|
Disclaimer:
+ Show Spoiler +The law of averages is actually not a law at all. It's a truism that has no truth stating that when you've flipped a coin and gotten heads three times in a row, well then tails is more likely to turn up on your next flip. I just think that "The Law of Large Numbers" is a horrible title.
Any personal injury you sustain as a result of things such as but not exclusively slaps from angry girls through the use of your newfound knowledge is not to be blamed on me. If you succeed in any of your endeavors however, it is.
The law of averages states that the more times you try something, the more likely your chance of success is. This seems like common sense but most people don't employ it in their everyday lives.
![[image loading]](http://i.imgur.com/hS8Cd.png) graphs make everything better...
chicken chicken
To better see how the law of averages can be applied take this real life example:
I recently found a job because I kept applying to different jobs. It doesn't really matter if you fail to get the job twenty, even thirty times in a row, as long as you keep trying to apply, and don't get discouraged by expected failures, you will eventually find a job.
Now what's even better is that the law of averages can be applied to everything!: women, starcraft, sales, etc...
If you go to a party, you can go hit on every single girl and eventually you'll get lucky enough to find one girl either drunk or horny enough.
You can generally improve your chances by using the law of averages selectively.
nsfw
+ Show Spoiler [girl at party] +![[image loading]](http://i.imgur.com/lEZqj.jpg) Depending on your goals, this can be a good sign
Note that I haven't actually tried the law of averages out on women.
When you apply the law of averages, you don't want to sacrifice quality for quantity. You don't want to go up to every girl at the party and start grinding up against her on the dance floor. You'll make whomever you go after next feel like sloppy seconds. When you're applying for jobs, you don't want to send in a nonchalant cover letter either.
Even if a business is desperate for work, they won't hire someone who greets them with "sup boss dude."
+ Show Spoiler +
The point of using the law of averages is not to sabotage your own efforts by working really hard on applying to only one job, or expecting to gain ranks in starcraft 2 through analyzing vods all day. You want to combine quantity with quality for maximum effectiveness, and utilizing the law of averages is the best way to do it.
  
|
i loled at "sup boss dude"
|
This is how my brother and I used to get women (when we were a tad younger). We would go to a club and hit on the hottest girl in the room (in a non weird way... just ask if they wanted a drink or something), slowly working our way down in hotness (yes im aware this is a very immature and disrespectful thing to do lol)...... all our friends would spend the whole night working up the courage to speak to one girl and me and my bro would already be on our way out the door! They could never figure out how we did it, and we never told them!
|
Well i guess its true, if a monkey types randomly on a keyboard for a very long time, eventually the monkey will write Shakespeare.
|
infinity21
Canada6683 Posts
This thread doesn't make any sense to me.
The law of large numbers is about how the average of many independent repeatable experiments will tend to its expected value. Your example about getting a job after applying to 30 different companies is simple math of taking the limit of the CDF of the geometric distribution as the # of trials approach infinity. Of course your chances are quite high if you apply to a ton of jobs. But if you tell yourself that you're bound to find a job because you've been rejected by 30 companies is naive because the geometric distribution is memoryless.
I guess you're just happy that you found a job but you should be aware of what you're telling people.
|
On May 21 2011 04:52 obesechicken13 wrote:
The law of averages states that the more times you try something, the more likely your chance of success is.
if you are defining success as rolling a 6, this isnt true. if you are defining success as managing to roll the dice at all, well, congratulations. success and failure type experiments dont work like this.
rephrase please <3
|
Funny read, also an interesting juxtaposition of a classy vodka and .. not so classy partiers 
|
On May 21 2011 06:16 turdburgler wrote:Show nested quote +On May 21 2011 04:52 obesechicken13 wrote:
The law of averages states that the more times you try something, the more likely your chance of success is.
if you are defining success as rolling a 6, this isnt true. if you are defining success as managing to roll the dice at all, well, congratulations. success and failure type experiments dont work like this. rephrase please <3
Dude... Read the OP again. If success is rolling a 6, the more times you roll a die, the higher the chance you get a 6. That's basic statistics. Roll 1 die one time,you have 1/6th chances. Roll 1 die two times. You get 1/6th chances of success at A, 1/6th chances of success at B, probability that either event occurs equals .306, which is quite a bit higher than 1/6. Cheers.
|
infinity21
Canada6683 Posts
On May 21 2011 06:57 Sleight wrote:Show nested quote +On May 21 2011 06:16 turdburgler wrote:On May 21 2011 04:52 obesechicken13 wrote:
The law of averages states that the more times you try something, the more likely your chance of success is.
if you are defining success as rolling a 6, this isnt true. if you are defining success as managing to roll the dice at all, well, congratulations. success and failure type experiments dont work like this. rephrase please <3 Dude... Read the OP again. If success is rolling a 6, the more times you roll a die, the higher the chance you get a 6. That's basic statistics. Roll 1 die one time,you have 1/6th chances. Roll 1 die two times. You get 1/6th chances of success at A, 1/6th chances of success at B, probability that either event occurs equals .306, which is quite a bit higher than 1/6. Cheers.
Chance of success in rolling a die typically refers to the chances of 1 independent event. And the law of large numbers has nothing to do with either of those, it's basic probability.
What the OP probably meant was that as the number of trials increase, the chances of seeing one 'success' increases.
|
On May 21 2011 07:14 infinity21 wrote:Show nested quote +On May 21 2011 06:57 Sleight wrote:On May 21 2011 06:16 turdburgler wrote:On May 21 2011 04:52 obesechicken13 wrote:
The law of averages states that the more times you try something, the more likely your chance of success is.
if you are defining success as rolling a 6, this isnt true. if you are defining success as managing to roll the dice at all, well, congratulations. success and failure type experiments dont work like this. rephrase please <3 Dude... Read the OP again. If success is rolling a 6, the more times you roll a die, the higher the chance you get a 6. That's basic statistics. Roll 1 die one time,you have 1/6th chances. Roll 1 die two times. You get 1/6th chances of success at A, 1/6th chances of success at B, probability that either event occurs equals .306, which is quite a bit higher than 1/6. Cheers. Chance of success in rolling a die typically refers to the chances of 1 independent event. And the law of large numbers has nothing to do with either of those, it's basic probability. What the OP probably meant was that as the number of trials increase, the chances of seeing one 'success' increases.
So if 6 is success and 54 times in a row you've gotten 1-5 and because the law of large numbers says everything will even out to the 1/6 eventually, you have a higher chance of getting a six the more you roll. I'm not actually familiar with the law, this is what I'm inferring
|
On May 21 2011 06:57 Sleight wrote:Show nested quote +On May 21 2011 06:16 turdburgler wrote:On May 21 2011 04:52 obesechicken13 wrote:
The law of averages states that the more times you try something, the more likely your chance of success is.
if you are defining success as rolling a 6, this isnt true. if you are defining success as managing to roll the dice at all, well, congratulations. success and failure type experiments dont work like this. rephrase please <3 Dude... Read the OP again. If success is rolling a 6, the more times you roll a die, the higher the chance you get a 6. That's basic statistics. Roll 1 die one time,you have 1/6th chances. Roll 1 die two times. You get 1/6th chances of success at A, 1/6th chances of success at B, probability that either event occurs equals .306, which is quite a bit higher than 1/6. Cheers.
Probability is a tricky thing. If you roll a dice your probability of getting a 6 is 1/6. The next time you roll the probability is the same! This is because the events are uncorrelated. When events are uncorrelated you actually don't increase your probability for success the next time you try. So if you rolled a dice hoping for a 6 say 100 times and didn't get it then this hasn't increased your chances the next time you try (but that you didn't roll a 6 with 100 tries is a very seldom event and one has to be very unlucky!). But your strategy is still the correct one since you have no chance if you don't try - same is true for girls.
Anyhow funny post.
|
While this method may help you develop the courage to talk to attractive women, I merely see it as just that, a method of motivation. While you may think that on average the more times you roll a dice, the more likely you are to get a six, all these events are independent of one another. While you do put simple probability on your side, that does not mean that it will actually work. You could be the unlucky guy to get struck down by every women you ask out. You could roll a dice 100 times and never get a 6, sometimes life is just tough.
EDIT: I liked the chicken chicken reference.
|
So let me check that I get this. Obviously each separate event in unrelated to the one before it but we know that the event has some finite chance of success and, if performed an infinite number of times, the success/failure rate will be whatever our finite chance is (probably why this is called the law of large numbers). So when we performed multiple trials with more failures than success we know that sometime in our infinite tests we'll have to get more success than usual to regress to the mean and thus our chance is higher even if the events have no impact on one another. Which is actually really intuitive and applicable but sounds paradoxical.
|
infinity21
Canada6683 Posts
On May 21 2011 07:23 n.DieJokes wrote:Show nested quote +On May 21 2011 07:14 infinity21 wrote:On May 21 2011 06:57 Sleight wrote:On May 21 2011 06:16 turdburgler wrote:On May 21 2011 04:52 obesechicken13 wrote:
The law of averages states that the more times you try something, the more likely your chance of success is.
if you are defining success as rolling a 6, this isnt true. if you are defining success as managing to roll the dice at all, well, congratulations. success and failure type experiments dont work like this. rephrase please <3 Dude... Read the OP again. If success is rolling a 6, the more times you roll a die, the higher the chance you get a 6. That's basic statistics. Roll 1 die one time,you have 1/6th chances. Roll 1 die two times. You get 1/6th chances of success at A, 1/6th chances of success at B, probability that either event occurs equals .306, which is quite a bit higher than 1/6. Cheers. Chance of success in rolling a die typically refers to the chances of 1 independent event. And the law of large numbers has nothing to do with either of those, it's basic probability. What the OP probably meant was that as the number of trials increase, the chances of seeing one 'success' increases. So if 6 is success and 54 times in a row you've gotten 1-5 and because the law of large numbers says everything will even out to the 1/6 eventually, you have a higher chance of getting a six the more you roll. I'm not actually familiar with the law, this is what I'm inferring
No, that's the popular misconception. What it really means is, even if you roll no sixes for your first 6 rolls, if you roll the dice 6000 times, you will see approximately 1000 6s because the chances of it diverging significantly from 1000 is very low.
Basically, having a bad start is insignificant when it comes to large # of trials as those few odd results will get dwarfed by the rest of the instances.
|
On May 21 2011 07:35 infinity21 wrote:Show nested quote +On May 21 2011 07:23 n.DieJokes wrote:On May 21 2011 07:14 infinity21 wrote:On May 21 2011 06:57 Sleight wrote:On May 21 2011 06:16 turdburgler wrote:On May 21 2011 04:52 obesechicken13 wrote:
The law of averages states that the more times you try something, the more likely your chance of success is.
if you are defining success as rolling a 6, this isnt true. if you are defining success as managing to roll the dice at all, well, congratulations. success and failure type experiments dont work like this. rephrase please <3 Dude... Read the OP again. If success is rolling a 6, the more times you roll a die, the higher the chance you get a 6. That's basic statistics. Roll 1 die one time,you have 1/6th chances. Roll 1 die two times. You get 1/6th chances of success at A, 1/6th chances of success at B, probability that either event occurs equals .306, which is quite a bit higher than 1/6. Cheers. Chance of success in rolling a die typically refers to the chances of 1 independent event. And the law of large numbers has nothing to do with either of those, it's basic probability. What the OP probably meant was that as the number of trials increase, the chances of seeing one 'success' increases. So if 6 is success and 54 times in a row you've gotten 1-5 and because the law of large numbers says everything will even out to the 1/6 eventually, you have a higher chance of getting a six the more you roll. I'm not actually familiar with the law, this is what I'm inferring No, that's the popular misconception. What it really means is, even if you roll no sixes for your first 6 rolls, if you roll the dice 6000 times, you will see approximately 1000 6s because the chances of it diverging significantly from 1000 is very low.
Isn't that what I said? If you don't get any 6's for some significantly large period and if you roll 1000 6's in 6000 you have to roll more 6's after the dry spell and thus have a slightly higher chance of success
|
Hopefully people don't confuse this with gambler's fallacy.
|
infinity21
Canada6683 Posts
On May 21 2011 07:32 n.DieJokes wrote:
So let me check that I get this. Obviously each separate event in unrelated to the one before it but we know that the event has some finite chance of success and, if performed an infinite number of times, the success/failure rate will be whatever our finite chance is (probably why this is called the law of large numbers). So when we performed multiple trials with more failures than success we know that sometime in our infinite tests we'll have to get more success than usual to regress to the mean and thus our chance is higher even if the events have no impact on one another. Which is actually really intuitive and applicable but sounds paradoxical.
No the chances of you rolling a 6 doesn't change based on previous results. That's why the events are called independent. By definition of independence, what happened in the past cannot influence the probability of your throw.
Here's some light reading:
http://en.wikipedia.org/wiki/Gambler's_fallacy
|
infinity21
Canada6683 Posts
On May 21 2011 07:38 n.DieJokes wrote:Show nested quote +On May 21 2011 07:35 infinity21 wrote:On May 21 2011 07:23 n.DieJokes wrote:On May 21 2011 07:14 infinity21 wrote:On May 21 2011 06:57 Sleight wrote:On May 21 2011 06:16 turdburgler wrote:On May 21 2011 04:52 obesechicken13 wrote:
The law of averages states that the more times you try something, the more likely your chance of success is.
if you are defining success as rolling a 6, this isnt true. if you are defining success as managing to roll the dice at all, well, congratulations. success and failure type experiments dont work like this. rephrase please <3 Dude... Read the OP again. If success is rolling a 6, the more times you roll a die, the higher the chance you get a 6. That's basic statistics. Roll 1 die one time,you have 1/6th chances. Roll 1 die two times. You get 1/6th chances of success at A, 1/6th chances of success at B, probability that either event occurs equals .306, which is quite a bit higher than 1/6. Cheers. Chance of success in rolling a die typically refers to the chances of 1 independent event. And the law of large numbers has nothing to do with either of those, it's basic probability. What the OP probably meant was that as the number of trials increase, the chances of seeing one 'success' increases. So if 6 is success and 54 times in a row you've gotten 1-5 and because the law of large numbers says everything will even out to the 1/6 eventually, you have a higher chance of getting a six the more you roll. I'm not actually familiar with the law, this is what I'm inferring No, that's the popular misconception. What it really means is, even if you roll no sixes for your first 6 rolls, if you roll the dice 6000 times, you will see approximately 1000 6s because the chances of it diverging significantly from 1000 is very low. Isn't that what I said? If you don't get any 6's for some significantly large period and if you roll 1000 6's in 6000 you have to roll more 6's after the dry spell and thus have a slightly higher chance of success
In my example, the expected value would be 999, which is basically going to be the same as having an expected value of 1000. So this seemingly significant event of not rolling a single 6 for 6 or even 12 throws is not significant over a sample of 6000 throws.
|
On May 21 2011 07:44 infinity21 wrote:Show nested quote +On May 21 2011 07:38 n.DieJokes wrote:On May 21 2011 07:35 infinity21 wrote:On May 21 2011 07:23 n.DieJokes wrote:On May 21 2011 07:14 infinity21 wrote:On May 21 2011 06:57 Sleight wrote:On May 21 2011 06:16 turdburgler wrote:On May 21 2011 04:52 obesechicken13 wrote:
The law of averages states that the more times you try something, the more likely your chance of success is.
if you are defining success as rolling a 6, this isnt true. if you are defining success as managing to roll the dice at all, well, congratulations. success and failure type experiments dont work like this. rephrase please <3 Dude... Read the OP again. If success is rolling a 6, the more times you roll a die, the higher the chance you get a 6. That's basic statistics. Roll 1 die one time,you have 1/6th chances. Roll 1 die two times. You get 1/6th chances of success at A, 1/6th chances of success at B, probability that either event occurs equals .306, which is quite a bit higher than 1/6. Cheers. Chance of success in rolling a die typically refers to the chances of 1 independent event. And the law of large numbers has nothing to do with either of those, it's basic probability. What the OP probably meant was that as the number of trials increase, the chances of seeing one 'success' increases. So if 6 is success and 54 times in a row you've gotten 1-5 and because the law of large numbers says everything will even out to the 1/6 eventually, you have a higher chance of getting a six the more you roll. I'm not actually familiar with the law, this is what I'm inferring No, that's the popular misconception. What it really means is, even if you roll no sixes for your first 6 rolls, if you roll the dice 6000 times, you will see approximately 1000 6s because the chances of it diverging significantly from 1000 is very low. Isn't that what I said? If you don't get any 6's for some significantly large period and if you roll 1000 6's in 6000 you have to roll more 6's after the dry spell and thus have a slightly higher chance of success In my example, the expected value would be 999, which is basically going to be the same as having an expected value of 1000. So this seemingly significant event of not rolling a single 6 for 6 or even 12 throws is not significant over a sample of 6000 throws.
lol kk. Then what is this blog saying?
|
infinity21
Canada6683 Posts
It's saying to never give up and keep working towards your goal because you'll eventually reach it.
|
Straight outta Johto18973 Posts
The statistical fallacies in this blog makes me sad. I understand how the message is supposed to work, but the execution was incredibly dubious.
|
Remake the blog, the more you try, the more likely you are to get it right sooner or later.
I learned that from a blog once.
|
Do you understand that there is a probability that you will never roll a 6 after an infinite amount of casts?
|
On May 21 2011 07:59 Lemonwalrus wrote:
Remake the blog, the more you try, the more likely you are to get it right sooner or later.
I learned that from a blog once.
Eh, you guys figured out what I was saying eventually.
On May 21 2011 10:24 EsX_Raptor wrote:
Do you understand that there is a probability that you will never roll a 6 after an infinite amount of casts?
No there isn't  . pfft infinity.
|
Guys, you really don't understand probability theory.
Two dice, any chance of getting one single 6 is the same as rolling one die twice to get a six.
6 * 6 = 36 Outcomes
Winning outcomes = [6 X] and [X 6] with [6 6] only counting once, obviously, so 11 wins.
11 / 36 = .306
You guys need to do better math. Rolling more dice DOES increase the probability of a success. Taking more chances increases the likelihood of a success as a strategy.
EDIT for clarification: No single roll EVER has a higher likelihood than 1/6, but you can only examine probability on the large scale. We are looking at combinations of chances, which mean that if you roll 1000 dice, your likelihood of getting a 6 is basically 1, but that says nothing of whether its the first roll or the last roll. Probability theory ONLY applies to SETS.
|
16953 Posts
On May 21 2011 10:24 EsX_Raptor wrote:
Do you understand that there is a probability that you will never roll a 6 after an infinite amount of casts?
The probability is zero, because you will almost surely roll a 6 in an infinite amount of casts.
The amount of statistical ignorance in the general population is staggering 
EDIT: Before someone replies, just because an event occurs with probability zero doesn't mean it can't occur. And I'm using the term "almost surely" in its technical meaning, in case there were any doubts.
|
Straight outta Johto18973 Posts
On May 21 2011 11:21 Sleight wrote:
Guys, you really don't understand probability theory.
Two dice, any chance of getting one single 6 is the same as rolling one die twice to get a six.
6 * 6 = 36 Outcomes
Winning outcomes = [6 X] and [X 6] with [6 6] only counting once, obviously, so 11 wins.
11 / 36 = .306
You guys need to do better math. Rolling more dice DOES increase the probability of a success. Taking more chances increases the likelihood of a success as a strategy.
EDIT for clarification: No single roll EVER has a higher likelihood than 1/6, but you can only examine probability on the large scale. We are looking at combinations of chances, which mean that if you roll 1000 dice, your likelihood of getting a 6 is basically 1, but that says nothing of whether its the first roll or the last roll. Probability theory ONLY applies to SETS.
I'm not sure what point you are trying to make with your first example of rolling two dice concurrently. But if you are saying that the chance getting a 6 when rolling two dice at the same time is different from the probability of obtaining a 6 when rolling a single die is different then the result is a given. In that instance you are comparing a joint probability distribution to a single variable probability distribution so the results will be skewed as a result.
You are misunderstanding the point that Infinity was making. There is nothing wrong with observing probability on a small scale. Statisticians do this all the time.
The "Law of Large Numbers" (I do detest the name but no matter) is a statement that a very large number of observations will tend to reflect the underlying probability of the random variable in question. On this instance, I believe we are all in agreement. If you flip a fair coin an infinite number of times, your observations will tend to a 50:50 split.
However, you are confusing this with the applications and caveats that come with independence and random sampling. The OP is not making a statement about deducing conclusion from an extremely large sample space. Instead, it is a case of repeated trials and commenting on the results of each observation. In this case, it is a conditional probability function.
For example, one example was continuing to hit on women at a party until you achieve success. It would be incorrect to state that we are making observations on a large scale. Instead, you are taking repeated samples from a state space and then commenting on each observation one at a time. The fallacy in his article was "Given that I have made advances on a woman and failed, I expect to have a success because the law of averages should even things out eventually." This is incorrect because independence between events means that the conditional probability of success given failure is still the same as the probability of success in a single event.
I would recommend you read the link that Infinity posted and not make claims such as "You don't understand probability theory".
|
I pointed out exactly how the Gambler's fallacy doesn't account for reality, friend. I also said probability only works on SETS as my last statement. Two instances is a set, even if its largely useless. Two hundred instances and we begin to find relevant information to inform a decision, typically.
Notice how I explicitly say your single instance probably NEVER CHANGES?
Did you read my post before saying the same thing in a different way and trying to sound contradictory? I said nothing about the OP's correctness in regards to his Gambler's Fallacy. I spoke only to the event that increased trials increases chance of success. He is right, statistically you will eventually hit it home, but he is wrong in that it is some magical self-correction. It's just probability, nothing more, nothing less.
Like I said and you said and other informed people have said.
|
Straight outta Johto18973 Posts
On May 21 2011 12:25 Sleight wrote:
I pointed out exactly how the Gambler's fallacy doesn't account for reality, friend. I also said probability only works on SETS as my last statement. Two instances is a set, even if its largely useless. Two hundred instances and we begin to find relevant information to inform a decision, typically.
Notice how I explicitly say your single instance probably NEVER CHANGES?
Did you read my post before saying the same thing in a different way and trying to sound contradictory? I said nothing about the OP's correctness in regards to his Gambler's Fallacy. I spoke only to the event that increased trials increases chance of success. He is right, statistically you will eventually hit it home, but he is wrong in that it is some magical self-correction. It's just probability, nothing more, nothing less.
Like I said and you said and other informed people have said.
There are two disagreements with your argument I have.
The first is "Rolling more dice DOES increase the probability of a success." If we consider rolling die to be a Bernoulli or Binomial process, then each roll has the same probability. The probability of success on each roll hasn't changed. Increased trials does not "increase the chance of success" as you say. It simply reduces the effect of randomness and chance from your samples in the long run. What repeated samples means is that the distribution of your sample results will converge to the underlying distribution.
The second disagreement is that "you can only examine probability on the large scale". You can model any event regardless of how large your sample size is.
|
You have quit reading what I said, clearly.
Again, for the record. Increasing trials DOES NOT increase the probability of ANY SINGLE INSTANCE.
Got it? I've said that each and every post. Rolling a die 10k times does not increase the chance of getting a 6 on any given trial.
You are wrong to say that the probability of getting a 6 rolling a dice once is the same as rolling a dice 10k times. Dead wrong. Which is what I have said over and over. You are COMPLETELY WRONG if defining this as a Bernoulli process formally changes ANYTHING from what I've said.
Let's define our process as such. If a die rolls a non-6, we score it a 0. If the die is a 6 it is a 1, p = 1/6. Let's run a trial on that for one time.
We will say P(1) = (1 choose 1)*.16^1*.84*0 = .16. The chance of getting a 6 in 1 roll.
We will ask what are the chances of getting at least 1 6 in 2 trials.
P(1)=(2 choose 1)*.16^1*.84^1 = .268.
Does this account for all outcomes? NO! We still need to consider 2 successes.
P(2)=(2 choose2)*.16^2*.84^1 = .022
P of at least 1 success = .268+.022 = .29 (i rounded .166 to .16, so its marginally lower).
You are unequivocally wrong that changing the model somehow stops more trials from increasing chance of success.
|
On May 21 2011 12:15 MoonBear wrote:Show nested quote +On May 21 2011 11:21 Sleight wrote:
Guys, you really don't understand probability theory.
Two dice, any chance of getting one single 6 is the same as rolling one die twice to get a six.
6 * 6 = 36 Outcomes
Winning outcomes = [6 X] and [X 6] with [6 6] only counting once, obviously, so 11 wins.
11 / 36 = .306
You guys need to do better math. Rolling more dice DOES increase the probability of a success. Taking more chances increases the likelihood of a success as a strategy.
EDIT for clarification: No single roll EVER has a higher likelihood than 1/6, but you can only examine probability on the large scale. We are looking at combinations of chances, which mean that if you roll 1000 dice, your likelihood of getting a 6 is basically 1, but that says nothing of whether its the first roll or the last roll. Probability theory ONLY applies to SETS. I'm not sure what point you are trying to make with your first example of rolling two dice concurrently. But if you are saying that the chance getting a 6 when rolling two dice at the same time is different from the probability of obtaining a 6 when rolling a single die is different then the result is a given. In that instance you are comparing a joint probability distribution to a single variable probability distribution so the results will be skewed as a result. You are misunderstanding the point that Infinity was making. There is nothing wrong with observing probability on a small scale. Statisticians do this all the time. The "Law of Large Numbers" (I do detest the name but no matter) is a statement that a very large number of observations will tend to reflect the underlying probability of the random variable in question. On this instance, I believe we are all in agreement. If you flip a fair coin an infinite number of times, your observations will tend to a 50:50 split. However, you are confusing this with the applications and caveats that come with independence and random sampling. The OP is not making a statement about deducing conclusion from an extremely large sample space. Instead, it is a case of repeated trials and commenting on the results of each observation. In this case, it is a conditional probability function. For example, one example was continuing to hit on women at a party until you achieve success. It would be incorrect to state that we are making observations on a large scale. Instead, you are taking repeated samples from a state space and then commenting on each observation one at a time.
Hmmm... ok... ok
The fallacy in his article was "Given that I have made advances on a woman and failed, I expect to have a success because the law of averages should even things out eventually." This is incorrect because independence between events means that the conditional probability of success given failure is still the same as the probability of success in a single event. I would recommend you read the link that Infinity posted and not make claims such as "You don't understand probability theory".
Hey I never said that!
+ Show Spoiler +
And I quote
The law of averages is actually not a law at all. It's a truism that has no truth stating that when you've flipped a coin and gotten heads three times in a row, well then tails is more likely to turn up on your next flip. I just think that "The Law of Large Numbers" is a horrible title.
My intention wasn't to say that I would have a greater chance of success on my next attempt, I was saying that if I got turned down 20 times, and then approached 50 more women, I would have a higher chance of success with at least one of those 50 women than if I felt discouraged and only approached two women. I'm trying to say don't give up. I understand "independece between events" or whatever
|
I was just going to type up something, then I saw two pages of comments
<3 TL
|
infinity21
Canada6683 Posts
You guys are literally arguing over the difference between 'chances of success' and 'chances of a success' lol
|
16953 Posts
On May 21 2011 14:44 infinity21 wrote:
You guys are literally arguing over the difference between 'chances of success' and 'chances of a success' lol
It's not worth trying.
|
You guys are missing the whole point of the blog. To use the dice analogy, the point is if you roll the die 100 times you're more likely to get a 6 in AT LEAST one of those 100 rolls than if you roll the die once. Which he applies to things like hooking up with girls, getting a job, etc.
|
16953 Posts
Right. But that's not what the law of averages is.
|
On May 21 2011 13:30 obesechicken13 wrote:Show nested quote +On May 21 2011 12:15 MoonBear wrote:On May 21 2011 11:21 Sleight wrote:
Guys, you really don't understand probability theory.
Two dice, any chance of getting one single 6 is the same as rolling one die twice to get a six.
6 * 6 = 36 Outcomes
Winning outcomes = [6 X] and [X 6] with [6 6] only counting once, obviously, so 11 wins.
11 / 36 = .306
You guys need to do better math. Rolling more dice DOES increase the probability of a success. Taking more chances increases the likelihood of a success as a strategy.
EDIT for clarification: No single roll EVER has a higher likelihood than 1/6, but you can only examine probability on the large scale. We are looking at combinations of chances, which mean that if you roll 1000 dice, your likelihood of getting a 6 is basically 1, but that says nothing of whether its the first roll or the last roll. Probability theory ONLY applies to SETS. I'm not sure what point you are trying to make with your first example of rolling two dice concurrently. But if you are saying that the chance getting a 6 when rolling two dice at the same time is different from the probability of obtaining a 6 when rolling a single die is different then the result is a given. In that instance you are comparing a joint probability distribution to a single variable probability distribution so the results will be skewed as a result. You are misunderstanding the point that Infinity was making. There is nothing wrong with observing probability on a small scale. Statisticians do this all the time. The "Law of Large Numbers" (I do detest the name but no matter) is a statement that a very large number of observations will tend to reflect the underlying probability of the random variable in question. On this instance, I believe we are all in agreement. If you flip a fair coin an infinite number of times, your observations will tend to a 50:50 split. However, you are confusing this with the applications and caveats that come with independence and random sampling. The OP is not making a statement about deducing conclusion from an extremely large sample space. Instead, it is a case of repeated trials and commenting on the results of each observation. In this case, it is a conditional probability function. For example, one example was continuing to hit on women at a party until you achieve success. It would be incorrect to state that we are making observations on a large scale. Instead, you are taking repeated samples from a state space and then commenting on each observation one at a time. Hmmm... ok... ok Show nested quote +The fallacy in his article was "Given that I have made advances on a woman and failed, I expect to have a success because the law of averages should even things out eventually." This is incorrect because independence between events means that the conditional probability of success given failure is still the same as the probability of success in a single event. I would recommend you read the link that Infinity posted and not make claims such as "You don't understand probability theory". Hey I never said that! + Show Spoiler +And I quote Show nested quote +The law of averages is actually not a law at all. It's a truism that has no truth stating that when you've flipped a coin and gotten heads three times in a row, well then tails is more likely to turn up on your next flip. I just think that "The Law of Large Numbers" is a horrible title. My intention wasn't to say that I would have a greater chance of success on my next attempt, I was saying that if I got turned down 20 times, and then approached 50 more women, I would have a higher chance of success with at least one of those 50 women than if I felt discouraged and only approached two women. I'm trying to say don't give up. I understand "independece between events" or whatever
Oh I wish you were right and that success in life was proportional to the number of times you try. The problem though is that if you fuck up badly then the probability to succeed next time you try will decrease. Or the other way around. Real life events usually aren't independent and you can fail if you behave stupidly :-).
EDIT: One may say that this is true for sc2 as well. Trying a lot is a good thing but if you don't analyze your games, optimize your build orders and create new ones well.. then there is a chance that you aren't improving or perhaps even get bad habits.
|
|
|
|
|
|
EPT![[image loading]](http://i.imgur.com/nEcOf.jpg)
![[image loading]](http://i.imgur.com/cz47t.jpg)
![[image loading]](http://i.imgur.com/6AVIV.jpg)
