Re: Boy Scouts make people nervous

On 20 Mar 2006 14:00:39 -0800, "Evil Shaman" <parker7273@xxxxxxxxxxx>
wrote:


hal@xxxxxxxxxx wrote:
On 20 Mar 2006 11:38:53 -0800, "Evil Shaman" <parker7273@xxxxxxxxxxx>
wrote:


If an opponent dodges right or left, 50% of the time he will dodge
left, 50% of the time he will dodge right. Right?


Not necessarily no. What makes you think each direction has an equal
probability of occurring? Do you know of any actual surveys or studies
backing this up? Do you know for a fact that hand dominance doesn't
play a role for instance?

No I do not know that.

Then you probably shouldn't state it as fact.

I didn't. I never said hand dominance would effect the outcome either
way. I am saying we don't know if it will, therefore we cannot assume
it does without adequate data. Do you have any such data?


In lieu of such information, you cannot assume
it will skew your results, therefore cannot assume a change without
sampling.

Assume a change from what?

skewing from the 50-50 predicted outcome.

In lieu of information you cannot assume
any probability. You just don't know.

Yes, you can, based on the number of possible outcomes. That would be
what statistics is for.


Now on the other hand, if in fact, dominance did have a
correlation to direction, and we knew that, and we knew the opps
dominance, then we would have adequate information to make a more
accuract prediction.

Really? Sounds like that might make an interesting topic of
discussion. Kirk should put it in his WOTT list.

Sure, if someone had a documented correlation between dominance and
movement. I just don't think there is any.


Since we don't know that, we must assume equal
likihood of occurance.

Even a coin toss isn't necessarily 50/50. I'm pretty sure I read
somewhere that there is like a 1% bias in favor of it landing on the
same side it was flipped from.

Yes, true, in the case of a real coin, but the variance in the real
coin is not really significant because the coin is simply being used
as the most common of the two outcome model.

The "fair coin" is used as an example of equal probability. Having two
outcomes does not necessarily mean equal probability.

Uh, yes, it does, sorry. Basic concept of probs. The simple fact
that you might not get both represented equally, does not mean that is
not the most likely outcome. There are exceptions, and there is a
margin of error.


I you want to keep it
exact, have a dummy coin exactly balanced, then it will be exactly
50-50 over an adequate sample size. The bigger the sample, the less
your error.


Yes, it is right, and if you do an adequate sample size, that WILL be
you're result, because both have an equal probability of occuring,
therefore they will be equal.


You know this for an absolute fact how?


Because probabilities tells us it will:

P(event)= The number of ways an event can occur
/The Total Number Of Possible Outcomes


You don't know that they have an equal possibility of occurring because
you don't know the distribution of outcomes. For example, if their are
blue and red marbles in the jar, seven red and three blue, there are
two outcomes but the probability of which one you will draw is not
50/50. It is 70/30. the greater your sampling the closer, you will
get to 70/30, not 50/50.

<sigh>

wrong model.

http://www.mathgoodies.com/lessons/vol6/intro_probability.html


In the dodging example, you can't even assume 2 possible outcomes.

True, but the premise was based on which way, left or right. If you
want to add multi axis movement, your chance of error just went way
up.

For
example on outcome could be a right hander dodging right, another a
left hander dodging right. Or maybe you have a male, left hander
dodging left against an incoming left hand jab. There are tons of
possible outcomes. If you just want to calculate right or left, you
have to have an idea how many times right or left occurs over all the
various possible outcomes.

It's still just either left or right. Mitigating factors certainly
are possible, but until we have hard data to support skewing our
results, P(left)=1/2=P(right).

Hal


Bryce

.

Relevant Pages

  • Re: Boy Scouts make people nervous
    ... any probability. ... Yes, true, in the case of a real coin, but the variance in the real ... outcomes does not necessarily mean equal probability. ... left hander dodging right. ...
    (rec.martial-arts)
  • Re: Boy Scouts make people nervous
    ... You keep conflating number of states with number of outcomes. ... It's the number of states that determine probability, ... Even with a coin flip you have ... Talk about more overcomplicating bullshit, ...
    (rec.martial-arts)
  • Re: Boy Scouts make people nervous
    ... You keep conflating number of states with number of outcomes. ... Even with a coin ... In principle you can work probability problems either way and get ... does not have the same symmetries. ...
    (rec.martial-arts)
  • Re: Boy Scouts make people nervous
    ... any probability. ... coin is not really significant because the coin is simply being used ... outcomes does not necessarily mean equal probability. ... left hander dodging right. ...
    (rec.martial-arts)
  • Re: Boy Scouts make people nervous
    ... any probability. ... coin is not really significant because the coin is simply being used ... outcomes does not necessarily mean equal probability. ...
    (rec.martial-arts)