Mitorikeiko vs. sabermetrics

[Lucien]

Hello and welcome

In kendo we are encouraged to do mitorigeiko, studying how other peole fence and stealing their technique, Our teachers also pass on the results of their long experience through anecdotes and instruction.

While I wouldn't call kendo a sport, we can find this emphasis on observation and experience in almost every sport. The biggest difference between kendo and other sports is that the other sports have started measuring their performance during training and contests to better plan their training and to employ on-the-field tactics more effectively. There is no such analytical data in kendo.

Of course, how we fight with sensei and how we fight our kohei is very different. There is an area of kendo where everyone fights the same and where statistical analysis would be most effective - shiai.

What I am proposing is a crowd-sourcing project where everyone contributes to a kendo database on shiai. You would go to YouTube and select a shiai. You record the general data (URL, who is fighting, where, for what, dan grades) and then watch and record all the actions of the two combatants (how many times attempted men, kote, do, tsuki, success rate). This is only a short list. I think there would be many data points.

The aim would be to build up large database from which we could extract useful data.

Let us assume that you are a member of your national team, a squad member as simply someone who aspires to do better in shiai. Or perhaps, like me, you are just a geek. You could use the data to analyse your opponents and analyse which techniques are the most effective. You could film yourself over a number of fights and use the data to better compare your own kendo to the kendo of your opponents. You could look at global figures, data for people from a certain country or even dojo. And so on.

I do not think this is a very radical idea. Every sport does this and shiai is where kendo is most sport-like.

What I am interested in knowing is:

- Would you help contruct such a database?
- What other datasources would you suggest?
- What problems do you anticipate?
- How can we make this the most useful?


Matt

[b8amack]

Did you just watch Moneyball?

[Lucien]

No, I am just a geek. :)

Actually I read Moneyball, Why England Always Lose and a few other books a while ago. The reason I thought about this is because I was talking to an English Go player on Saturday who had carried out similar analysis on Go. Interestingly - assuming you have no friends and an unhealthy interest in numbers - they found some errors in 600-year-old Go texts.

I also think that some time ago we reached the point where there is sufficient data and easily accessible tools to do this kind of analysis. I think a collaborative project could produce interesting results.

And, yes, I will try to get out more.

[turboyoshi]

1 problem I see is that even knowing what waza was successful for a certain competitor, it doesn't tell you much because seme is really a critical factor and difficult to measure. It might be useful to include example clips or other wise describe how a certain competitor sets up certain waza. This sounds like quite a geek-worthy project.

[Missingno.]

If you make up a spreadsheet with all of your parameters, I'll give a few hours to data collection. Sounds like an interesting idea.

[b8amack]

No, I am just a geek. :)

Actually I read Moneyball, Why England Always Lose and a few other books a while ago.

I asked because I had just watched Moneyball. It's the only reason I knew what sabermetrics were.

Personally, I think this puts the emphasis on the wrong thing. Our aim is not winning, but improving. Knowing statistically that we are more likely to win against player X with debana kote, we should hit anything but debana kote. Unless we are instructing X. Sabermetrics is backwards that way, indeed would lead to a lot more stalling in tsubazeriai for whole matches when there are teams. Of course shiai is not keiko, but that should be read not in the sense of "tigers are not dragons" but rather "shit is not cuisine".

[Bokushingu]

An interesting idea, but i agree with b8amack. Goal is improvement over a lifetime and the purpose of shiai is to measure & test one's training.

[b8amack]

I think I waxed a little grandiose there with my analogies.

[turboyoshi]

Personally, I think this puts the emphasis on the wrong thing. Our aim is not winning, but improving. Knowing statistically that we are more likely to win against player X with debana kote, we should hit anything but debana kote.

That's not only unrealistic but I'm not sure I agree that shiai is the place to try to improve. I think shiai is supposed to be a fun social event as well as a time for measuring how well your training is paying off. I'm not disagreeing that our long term focus shouldn't be on winning shiai but during shiai, we should do the best kendo we can that is most likely to win. Especially if you are in a team match, you should not be selfishly pursuing your own goals but do whatever is best for the team.

[DigitalDowntown]

Hmm, this is a very interesting idea.
Although I agree with the others here, that kendo is about personal growth and not about trying to to excel in shiai, having such statistics would simply be interesting/fun if for no other reason than 'water-cooler talk' at the dojo. My fear is that many would use these kind of statistics as ammunition to turn shiai into more of a 'competitive sport' (think of the controversy around kendo in the Olympics), maybe even leading to gambling (sumo?) and certainly no one serious about their budo wants to go down that road. Still, there's a side of human nature that's curious about such data.

[b8amack]

That's not only unrealistic but I'm not sure I agree that shiai is the place to try to improve. I think shiai is supposed to be a fun social event as well as a time for measuring how well your training is paying off. I'm not disagreeing that our long term focus shouldn't be on winning shiai but during shiai, we should do the best kendo we can that is most likely to win. Especially if you are in a team match, you should not be selfishly pursuing your own goals but do whatever is best for the team.

Read the title again Turbo. Mitorigeiko is for us to learn from our fellow athletes, not to analyze for weaknesses so we can beat them. My earlier examples related to keiko, which is what the person watching would be seeing. This approach fundamentally turns everything you do in kendo upside down.

[Anime12478]

I agree that this is quite an interesting project to take part in. Of course, it could take some time to come up with a template that will be useful to kenshi that engage in tournaments to use. On that note, when it comes to participating in tournaments and mitori-geiko, I have a view that might be unpopular to some.

When I'm in a tournament, I'm there to get that gold metal. Before I go, I work hard on my physical and mental abilities that could help me get to the top. I used to think that participating in tournaments wasn't all about winning, but I guess I've taken a more relaxed view on the situation. If I win, then that means that my training methods worked and I was able to best the people that I was fighting against at that particular tournament on that particular day. After it's over, I take that data to make sure I continue improving and remain on top. If I lose, then that obviously means that I was deficient in something. The next step is to examine the weaknesses I had at the tournament on that day and improve on them so I can perform better at the next tournament.

For mitori-geiko, I think we can all agree that it's a great chance to learn from our fellow kenshi. We can see how people are performing the exercises, and then taking that information and applying it to how we would theoretically handle it based on our skill set and knowledge of Kendo. However, I also think that it's a nice time to analyze the weaknesses in others to see how we can take advantage of that. I'm looking to improve my skill, and the best way to do that is to learn how to win against others. If I win, then that causes the other person to look at why they lost, and then they try to improve and then I need to look at how to deal with a new set of circumstances. Also, mitori-geiko doesn't only happen at general practice, it can also happen at tournaments. If I'm waiting for the next round, I look at the people I may be fighting and then finding out whether or not I can handle them and what I need to do based on the skills I need.

Now, keep in mind that I'm not the kind of person that takes losing situations badly. If I lose, I'm motivated to cut my losses and improve on what I already know. If I happen to win (which, admittedly, is rare) then I still improve on what I already know to make sure I continue to grow and win against people. At first glance, it seems like it's a horrible way to think about Kendo. But then, trying to improve myself to get to the top of tournaments or move up in rank is the basis on why I continue practicing. How one gets there is a matter of personal reflection, and some debate.

[Lucien]

The people who contributed to the data collection would be doing mitorigeiko. I have not yet attempted it, but I think watching a shiai almost frame by frame would be very illuminating.

It is probably worth remembering that video analysis is a very common practice in Japan. I have also heard of at least one national coach in Europe who does this kind of thing already albeit with a pencil and paper. The project would be a way to add some scientific rigour to the process. (I am not a scientist, btw, just a humble arts grad.)

My inital idea was to watch a fight from red's perspective and note down all the discrete actions he did and then do the same for white. Having thought about it further it might be worth recording the actions sequentially to see how points are built. This will probably be a headache to implement but would add some depth to the analysis.

I will come up with some parametres and post them here for further discussion. This will most likely be in the new year. In the meantime I am interested in any and all feedback.

[Bokushingu]

I suppose if you are doing it as a hobby or for entertainment reasons...it could be worth doing. But that type of Data would be distracting to your or others training. In boxing, we used fighter analysis to prep for the fight. But as a boxer winning is most important.

Lol I remember a kyu was so frustrated with her kendo. i asked her why. And she said, she had been keeping an exact log of her accuracy. And she wasn't happy with her "miss to hit" ratio, and it was becoming hard to improve. I told her to throw it all away and don't keep track anymore! She said why . And i said " you shifted your focus from improving to results." So i told her if she miss, think of why she missed that moment...then move on, but try to correct the error if she could tell what it was, if she couldn't, ask our Sensei. After that, she said she felt better and less frustrated....and she felt she was improving now.

So data collection can be a distraction...especially for something as dynamic as Kendo.

[rfoxmich]

I recall some time ago statistics being published about the percentage of tournament points that were men, kote, do and tsuki. But to put a cap on the discussion so far see:

http://radgeek.com/gt/2006/04/28/ItMeansYou%27reHavingMoreFun.png

Which puts this all in perspective perhaps.

[b8amack]

I agree pretty strongly with Bokushingu. Of course we all do this somewhat for shiai, but I feel ultimately it's a step in the Olympicization of kendo.

[Jiyoui]

It's not about the data you collect but why you collected it.
Things we do cannot be explained in a binomial. It's a multivariate with none specific weights to each variable.

Certain things in life can be boiled down to an equations. Life itself cannot.
It might be worthwhile to go google why the game 'go' is so hard to teach a computer. Kendo is very much like this.

[jjcruiser]

It's an interesting idea, but the reason sabermetrics (and before that Bill James's Baseball Prospectus projections) were applied was to quantify baseball (and other sports') player's abilities in an economically efficient and predictive manner. That's inapplicable in Kendo. It's not like you are trying to hire certain good players, and you don't want to pay them as much as someone else, so you think jodan is undervalued, and go sign several jodan players to under market contracts; knowing that their men may not be as successful as their kote but that kote scores in higher percentages so that's okay.

[turboyoshi]

It might be worthwhile to go google why the game 'go' is so hard to teach a computer. Kendo is very much like this.

It's not the same. Go has such complexity because there are more possible permutations than a machine can reasonably compute. People behave in much more predictable ways. This is why statistics and profiling are effective.
I wouldn't go to this much trouble to analyze kendo but I can't understand why anyone would object. If you're the kind of person interested in personal growth, then what other people do with their own kendo shouldn't affect you. The rest of you would be hypocritical to object because you do the same thing, when you go to a taikai and watch the people you're likely to compete against and try to figure out what you'll need to do to beat them. You're just doing it on a much less structured and organized scale. I wouldn't even worry about it being used to olympicize kendo. The governing bodies for kendo are going to be the ones who determine whether things are allowed to go that far and we just have to trust them to keep kendo from going down that path. Most os us don't really have a say in such matters though so it's generally irrelevant.
Besides, if you really care about developing your kendo, the fact that others are looking at your weaknesses will force you to become better to stay competitive. There's no downside here, except to your own ego.

[b8amack]

It's not the same. Go has such complexity because there are more possible permutations than a machine can reasonably compute. People behave in much more predictable ways. This is why statistics and profiling are effective.
I wouldn't go to this much trouble to analyze kendo but I can't understand why anyone would object. If you're the kind of person interested in personal growth, then what other people do with their own kendo shouldn't affect you. The rest of you would be hypocritical to object because you do the same thing, when you go to a taikai and watch the people you're likely to compete against and try to figure out what you'll need to do to beat them. You're just doing it on a much less structured and organized scale. I wouldn't even worry about it being used to olympicize kendo. The governing bodies for kendo are going to be the ones who determine whether things are allowed to go that far and we just have to trust them to keep kendo from going down that path. Most os us don't really have a say in such matters though so it's generally irrelevant.
Besides, if you really care about developing your kendo, the fact that others are looking at your weaknesses will force you to become better to stay competitive. There's no downside here, except to your own ego.

I think of the line from the movie Moneyball, where David Justice is being told to stop stealing bases. (Quotation from memory, and so not exact)



DJ: But that's what I do. That's why you hired me.
BB: We hired you to get to first base, not to get thrown out at second.

In the context of baseball, this is absolutely a sound philosophy. In the context of kendo, this kind of thinking is antithetical to core concepts, like sutemi, and personal growth. DJ was essentially being told to stop trying, for the good of the team.

This has nothing to do with ego in the sense that you mean it. I've seen this shit in fencing. Statistically, it's easier to score with flicks than realistic motions based on how one would use an actual sword, and so the martial aspect of the sport is lost. You're deluding yourself to say there's no downside to approaching kendo this way.

[Missingno.]

I think of the line from the movie Moneyball, where David Justice is being told to stop stealing bases. (Quotation from memory, and so not exact)



In the context of baseball, this is absolutely a sound philosophy. In the context of kendo, this kind of thinking is antithetical to core concepts, like sutemi, and personal growth. DJ was essentially being told to stop trying, for the good of the team.

This has nothing to do with ego in the sense that you mean it. I've seen this shit in fencing. Statistically, it's easier to score with flicks than realistic motions based on how one would use an actual sword, and so the martial aspect of the sport is lost. You're deluding yourself to say there's no downside to approaching kendo this way.

So you think gathering some data is going to make people stop trying in kendo? I strongly disagree for a few reasons.

First, no matter how much data is collected, we will always need to fulfill the requirements for yuko datotsu in order to score. This discounts the possibility of scoring with flicks like we do in fencing, where all you need to do is press a button on the tip of the blade.

Second, I don't know how your dojo operates, but in my dojo we don't primarily learn how to score points. Our sensei tries to teach us to do beautiful kendo, not primarily shiai effective kendo, and with good reason. Namely, shiai is only a small part of kendo. If this data changes the way people approach their kendo, then they probably weren't approaching it in the right way to begin with.

Third, this just going to be a private project done for fun. It's not like anyone's proposing that we get federations to hire people to take statistical data on every match, it would just be a small sample of selected matches that would be taken for fun. There is literally no way that this could lead to the olypicization of kendo.

[turboyoshi]

I've seen this shit in fencing. Statistically, it's easier to score with flicks than realistic motions based on how one would use an actual sword, and so the martial aspect of the sport is lost.u're deluding yourself to say there's no downside to approaching kendo this way.

Now you're just being contrary. The idea of flicky kendo wasn't even brought up and this type of data analysis doesn't change the way people will execute waza. First of all, it's nothing different than what people already do mentally, just more structured and on a wider scale. Secondly, flicky strikes won't score unless the shinpan award them. You want to stop flicky strikes, then as a shinpan, the burden is on you to not award those kinds of points. I would love that since I hate small waza and I have some confidence in my ability to execute large strikes although my seme sucks and I can't see how I'd create an opening from that maai. If you're not a shinpan, you really don't have much influence in the matter so best not to even worry about it. Thirdly, you can't stop it. People, especially the competitive ones, will do it, whether or not you object to it. I understand the philosophical/moral counterargument and even agree with it. Maybe I lack foresight but I just can't see this kind of thing causing any harm. IMO, the biggest threat to keeping kendo pure is money. Once people start making money on kendo, either through sponsorships or what have you, then you attract the unscrupulous types and it's all downhill from there. I'm sure eventually that will happen to kendo, because that's just human nature but I just hope it isn't during my lifetime.

[Fudo-Shin]

Wow, this debate is paradoxical. There is clear polarization here, yet I agree wholeheartedly with both sides. In essence I think all are united with keeping the purity of kendo alive, but how do we go about it?

Personally I don't see a need to over-analyze but rather just practice hard and do our best to keep our kendo 'beautiful'. Shiai etc will sort itself out.

That said, I completely agree with Turbo that money would be the biggest threat to kendo and I believe that is at the core of kendo remaining pure.

However, while an exercise like this may be harmless, I just don't see how it is worth it's weight either as it doesn't shed light on the non-tangible aspects of kendo which are all important.

[JSchmidt]

A couple of problems with this:
The reason for the attack is as important as the attack itself. Often that will be very hard for others to judge. How do you quantify that in statistics?

Sabermetrics also works because of the large amount of games that they play in MLB. I'm not sure of the average number of at-bats per player, it'll be a high percentage of 3-4 at bats * 161 games, so for arguments sake, lets call it 500 at bats per season.

How many fights does the keen (but average) tournament kendo player have per year?. 20? Even if we then make that a successful keen tournament player and include team matches, you'll still struggle to get that number over 50. You then add the regular (casual?) player and the average amount of matches per year probably drops below 10.

[JSchmidt]

DJ: But that's what I do. That's why you hired me.
BB: We hired you to get to first base, not to get thrown out at second.

In the context of baseball, this is absolutely a sound philosophy. In the context of kendo, this kind of thinking is antithetical to core concepts, like sutemi, and personal growth. DJ was essentially being told to stop trying, for the good of the team.


What the role of the chuken in a team match?.

[turboyoshi]

A couple of problems with this:
The reason for the attack is as important as the attack itself. Often that will be very hard for others to judge. How do you quantify that in statistics?

Sabermetrics also works because of the large amount of games that they play in MLB. I'm not sure of the average number of at-bats per player, it'll be a high percentage of 3-4 at bats * 161 games, so for arguments sake, lets call it 500 at bats per season.


I agree with both these points. That just adds more weight to why I can't be too concerned with people taking on geek projects like this. They might change things for a relatively few people but on the whole, I just don't see how this kind of project would have a serious impact on the kendo world.

[b8amack]

Now you're just being contrary. The idea of flicky kendo wasn't even brought up and this type of data analysis doesn't change the way people will execute waza. First of all, it's nothing different than what people already do mentally, just more structured and on a wider scale. Secondly, flicky strikes won't score unless the shinpan award them. You want to stop flicky strikes, then as a shinpan, the burden is on you to not award those kinds of points. I would love that since I hate small waza and I have some confidence in my ability to execute large strikes although my seme sucks and I can't see how I'd create an opening from that maai. If you're not a shinpan, you really don't have much influence in the matter so best not to even worry about it. Thirdly, you can't stop it. People, especially the competitive ones, will do it, whether or not you object to it. I understand the philosophical/moral counterargument and even agree with it. Maybe I lack foresight but I just can't see this kind of thing causing any harm. IMO, the biggest threat to keeping kendo pure is money. Once people start making money on kendo, either through sponsorships or what have you, then you attract the unscrupulous types and it's all downhill from there. I'm sure eventually that will happen to kendo, because that's just human nature but I just hope it isn't during my lifetime.

That's the second time you've told me I'm "just being contrary". Look it up. Contrary does not mean "does not agree with Turboyoshi." I'm being consistent.

Don't deliberately dumb yourself down, in order to win an argument. I mentioned flicky in fencing as an example of what an overemphasis on "barest minimum to win" thinking can do. I allowed myself to think you had the imagination to extend that example to kendo without just leaping on the most literal way to refute it. Apparently, I was wrong.

[b8amack]

A couple of problems with this:
The reason for the attack is as important as the attack itself. Often that will be very hard for others to judge. How do you quantify that in statistics?

That's not a problem, but an agreement with what I wrote. I think statistics is a much broader term than sabermetrics. (Otherwise, we'd just see the former in the title). We can not quantify the reason for the attack statistically, but if the reason for our attack is born out of statistics... well, why would you do jodan, for example? Statistically, not as successful as chudan, and you know it. There are all sorts of reasons why we could use sabermetrics to convince you to stop using jodan, for the good of the team. Judges statistically award less points, especially at smaller tournaments, and misjudge the points they give chudan opponents more than you would see against chudan... I'm sure I could go on. Let's be clear, Billy Beane would likely have cut you. Certainly me. Ah, but I shouldn't make jokes like that, because Turbo is waiting to leap in and say "See! It's about your ego!"


Sabermetrics also works because of the large amount of games that they play in MLB. I'm not sure of the average number of at-bats per player, it'll be a high percentage of 3-4 at bats * 161 games, so for arguments sake, lets call it 500 at bats per season.

How many fights does the keen (but average) tournament kendo player have per year?. 20? Even if we then make that a successful keen tournament player and include team matches, you'll still struggle to get that number over 50. You then add the regular (casual?) player and the average amount of matches per year probably drops below 10.

Right. I don't disagree with this at all. In fact, like the above, it only furthers my point. Sabermetrics does not work for kendo. Not unless you also include keiko, which then would skew both keiko and the statistics.

[b8amack]

What the role of the chuken in a team match?.

Is it "Doing the barest minimum"? If not, then ask yourself if that boldening really means what you want it to mean.

[b8amack]

So you think gathering some data is going to make people stop trying in kendo? I strongly disagree for a few reasons.

First, no matter how much data is collected, we will always need to fulfill the requirements for yuko datotsu in order to score. This discounts the possibility of scoring with flicks like we do in fencing, where all you need to do is press a button on the tip of the blade.


If you do a lot of HD slo-mo analysis, you'd probably find that a lot of ippon are awarded without really fulfilling the requirements for yuko datotsu. Just personally, I've had ippon scored off the front of my kote (and by front, I mean the front of my hands) and my right elbow. Both in tournaments this month. I suspect I'm hardly unique in this. Again... analogies require imagination.



Second, I don't know how your dojo operates, but in my dojo we don't primarily learn how to score points. Our sensei tries to teach us to do beautiful kendo, not primarily shiai effective kendo, and with good reason. Namely, shiai is only a small part of kendo. If this data changes the way people approach their kendo, then they probably weren't approaching it in the right way to begin with.

If I was going to be overly literal, as you are, I would argue "then why are you disregarding your sensei's example by approaching mitorikeiko this way"? But I'm not, as you clearly misunderstand me. It is not the data that I object to.


Third, this just going to be a private project done for fun. It's not like anyone's proposing that we get federations to hire people to take statistical data on every match, it would just be a small sample of selected matches that would be taken for fun. There is literally no way that this could lead to the olypicization of kendo.

Read the OP again. You see how he mentions "member of a national team"?

I think the problem with some of you guys is that you are living in countries where there is NOT a huge movement to bring kendo to the Olympics, and so you don't take this very seriously. I do live in such a country, and so this is not just a reason for me to be "contrary", as Turbo puts it. There are very solid reasons why I am opposed to this way of looking at/practicing kendo. That you yourself may only be doing this "for fun" changes that not at all.

[b8amack]

*Apologies to Jakob... I got on a typing jag there, and didn't realize I was reading your posts wrong (probably because I read them in reverse order). Sorry about that. I'd edit my earlier post in response to yours, but you know, past the time limit.

[JSchmidt]

Is it "Doing the barest minimum"? If not, then ask yourself if that boldening really means what you want it to mean.

First part is: Don't lose!. Senpo, jiho, can afford to take risks. Chuken has to be able to, at the very least, get a draw. So..essentially...get on to first base and don't get thrown out trying to steal second...for the good of the team.

(And yeah, figured you caught that post wrong...no worries :) )

[b8amack]

First part is: Don't lose!. Senpo, jiho, can afford to take risks. Chuken has to be able to, at the very least, get a draw. So..essentially...get on to first base and don't get thrown out trying to steal second...for the good of the team.

If you're aiming for a draw, then it needs to be interpreted defensively. With apologies to Minnie Driver in Good Will Hunting, I'm just going to mix my sporting metaphors. In a five-man team (kendo), that's more along the lines of allowing no runs in (baseball) for nearly two innings. That's significantly more than walking to first!


(And yeah, figured you caught that post wrong...no worries :) )

Hardly the first time I've typed my foot into my mouth here. Wish I could say it'll be my last. ;)

[Missingno.]

If you do a lot of HD slo-mo analysis, you'd probably find that a lot of ippon are awarded without really fulfilling the requirements for yuko datotsu. Just personally, I've had ippon scored off the front of my kote (and by front, I mean the front of my hands) and my right elbow. Both in tournaments this month. I suspect I'm hardly unique in this. Again... analogies require imagination.

I'm not sure what the point of you talking about shinpan quality is. Your point was that fencers flick because it's the bare minimum required for scoring and you don't think flicking is pretty, and my point was that our minimums for scoring are more strict. Slow motion analysis and shinpan quality is irrelevant to the point. Sure you may score something that misses a little bit that the shinpan might make a mistake on, but that doesn't change the requirements themselves. Even if it does miss, your attack is probably not going to be ippon unless it shows legitimate intent and (even though it's not actually specified in the rules) a decent amount of power.

So your new point is that you think people will strive to not fulfill the requirements for yuko datotsu because they saw data on slow motion analysis that says attacks miss sometimes. I'm not really sure how you think that would change anything. If someone scores like that, it's because they sold their attack really well. The hit itself might not have been there, but everything else was very strong. There was a thread about this actually that came up after last year's AJKC about whether hitting datotso-bu was really necessary for yuko datotsu if all the other elements are there (I personally think it absolutely is the most important part, but mistakes can be made even at the highest levels).

Furthermore, the governing bodies of kendo tend to change the rules when kendo isn't going in the direction that it wants. For example, the introduction of munezuki in the 80s to discourage jodan kenshi, which worked very well. If people missing on purpose really becomes a problem, then a rule change will absolutely be made.


If I was going to be overly literal, as you are, I would argue "then why are you disregarding your sensei's example by approaching mitorikeiko this way"? But I'm not, as you clearly misunderstand me. It is not the data that I object to.

I know, you object to people changing the way they approach kendo, which is what I'm trying to demonstrate is an irrational objection.


Read the OP again. You see how he mentions "member of a national team"?

And? Why should a member of a national team be barred from looking at data compiled from a private project taken for fun?


I think the problem with some of you guys is that you are living in countries where there is NOT a huge movement to bring kendo to the Olympics, and so you don't take this very seriously. I do live in such a country, and so this is not just a reason for me to be "contrary", as Turbo puts it. There are very solid reasons why I am opposed to this way of looking at/practicing kendo. That you yourself may only be doing this "for fun" changes that not at all.

That's true, I do, and I don't. I don't take it very seriously because it's not very serious. Your problem is you're dodging arguments. My argument is that gathering and analyzing data will not change the way people will look at or practice or approach kendo. My reasoning, which you have dismissed, is that shiai is a very small part of kendo and what we learn in the dojo does not revolve around doing well in shiai.

[DigitalDowntown]

I think b8amack does make some valid points, many dojo (not just in S. Korea & Japan) do put more emphasis on scoring points in shiai. Some may even already be doing their own statistical analysis of shiai internally; particularly school teams. The data sample may not be as large for kendo as it is for MLB, but that doesn't change the fact that data can still be gathered, and statistics can still generated.

As Schmidt-sensei reminded us, strategy for the 'good of the team' already exists outside of the 'just do the your best, fight your best kendo' mentality. It regularly goes beyond the chuken being told to lock up a draw; what about teams that swap their 'best' player in as jiho? What about 'pushing' your opponent out of bounds instead of going for ippon? Is this not a strategy to win over growing one's self through application of the sword? (I guess you could argue this grows one's self-ego.)

I agree that making money through kendo is the greatest risk to 'keeping the art pure,' and I do think there is a valid concern in making this kind of statistical data widely available; after all, such data can lead to organized gambling among other things. I just don't see today's kendo attracting that side of society the way perhaps, MMA fights can. Of course, tomorrow's kendo, possibly heavily laden with statistical data, might become something different altogether.
I for one am, however, interested in such data; perhaps like Lucien, more just to satisfy my curiosity than to help 'stack my team' for the win.

[turboyoshi]

I agree that making money through kendo is the greatest risk to 'keeping the art pure,' and I do think there is a valid concern in making this kind of statistical data widely available; after all, such data can lead to organized gambling among other things.

The thing is, this data is easily collectable and I'm sure already done by the types of people who would put down a bet on a coin flip. I'm sure there are people who already gamble on kendo and I bet they do some very detailed statistical analysis to determine odds and payouts. As long as kendo doesn't become so popular that it attracts the average gambler, we'll be all right. b8amack keeps introducing new arguments instead of actually countering the arguments I and others have already made. I believe he is irrationally overreacting to a perceived threat rather than an actual threat precisely because he lives in a country where many would like to olympicize kendo. His arguments are founded on emotion rather than logic. That isn't wrong in itself, but I just find it unconvincing for this particular case.

[JSchmidt]

The thing is, this data is easily collectable

Again..all you can reliably collect, is what point is awarded. In order for that data to have any relevance, you need to know why that point was scored and good luck in collecting that.

[b8amack]

The thing is, this data is easily collectable and I'm sure already done by the types of people who would put down a bet on a coin flip. I'm sure there are people who already gamble on kendo and I bet they do some very detailed statistical analysis to determine odds and payouts. As long as kendo doesn't become so popular that it attracts the average gambler, we'll be all right. b8amack keeps introducing new arguments instead of actually countering the arguments I and others have already made. I believe he is irrationally overreacting to a perceived threat rather than an actual threat precisely because he lives in a country where many would like to olympicize kendo. His arguments are founded on emotion rather than logic. That isn't wrong in itself, but I just find it unconvincing for this particular case.

Actually, your argument ("It's fun and it means nothing") is an emotional argument. My argument (Which has not changed... and remains that approaching mitorigeiko in this way emphasizes the wrong thing: winning vs. improvement) may well be emotional. But I have provided logical reasons why I hold to that. I know your strategy of arguing, which is to complain about the argument and the arguer (he's irrational! He's overreacting! It's about his ego!) rather than address it... weak.

[b8amack]

I'm not sure what the point of you talking about shinpan quality is.

Statistical relevance... it's right there.


Your point was that fencers flick because it's the bare minimum required for scoring Correct


and (you don't think flicking is pretty,)

Incorrect and dismissive.

and my point was that our minimums for scoring are more strict.

No, you said they will all be met for ippon, nor did you mention them as "minimums", but absolutes. (Who is changing his argument?) Which is ideally true but extremely difficult to ensure. False, to put it simply.


Slow motion analysis and shinpan quality is irrelevant to the point.

How is it irrelevant when the OP mentions watching videos in slo-mo/in-depth to see points? Am I supposed to ignore that? And how can shinpan quality be irrelevant? If you were going to do any serious statistical analysis of kendo shiai, you would HAVE to factor in the shinpan themselves, or the data would be meaningless.


Sure you may score something that misses a little bit that the shinpan might make a mistake on, but that doesn't change the requirements themselves. Even if it does miss, your attack is probably not going to be ippon unless it shows legitimate intent and (even though it's not actually specified in the rules) a decent amount of power.

You just changed what you said above.


So your new point

Honestly. Nice try, princess. My point remains the same.


is that you think people will strive to not fulfill the requirements for yuko datotsu because they saw data on slow motion analysis that says attacks miss sometimes. I'm not really sure how you think that would change anything. If someone scores like that, it's because they sold their attack really well. The hit itself might not have been there, but everything else was very strong. There was a thread about this actually that came up after last year's AJKC about whether hitting datotso-bu was really necessary for yuko datotsu if all the other elements are there (I personally think it absolutely is the most important part, but mistakes can be made even at the highest levels).

You're changing your earlier statement again. YOU were the one who brought in the yuko datotsu as an absolute, not me. Don't explain how it's not now and act like that wasn't what I just said.


Furthermore, the governing bodies of kendo tend to change the rules when kendo isn't going in the direction that it wants. For example, the introduction of munezuki in the 80s to discourage jodan kenshi, which worked very well. If people missing on purpose really becomes a problem, then a rule change will absolutely be made.

For a person who just said that there is no way this would never impact kendo, this is a strange point to make. You assume that all rule changes are to the benefit of kendo. That is an indefensible position. The munezuki rule was NOT beneficial, which is why it's gone. The damage has been done, however.




I know, you object to people changing the way they approach kendo, which is what I'm trying to demonstrate is an irrational objection.

Not to a traditionalist, no.




And? Why should a member of a national team be barred from looking at data compiled from a private project taken for fun?

This is a public and global forum. What was proposed was not a private project, by any definition of that word.




That's true, I do, and I don't. I don't take it very seriously because it's not very serious. Your problem is you're dodging arguments. My argument is that gathering and analyzing data will not change the way people will look at or practice or approach kendo. My reasoning, which you have dismissed, is that shiai is a very small part of kendo and what we learn in the dojo does not revolve around doing well in shiai.

You're dodging arguments, not me. I have addressed all your points very specifically. My problem with this last is that it's hypocritical to argue for an emphasis on one thing (shiai) while proclaiming its non-emphasis. "Gambling on kendo doesn't matter because nobody gambles on kendo. So it's ok if I do it!" That's the exact same argument you are making. Don't believe me, substitute "Emphasis on shiai" for "Gambling on kendo". In fact substitute any old thing in there: "Eating rocks"... "Wearing linoleum"... always comes out a nonsense argument.

Of course, an emphasis on shiai is NOT what I'm opposed to, as I said last time. But of course, you ignored that. And in any case, what you guys do I have no control of. You will do as you please. Why am I the one being accused of overreacting?

[b8amack]

I should add that I do not consider it wrong to gamble on shiai. I used that as an example only because it was just mentioned. For that matter, I don't consider it "wrong" to wear linoleum. Strange and impractical, and probably uncomfortable... certainly unfashionable, but not "wrong." Eating rocks, however, is wrong.

[Missingno.]

Statistical relevance... it's right there.

You used it to argue against my point about the minimum requirements being more strict in kendo then in fencing. They still are, regardless of shinpan quality.


No, you said they will all be met for ippon, nor did you mention them as "minimums", but absolutes. (Who is changing his argument?) Which is ideally true but extremely difficult to ensure. False, to put it simply.[quote]

Here is what I said:

[quote]First, no matter how much data is collected, we will always need to fulfill the requirements for yuko datotsu in order to score. This discounts the possibility of scoring with flicks like we do in fencing, where all you need to do is press a button on the tip of the blade.

I made a comparison between the minimum requirements needed to score in kendo and the minimum requirements needed to score in fencing. The fact that you didn't understand this when you first skimmed it is not my problem.


How is it irrelevant when the OP mentions watching videos in slo-mo/in-depth to see points? Am I supposed to ignore that? And how can shinpan quality be irrelevant? If you were going to do any serious statistical analysis of kendo shiai, you would HAVE to factor in the shinpan themselves, or the data would be meaningless.

It's irrelevant to my point about minimum requirements. See above.


You just changed what you said above.

No I didn't I was explaining how it was irrelevant to my point. You're mistaking my explanations of my points for points themselves. I'll post what this is all referencing again, just so we're clear.


First, no matter how much data is collected, we will always need to fulfill the requirements for yuko datotsu in order to score. This discounts the possibility of scoring with flicks like we do in fencing, where all you need to do is press a button on the tip of the blade.


Honestly. Nice try, princess. My point remains the same.

How mature.


You're changing your earlier statement again. YOU were the one who brought in the yuko datotsu as an absolute, not me. Don't explain how it's not now and act like that wasn't what I just said.

I made a comparison. See above.


For a person who just said that there is no way this would never impact kendo, this is a strange point to make. You assume that all rule changes are to the benefit of kendo. That is an indefensible position. The munezuki rule was NOT beneficial, which is why it's gone. The damage has been done, however.

It did its job.



Not to a traditionalist, no.

Ok... Not really an argument, so I can't respond to it. I guess you win?


This is a public and global forum. What was proposed was not a private project, by any definition of that word.

Semantics. You understood what I meant.


You're dodging arguments, not me. I have addressed all your points very specifically. My problem with this last is that it's hypocritical to argue for an emphasis on one thing (shiai) while proclaiming its non-emphasis. "Gambling on kendo doesn't matter because nobody gambles on kendo. So it's ok if I do it!" That's the exact same argument you are making. Don't believe me, substitute "Emphasis on shiai" for "Gambling on kendo". In fact substitute any old thing in there: "Eating rocks"... "Wearing linoleum"... always comes out a nonsense argument.

I'm not arguing for emphasis on shiai. Never in any of my posts have I argued for emphasis on shiai. Just because I am in favor of a project that would involve shiai doesn't mean I'm arguing for emphasis on shiai. Wanting to gather data does not imply wanting to change the way I approach kendo, it implies wanting to gather data. This is an important distinction that you have not been able to make, which is leading you to believe things like the idea that I'm arguing for emphasis on shiai.


Of course, an emphasis on shiai is NOT what I'm opposed to, as I said last time. But of course, you ignored that. And in any case, what you guys do I have no control of. You will do as you please. Why am I the one being accused of overreacting?

I don't recall making such an accusation. I apologize if that's how I came across.

[Lucien]

As the OP i wanted to say that the conversations in this thread are vry interesting. I would like to contribute more but am on holiday. I'll write more in the New Year. I have started working on the data collection tool and hope to finish it mid -Jan.

[Maro]

Sabremetrics. Has there been a more try-hard name?

[JSchmidt]

Except that Sabermetrics (http://en.wikipedia.org/wiki/Sabermetrics) comes from baseball and is not related to sword arts at all.


The term is derived from the acronym SABR, which stands for the Society for American Baseball Research.

[futabachan]

The whole point of sabermetrics was that the traditional teachings of baseball (the Triple Crown stats) were wonky, and obviously so. Is that what you're trying to say about kendo, Lucien?

[Maro]

Except that Sabermetrics (http://en.wikipedia.org/wiki/Sabermetrics) comes from baseball and is not related to sword arts at all.

I know it doesn't come from Sword.

You can bet the person who thought of it nearly passed out with joy when he worked the name though.........

[Jiyoui]

holy... this thing is still going?
I say, go for it, good luck. It'd be interesting to see what you find, but you really have to set out some defined variables.

In baseball, a swing is a swing and a strike is a strike. ta-da batting average
kendo now - what counts miss when someone goes for a men. just a straight men, what about men-kaeshi-men if the opportunity was good is that still a miss? Or in the case of kote, what is a kote-men, is it a kote swing or a men swing.

Not saying that it can't be done, but applying math to the real world is tricky thing to do when the environment isn't controlled, or not as controlled as you think it is.