So today my library book came in. It would have cost $65 via Amazon, more elsewhere, if even available. It was written for Engineers working on their Masters degrees. Luckily, this book is fairly well written and I am following it fairly easily so far. I expect that after mastering this book I shall have to move on to more advanced books on the subject… but also that after mastering this book I will be able to move on to more advanced books on the subject. It is A Primer on the Taguchi Method. Once I have the Taguchi method in my brain I will be able to … well, I will know how to … Well, to select a small set of experiments that will give statistical data enabling the analasys of a large number of variables, and to know the statistical methods for interpreting that data and understanding the outcomes… in such a way as to optimize performance and reliability in a process (manufacturing or otherwise) while at the same time minimizing the cost of production and of error, as well as of the actual experimentation itself. Yes. That is it.
I actually have a fair grasp of the basics of the Taguchi Method already, having got through nearly 5 full chapters in a couple of short hours, and could, if pressed, work out the rest on my own… but .. Oh, I’m lazy; I’ll just keep reading.
But … oh, there was a point to saying that last bit, and it is that upon learning the basics of the Taguchi Method and its applications and limitations and begining to try to dream of how I might utilize it, it occurred to me that I may never have a use for it. Still, I may subject you to it via a survey or two regarding web site design or … I don’t know … comics art style… I keep thinking that it would be really interesting to be able to follow an experimental process to optimize an essentially creative act in some way… to scientifically engineer good art or compelling characters. Just talking about it, I’m thinking about how it may be done… hmmm… Except … it will require actual user feedback from people … something I’ve never quite been able to garner.
I wonder if I could get good data from people about their preferences about comics I do. Thinking about it, I would probably have to (using the Taguchi Method to select the parameters) create at least 16 distinct comics, several ‘strips’ or ‘episodes’ long for people to give feedback on… then, using statistical analysis, I could generate the optimal comic for my audience! As the audience grows and changes, the experiments could be repeated to generate the ideal comic for the new audience!
The trick to using the Taguchi Method, for me, is that in order to use it to generate a certain result, you must know the result you want to generate. In the case of an online comic, the desired result is that readers like the comic, and ‘the more the better’. The Taguchi Method will work with this. But how can I apply it towards my paintings? What is the desired result? My own satisfaction with the work? Volume of sales? Average price of sales? Public adoration? Groupies who trade sex for paintings?
Theoretically, the Taguchi Method could be used towards obtaining any result… as long as you were aware of the desired result and the controllable factors affecting it… I may try to optimize the Buccala recipe a little this year, taking into account size and saltiness of fish pieces, and perhaps re-checking that egg matter… as well as doing an actual comparison of the two yeasts available and single vs. multiple rises of the dough. Any other factors you can think of that might be testable? I could have two more variables without increasing the number of experiments necessary to create an optimal result. At least with Buccala, my family is around to give me feedback on optimal Buccala. I know how much they can and can’t eat… I actually managed to make a little too much last year… as long as by a little, you mean several pounds too much.
Anyway… what else needs to be optimized around here? I’m itching to put this to use… Too often my brain decides to learn a thing and I have no idea why, or how I’ll use it. I figure someday I’ll get stuck in the equivalent of an adventure game and all these strange and curious things I’ve been learning will add up to survival, while my vast ignorance of many of the subjects that seem to be ‘common’ or at least ‘classic’ general knowledge won’t make a smidge of difference. So what if I’ve never read any Dostoyevsky? I know how to optimize things with complex variables using a small number of controlled experiments! Hah!