The tree training method

[Monkey Turned Wolf]

I think this is in the same concept of trying to map out a conversation-an impossible task. With a conversation, I say "Hi how are you", there are 50 different ways someone might respond, if I have a response to each (or even group them together to 10 ways), then they respond to that in 50 different ways. And it seems like an impossible task. Yet pretty much any of us can carry out a conversation.

It does much more to teach people general principles and give them experience to learn how to adapt and choose from the situation that they're in, then it does to have them memorize specific responses to specific attacks/counters. The conversation version of this would be to teach them basic counters and blocks, have people throw punches/kicks/try to grab them, and they learn to respond to those, the same way kids learn to respond to unexpected (to them) responses to conversations.

It's why pressure testing is so important.

 

[wab25]

It should be a forest (a set of trees) and not just a single tree.

- Any missing tree branch may indicate that something is missing in the system.
- Any wrong tree branch may indicate that something is wrong in the system.

I believe it may take someone lifetime to develop it. One may develop only 1 or 2 trees in his life. One may never develop any tree.

The advantage of this approach is you can use 1 technique to link many techniques into a group (tree). It's easy to learn, teach, train, remember ....

For example, if your system has 200 techniques, how do you remember those techniques? You can remember

1. one after another.
2. as a set of groups (trees).

IMO, 1 < 2.

Okay... so show me the complete forest... or one complete tree for that matter... I'll wait....

Whats interesting is that every time someone brings up a different view, you respond with a list of techniques that can come from and attack or that can follow up an attack... and many times you toss in boolean operators to show relationships...

This kind of organization is great.... if you are writing computer code for a fighting video game. In fact, these kind of trees and relationships are necessary to write code for a video game. However, people operate in a totally different way. What works to make a playstation know how to fight, does not work for people. Further, even though the video game "knows" the trees and relationships... it does not know how to fight, or how to actually apply the techniques... it simply knows the tree. Even in a video game, where the things the other guy can do are finite in number, and completely known... these systems still have issues when the other guy does something he is not supposed to do. There are many edge cases, where the trees and relationships do not have the answer... video games handle this by defining some kind of default behavior... but that default behavior is there to cover a defect, not because it is a useful response. This is why it is one thing to be good against the computer and quite another to be good against another human player. Humans learn how to use those relationship trees in different ways and can become much better at the game, than the computer can... and many humans do this, with out ever learning, memorizing or looking at the technique trees used to write the code. They learn a few core things, that can be applied to many situations, in different ways and they get good at using and understanding that subset.... many times humans find a better way to use and apply these techniques than the relationship tree.

What I find interesting is that you think that this is some great way to teach... yet when I ask to see your tree for your art... you hand wave about how its a forest, where every tree, branch and root must be perfect... yet you can't show the one you use...

 

[Kung Fu Wang]

yet you can't show the one you use...

There is no way that anybody on earth can show the complete tree (or forest). The idea is to use technique 1 to set up technique 2, to use technique 2 to set up technique 3, ... Most people may go down to 3 levels deep. Very few people can go deeper than that.

The question is whether there is any value to train combo (such as groin kick, face punch). The issue is if one doesn't train combo, when his opponent escapes his 1st attack, he may not be able to continue his 2nd attack.

In the following 15 seconds short clip, he made the following combo worked twice back-to-back. It proves that he had trained this combo before. He didn't just pull it out by accident.

1. right roundhouse kick.
2. right side kick.
3. left hook.
4. right hook.
5. underhook clinch.
6. right knee strike.
7. right outer hook throw.

In the following demo clip,

1. he uses shin bite to attack his opponent's right leg.
2. When his opponent escapes, he uses reverse shin bite to attack his opponent's left leg.
3. When his opponent escapes again, he uses foot sweep to attack his opponent's right leg again.

 

[skribs]

There is no way that anybody on earth can show the complete tree (or forest). The idea is to use technique 1 to set up technique 2, to use technique 2 to set up technique 3, ... Most people may go down to 3 levels deep. Very few people can go deeper than that.

If you can't show it, then it's not a training method. It's a nebulous concept used to create an air of superiority. You prove it in the second part of this quote. You say most people can't go more than three levels deep. Maybe they can, but they don't, because they don't think it makes sense to go deeper than that.

The question is whether there is any value to train combo (such as groin kick, face punch). The issue is if one doesn't train combo, when his opponent escapes his 1st attack, he may not be able to continue his 2nd attack.

The fact that combinations are a thing doesn't validate this training model. There are many other ways of looking at combinations.

1. Your combination assumes a specific reaction, and if your opponent reacts differently you break it off. You create your next combination based on what you learned about your opponent.
2. Combination training is useful for learning how to combine techniques together (for example, a 1-2 punch instead of a jab, and later a cross). Knowing which techniques to use in the combination is a matter of reading your opponent.

I'm not saying the model is invalid. I'm just saying this doesn't prove the model is valid.

Using your example, let's say I go for a (groin kick, straight face punch) combo, under the assumption that the groin kick will drop my opponent's hands. My opponent uses his hip to block and keeps his hands up. Let's look for each option:

1. I may greedily go for the combo, in which case I will throw the straight punch into his guard. The next time I go for this combo, I will know he keeps his guard up, and opt for a different attack (such as a hook punch).
2. I may go for the combo, and read that my straight punch won't hit through the guard, and change to a hook punch. This isn't based on a tree starting from groin kick, this is just based on "guard beats straight punch". I'm not even going 2 levels deep, let alone 3 or more, because each technique is assessed on its own.

In the following 15 seconds short clip, he made the following combo worked twice back-to-back. It proves that he had trained this combo before. He didn't just pull it out by accident.

This is the opposite of the tree method. He does the same combo twice. This is rote work and not branching decisions.

 

[wab25]

There is no way that anybody on earth can show the complete tree (or forest). The idea is to use technique 1 to set up technique 2, to use technique 2 to set up technique 3, ... Most people may go down to 3 levels deep. Very few people can go deeper than that.

So then, there really is no way to verify this:

- Any missing tree branch may indicate that something is missing in the system.
- Any wrong tree branch may indicate that something is wrong in the system.

Since we can't have the whole forest or the whole tree... how would we know what was missing or what was wrong?

The question is whether there is any value to train combo

This is a different question. Of coarse you should train combos. Even better, you should learn how to make your own combos.

The problem with bringing everything down to absolutes "a > b" is that people are not machines. Every time you come up with an absolute "a > b" people can immediately come up with cases where "b > a" for a certain set. Idea is to really learn "a" and really learn "b" and learn how they can work together and what they do to both you and the other guy. Fights rarely end because you did "a" and I did "b." They end because you were able to apply "a" better than I was able to apply "b." There is a subtle but important difference in those last two sentences.

 

[Kung Fu Wang]

This is a different question. Of coarse you should train combos. Even better, you should learn how to make your own combos.

The tree training method is the same as the combo training method.

If you agree with the combo training, we may agree more than we may dis-agree.

1. A throws a groin kick at B.
2. If B drops his arm to block it, A punches on B's face.
3. If B raise leg to block it (still on guard), A uses punch to pull B's guard apart, A then throw another punch at B's face.

You can use a groin kick to set up a face punch. Whether your face punch will require a pull guard or not depends on whether your opponent may drop his arm to block your kick or not.

 

[skribs]

The tree training method is the same as the combo training method.

No, it really isn't. It's just one way of looking at combinations.

 

[Kung Fu Wang]

No, it really isn't. It's just one way of looking at combinations.

What's the other way to look at combination?

- You make a move.
- Your opponent responds to your move.
- ...

 

[skribs]

What's the other way to look at combination?

- You make a move.
- Your opponent responds to your move.
- ...

I already said it in a previous comment. I'll just copy+paste it here since you missed it last time:

There are many other ways of looking at combinations.

1. Your combination assumes a specific reaction, and if your opponent reacts differently you break it off. You create your next combination based on what you learned about your opponent.
2. Combination training is useful for learning how to combine techniques together (for example, a 1-2 punch instead of a jab, and later a cross). Knowing which techniques to use in the combination is a matter of reading your opponent.

I'm not saying the model is invalid. I'm just saying this doesn't prove the model is valid.

Using your example, let's say I go for a (groin kick, straight face punch) combo, under the assumption that the groin kick will drop my opponent's hands. My opponent uses his hip to block and keeps his hands up. Let's look for each option:

1. I may greedily go for the combo, in which case I will throw the straight punch into his guard. The next time I go for this combo, I will know he keeps his guard up, and opt for a different attack (such as a hook punch).
2. I may go for the combo, and read that my straight punch won't hit through the guard, and change to a hook punch. This isn't based on a tree starting from groin kick, this is just based on "guard beats straight punch". I'm not even going 2 levels deep, let alone 3 or more, because each technique is assessed on its own.

 

[Kung Fu Wang]

I already said it in a previous comment. I'll just copy+paste it here since you missed it last time:

There are many other ways of looking at combinations.

1. Your combination assumes a specific reaction, and if your opponent reacts differently you break it off. You create your next combination based on what you learned about your opponent.
2. Combination training is useful for learning how to combine techniques together (for example, a 1-2 punch instead of a jab, and later a cross). Knowing which techniques to use in the combination is a matter of reading your opponent.

I'm not saying the model is invalid. I'm just saying this doesn't prove the model is valid.

Using your example, let's say I go for a (groin kick, straight face punch) combo, under the assumption that the groin kick will drop my opponent's hands. My opponent uses his hip to block and keeps his hands up. Let's look for each option:

1. I may greedily go for the combo, in which case I will throw the straight punch into his guard. The next time I go for this combo, I will know he keeps his guard up, and opt for a different attack (such as a hook punch).
2. I may go for the combo, and read that my straight punch won't hit through the guard, and change to a hook punch. This isn't based on a tree starting from groin kick, this is just based on "guard beats straight punch". I'm not even going 2 levels deep, let alone 3 or more, because each technique is assessed on its own.

You may mis-understand what I was trying to say. I don't expect anything. If my 1st attack fail, I want to continue with my 2nd attack. My 2nd attack depends on the opportunity that you may give to me.

I still think we are talking about the same thing here. Let's use a more abstract example.

- You attack me.
- I step back.
- You step in, and attack me again.

What's the difference between my method and your method?

 

[skribs]

You may mis-understand what I was trying to say. I don't expect anything. If my 1st attack fail, I want to continue with my 2nd attack. My 2nd attack depends on the opportunity that you may give to me.

I still think we are talking about the same thing here. Let's use a more abstract example.

- You attack me.
- I step back.
- You step in, and attack me again.

What's the difference between my method and your method?

You're enumerating it and connecting everything together. I'm not. I'm not trying to draw a diagram of every possible reaction you can make, and then go into if I do A and you do A1 then I can do A1A, A1B, or A1C.

You stepped back. That is now irrelevant of whether I did technique A, B, C, D, E, F, or G. You stepped back. Now the position is that you're out of range, and I need to either close the range or get you to close the range. So you're in position 1. I can use A again, or B, or C, or any of the above. The difference is in how this sequence would look if analyzed:

	Enumerated Tree Method	Response
Attack 1	A	A
Response 1	A1	A1
Attack 2	A1B	1B
Response 2	A1B2	B2
Attack 3	A1B2C	2C
Response 3	A1B2C3	C3
Attack 4	A1B2C3D	3D
Response 4	A1B2C3D4	D4
Attack 5	A1B2C3D4E	4E

This is also why I don't think most people go more than so many levels deep. Because you only need a couple of layers, and then you string those together.

 

[Kung Fu Wang]

You're enumerating it and connecting everything together. I'm not. I'm not trying to draw a diagram of every possible reaction you can make, and then go into if I do A and you do A1 then I can do A1A, A1B, or A1C.

You stepped back. That is now irrelevant of whether I did technique A, B, C, D, E, F, or G. You stepped back. Now the position is that you're out of range, and I need to either close the range or get you to close the range. So you're in position 1. I can use A again, or B, or C, or any of the above. The difference is in how this sequence would look if analyzed:

	Enumerated Tree Method	Response
Attack 1	A	A
Response 1	A1	A1
Attack 2	A1B	1B
Response 2	A1B2	B2
Attack 3	A1B2C	2C
Response 3	A1B2C3	C3
Attack 4	A1B2C3D	3D
Response 4	A1B2C3D4	D4
Attack 5	A1B2C3D4E	4E

This is also why I don't think most people go more than so many levels deep. Because you only need a couple of layers, and then you string those together.

Let me use another simple example.

- You pull me (technique 1).
- If I resist, you take advantage on my resistance force, change your pull into a push (technique 2).
- If I yield, you take advantage on my yielding force, you pull me even harder (technique 2).

I don't think your approach and my approach are any different here.

Of course your technique 2 may depend on my respond. But that's just detail. The concept should be the same.

In other word, your technique 2 depends on my respond.

 

[skribs]

Let me use another simple example.

- You pull me (technique 1).
- If I resist, you take advantage on my resistance force, change your pull into a push (technique 2).
- If I yield, you take advantage on my yielding force, you pull me even harder (technique 2).

I don't think your approach and my approach are any different here.

Of course your technique 2 may depend on my respond. But that's just detail. The concept should be the same.

In other word, your technique 2 depends on my respond.

Yes. But then you go a step further and draw out the entire fight based on every possible response.

 

[Kung Fu Wang]

Yes. But then you go a step further and draw out the entire fight based on every possible response.

So, you agree with the concept. You just don't agree with going to the detail.

IMO, the detail can make us to understand MA better.

 

[skribs]

So, you agree with the concept. You just don't agree with going to the detail.

IMO, the detail can make us to understand MA better.

When you reduce anything to something ubiquitous, then of course everyone's going to "agree".

 

[punisher73]

Which one is better? Your opponent punches at you. You

1. dodge or block his punch.
2. attack his punching arm.

IMO, 1 < 2.

How long will it take for a beginner to figure this out?

Ugghh....

Artificial either/or response. That's the WHOLE point of training.

Which is better? It all depends on other factors.

Yes, option #2 would be the best because it ASSUMES a lot in combat already. That you were ready for the confrontation and were able to keep enough of a reactionary gap to be able to read/respond to the attack in time to attack the limb yourself. Option #1 is a better option when you don't have those combat assumptions already there like in #2.

 

[punisher73]

So, you agree with the concept. You just don't agree with going to the detail.

IMO, the detail can make us to understand MA better.

To quote Heraclitus, "No man ever steps in the same river twice, for it's not the same river and he's not the same man.”

What you are trying to do is assume that once an opponent "does something" against your attack you can keep planning out responses like a computer program of "if this=then this". WHY? It is pure mental masturbation. It might feel good doing this, but doesn't accomplish anything.

You can't plan on ALL contingencies. That is why systems are based on principles/strategies to implement the specific techniques/tactics. When you understand the design of the system, it allows you to create on the fly answers to those problems.

 

[isshinryuronin]

Idea is to really learn "a" and really learn "b" and learn how they can work together and what they do to both you and the other guy. Fights rarely end because you did "a" and I did "b." They end because you were able to apply "a" better than I was able to apply "b.

I fully agree with this. There is more than just "a" and "b." Which letter you choose may depend on your individual skills or inclination. Also, your hand and foot position or your balance at that particular second can play a major role in you first or second move. The are several ways to win in most any situation, usually there are multiple paths to victory.

You can't plan on ALL contingencies.

Right. This is impossible. But, IMO, it's good to drill for the high probability ones. Then, again, you want to be able to adapt to contingencies. What will happen if "a" misses - where will you be vulnerable? What if the opponent's "b" is a counter strike instead of the expected block? It's good to be able to change one's strike into a defensive move during delivery. Better not to get too hung up on specific letters.

"Expect nothing, be able to react to anything." Knowing the opponent (if and where possible), a relaxed spirit, muscle memory thru much drilling, and understanding the technique's physical properties and the lines of attack they may afford the opponent are all things that will help dealing with combat's serendipity.

 

[punisher73]

I fully agree with this. There is more than just "a" and "b." Which letter you choose may depend on your individual skills or inclination. Also, your hand and foot position or your balance at that particular second can play a major role in you first or second move. The are several ways to win in most any situation, usually there are multiple paths to victory.

Right. This is impossible. But, IMO, it's good to drill for the high probability ones. Then, again, you want to be able to adapt to contingencies. What will happen if "a" misses - where will you be vulnerable? What if the opponent's "b" is a counter strike instead of the expected block? It's good to be able to change one's strike into a defensive move during delivery. Better not to get too hung up on specific letters.

"Expect nothing, be able to react to anything." Knowing the opponent (if and where possible), a relaxed spirit, muscle memory thru much drilling, and understanding the technique's physical properties and the lines of attack they may afford the opponent are all things that will help dealing with combat's serendipity.

We are in agreement on high percentage probabilities and short term responses in training. I am just in disagreement taking it out so far out and mapping it all out to the Nth degree.