.

Loch Moy

Anyway, our walk-trot and trot-walk transitions seem to be in pretty good shape, tho a nearly 1 point difference in the free walk scores between BN A and BN B seems odd. Probably it's attributed to doing the free walk on the short diagonal in test B, giving us a better shot at faking it. Considering that's usually the coefficient score I'd like to see improvements there regardless.

Our turns up the center line and halts are generally strong. The halts surprise me because we are rarely square and often not straight. Meaning I could see some serious improvement in scores by fixing that. The center lines are another focus with trainer C and the averages look good. (the spark line for the walk trot transition to turn up center line trends downward bc the last time we rode that test, at Fair Hill, we biffed the transition, thus killing the score).

Jenny Camp 2015

So. There's some data analysis for ya. Mayyyybe not what you expected on a Saturday morning, but whatevs. This little exercise tells to me that we're quite good at trotting (look at all those 7s or almost-7s!!). We need to improve the canter. It is happening, but it's slow and painstaking work.

Perhaps the easiest ways to quickly earn points will be to polish the halts and free walk. Those already score well despite getting very little attention in schooling.

And submission... hahaha... ehhhh. You'll have to talk to Isabel about that one ;)

Anyway this was fun. Have you ever done a similar analysis of your riding? How do you decide what needs work and what's already doing well?