A notion on testing martial arts techniques

[BostonRedBaron]

So while on vacation with some friends, I got into a discussion about what it means to test oneself with regards to martial arts skill/technique. As a professional test engineer, I tried to explain some basic test engineering methodology that I go about testing software/hardware on the job and how to relate that to martial arts. So while I don't write/blog/post too much on martial arts as I spend most of my day writing test cases, results etc, I actually kinda like what I wrote/discussed and would like some feedback on it. Also, I do need some better examples as I'm a little light on a couple of them. I'm hoping this prompts some intelligent discussions so I thought I'd throw it up here and see what people think. I hope someone finds this useful. http://www.theredbaron.net/RedBaron/Reds_Blog/Entries/2011/9/17_Reproducibility_In_Martial_Arts.html FYI, I put it up on my blog as its too long to post here on the forum. I hope this doesn't violate some forum rules about posting links.

​

 

[clfsean]

Go see my buddy Jeff Burger there in Boston... he'll help you figure it. ;-)

 

[BostonRedBaron]

I know lots of people don't favor forms training too much, but I personally like the way it gives me a CLF aftertaste...

Trains at BTT. I know some guys that train down there. A quality gym.

 

[clfsean]

Trains at BTT. I know some guys that train down there. A quality gym.

I think. I'm not 100% where he spends his training time, but I'm most familiar hearing about his KICK program.

But he can give you first hand experience in pressure testing MAs from a variety of stand points & situations.

 

[Jenna]

A cool idea I think what you are proposing is entirely doable as an exercise in martial science. I think though that there are a colossal amount of variables which either need to be accounted for, averaged or assumed out, otherwise you may end up with a set of results which are so specific as to negate their worth.

I think it is equivalent to testing your server (or whatever) to cope with one piece of code running from one user. That, as you know is not a test from which you can draw conclusions that can be extrapolated across the entire system. You must have more realistic conditions in order to satisfy your quality assurance protocols.

In this case, say Test Case 1.. (knockout blow) you must account for all variables, opponent size, weight, their skill in defending and in blocking your strike. Even after all that, we can still only cite results in ONE test case.

So yes, I think it is possible to test the art as science and but I think it would require a significant deal of depth in the test criteria and conditions to be useful.

 

[BostonRedBaron]

In this case, say Test Case 1.. (knockout blow) you must account for all variables, opponent size, weight, their skill in defending and in blocking your strike. Even after all that, we can still only cite results in ONE test case.

One test case BUT can and should be tested multiple times to ensure validity. Hence the notion of Reproducibility. But I tried to dumb it down to having one general test case. I think I will re-write that part to state that the test cases are generalizations and should be tested with varying partners and reproduced, which is how I usually approach a technique as its the same when I am testing code. I cant account of every possible combination but I can get a decent spread of the variables to be included. The colossal amount of variables can be generalized down to something simpler such as varying body types (shorter, taller, same height etc.).

I think it is equivalent to testing your server (or whatever) to cope with one piece of code running from one user. That, as you know is not a test from which you can draw conclusions that can be extrapolated across the entire system. You must have more realistic conditions in order to satisfy your quality assurance protocols.

Absolutely true, which I think is the hard part about this notion as there almost needs to be a use case test first to define what the conditions should be for the unit test success criteria. Good point I am going to have to edit the blog to include that notion as well. Also why i stated at the end that I thought regression testing was the most important aspect of this approach to testing martial arts as it allows someone to see what the affects of the new change was across the system as a whole.

 

[Jenna]

I cant account of every possible combination but I can get a decent spread of the variables to be included. The colossal amount of variables can be generalized down to something simpler such as varying body types (shorter, taller, same height etc.).

Exactly so! I think it is imperative that we each of us practice our technique (or reproduce the test with different initial conditions or variables) for as many combinations as is feasible. No we cannot test exhaustively and but as you say we can approximate exhaustive testing by using variables that typify each of the subsets of variables we are dealing with - as you say: using our full power strike as a baseline unchanging constant for our full-power test, we can get data (theoretically) for differing height, weight, defence skill level, speed, agility and ability to handle strikes of our opponent. This is how I always tried to train - narrow down the infinite amount of combinations to a testable subset which were representative of the whole. I think that is probably where it breaks down from your QA stuff which demands tests for every scenario? Perhaps that is not how it works Excuse my ignorance, I am just a diddler when it comes to science or tech things I like your idea though. I would like to see a full-blown out plan even for one technique as I think it would give a scientific framework that has possibly not been done before? It is interesting, thank you.

 

[BostonRedBaron]

Ok. Let me go back and redo the test cases on the blog then and start with one technique and use that for each one of the QA methods of testing. I'll repost when i'm done editing.

Thanks for the input. That gave me a lot to think about.

 

[Jenna]

Ok. Let me go back and redo the test cases on the blog then and start with one technique and use that for each one of the QA methods of testing. I'll repost when i'm done editing.

Thanks for the input. That gave me a lot to think about.

No do not go!! You may get swallowed up in your conditions and never return and you have not introduced yourself yet or explained how you do your JKD! Also, there are plenty of science and engineering people here on board I think too - if you phrased your title differently to reflect the nature of your interest, maybe their radar would pick up your thread in numbers! Good luck though, hope you stay around and welcome to MT!!

 

[BostonRedBaron]

No do not go!! You may get swallowed up in your conditions and never return and you have not introduced yourself yet or explained how you do your JKD!

Heh. You have no idea how true that actually is.

Well, I'm a test engineer for a green company in Boston so i spend 40-60 hours a week in a lab implementing test cases I wrote/designed. I test both hardware and software. I have been training various martial arts since I was a kid. Currently train Muay Thai, Boxing, BJJ, Judo and Dog Brothers FMA. I have studied Wing Tsun/Ving Tsun and some Shotokan karate in the past. I enjoy doing really odd little experiments in the gym like sparring thai boxers but out of a boxing stance and not throwing any kicks. Or vice versa, i spar boxers out of a thai stance & thai stalk (and obviously dont throw any kicks, knees elbows). It has surprised me what i find to work and not work when doing these odd (and painful) little tests of mine, but I seem to "get it" more when stuff fails and i understand why it fails then when stuff just works.

 

[Carol]

Welcome aboard! Greetings from the 603, good to see another New Englander here

Hope you stick around and chat with us, I liked reading your blog entry a lot.

 

[Buka]

Greetings, my friend. Welcome aboard.

As for test casing martial arts techniques - one can do all the tests, simulations, drilling etc that one wants, but it's all for show. To be tested in a martial arts environment, it has to be tested against others that do not train with you. But, again, that's show/sport. What works in a fight, works in a fight, what doesn't, doesn't.

But that's a pretty nice blog you put together there.

 

[BostonRedBaron]

To be tested in a martial arts environment, it has to be tested against others that do not train with you.

I agree. Use-case testing is what you are referring to in the QA world, or more specifically beta-testing, where we place the hardware/software in the real world and see how it performs prior to releasing it as an actual product (need to write that section next). Competition is one form of use-case testing as I addressed in the Use-Case Testing on Multiple Environments section of the blog. Often times its what fails that is more important to analyze than what works. The QA process includes what your are referring to, but its only one piece of the process of testing a system.

I'm glad you like what I wrote and will be updating it shortly to reflect on the input i've gotten from everyone.

Thanks