Vlar's Offical List of Memorable IRC Quotes!

Averages are a great way to analyze the past to calculate probability for the future. The only guarantee they provide however is the low and the high end of the spectrum. So you can be sure your weapon won't hit less than 5 or greater than 45 but you have to swing the weapon and watch the screen for the result in the actual instance.

I had this argument with my D&D friends a long time ago when my DM got mad at me for rolling the 1's out of my dice.

But in all seriousness math can go die in a mongbat felching incident.

"Are you such a dreamer, to put the world to rights? I'll stay home forever, where 2 and 2 Always makes a 5."

I wonder if @Gnarl considers Thom Yorke a poet... I think he's cross eyed or something. Thom Yorke not Gnarl...

Please go on test and show me any 1000 swing samples of the same weapon type that have anywhere NEAR that much variance. The larger the sample size the closer the results will be to the mean. Yes of course a single swing can be anywhere in the dice roll range, hence large sample sizes. Just test it and post the results and you will be redeemed!

Now there's two people who are wrong here :/

I'll take this from a few angles, and if this doesn't work I'm calling in Gideon.

1) Why would you use a sample of 1000 swings, if the result is completely random? The whole point of using 1000 swings as a sample (instead of 1 swing) is because it gives you an average that is less prone to variance.

2) Why do pollsters waste time calling 1000 people to ask who they're going to vote for in the presidential election? If you were right, they'd just ask one person and call it a day.

3) Why 1000 swings, and not 10,000? Or 1,000,000? If you took two samples of 1,000,000 swings, would you still get an average of 30 in the first one and an average of 15 in the second? Because if you actually think that, then you literally cannot tell one weapon apart from another. Just because a halberd hit harder than a kryss the last 1,000,000 times they were each swung, wouldn't guarantee that would happen in the future.

4) Look at baseball players. Guy hits .300 for the first half of the season. What is he going to hit in the second half? About .300. Maybe .250, maybe .350. Not .700.

5) Your entire argument can be disproven in about 5 minutes on Test Center. Get a Combat Log Book, get two identical weapons, and log 100 hits with each one. Your average damage will be within a few points, guaranteed.

6) if you really believe what you said, you'd put your money where your mouth is. You pick the weapon that does 30 damage for 1,000 hits (let's say halberd). Then we log it once more. If the next 1,000 is between 28 and 32, I win. If it's ANY OTHER NUMBER, you win. How much shall we wager?

Ok we arent calculating the probability of a hispanic walking into a walmart in Rio Rancho on saturday at 3:00 p.m. we are talking about averaging random rolls. You can average randomness with larger and larger samples but the fact of the matter is each time you run a number it is still going to be random..

Ok so let's let actual math do the talking here.

I ran data sets of 1, 10, 100, and 1000 random rolls with a minimum of 10 and maximum of 50. Two samples. Here are the averages of those results;

1 roll:
A. 12
B. 31

10 rolls:
A. 22
B. 28

100 rolls:
A. 34
B. 25

1000 rolls:
A. 32
B. 36


After this once could maybe average a and b to get double the sample size? I guess the point here is fluctuation. There is fluctuation with all sample sizes.

I'm not saying 1 swing is as accurate as average of 1000 swings I'm saying merely that deriving a static dps figure off a random number generator is near impossible. I'd be curious to see what 1 million swings would be like.. but if u sampled 1 million swings a week later the number may still fluctuate.

If you want to sit and watch and document 1 million swings go for it. You obviously have the time on hand since you won't stop trying to save face


As for the baseball player analogy.. batting averages arent random and neither are political opinions.. those can't be used to argue rng.

1) Never claimed result was random I said it would be in the middle.

2) lol I didn't disagree with this either. All I said was after calling 1000 people, they could not predict what the 1001st person would say.

3) why not 1k vs 10k vs 100k? no reason other than it's a lot of work i suppose... should I repeat my point about what any of this says for the next swing?

4) I agree.

5 and 6 again have nothing to do with what I said and I believe you are selectively reading Hammers argument. Hammers argument rests on "may" and not "will". He says "it may result in widely varying results".

Massive derail....

Halabinder, the only thing you said that was wrong was that he wasn't wrong. My post wasn't aimed at you.

I'll call BS on this.

If you seriously think there's some difference between doing 1,000 RNG and doing 1,000,000 RNG with any modern machine, you're even more delusional than I thought. Here's a bunch of trials which will show you why I think you made an error in how you set this up, as 32 and 36 within 1,000 samples is an absurdly unlikely result. Here's me running sample "A" and "B" a bunch of times:

1 sample 44 15
10 samples 32 30
100 samples 30 29
1000 samples 30 30

1 sample 34 48
10 samples 32 34
100 samples 30 29
1000 samples 30 30

1 sample 46 14
10 samples 30 36
100 samples 28 31
1000 samples 29 30

1 sample 16 43
10 samples 26 29
100 samples 30 27
1000 samples 29 30

1 sample 46 49
10 samples 30 36
100 samples 29 30
1000 samples 30 30

Finally, I'll repeat this:

I'm sure we can come up with a wager that's worth your time.

OK that's fair you guys can argue all you want about variance. I'm gonna find a post by HateCrime and stare at his signature for 15 minutes. Mordechai's new one is almost as good but still...

.

.

I did above (well, Excel). I never got more than a difference of 1 in the averages. So I'm not sure how he got 4 in his example. It was either an absurdly unlikely result that he got, or either he or I messed up our sims.

You misunderstand my comment. I don't disagree with what you said above. I meant if Dalavar and Hound want to argue about something that is that obvious I would rather stare at an entertaining GIF. I thought the original argument was about predicting the next swing not about the variance of the average damage of 2 large samples of data.

WHAT THE IS HAPPENING HERE

It was a memorable quote... Sorry that's my fault. I fix I fix http://uorforum.com/threads/alternate-thread-for-arguing-about-variance.8338/

Hold on hold on.. before we go a minute further is this an African scimitar or a European scimtar????

.

I just say stuff and then Kal Vas Gideon Jura and have you tell me the actual names of the things I'm trying to talk about.

Wasn't this thread about quotes or something?

[17:50:54] <Wise> but she say hes just a friend
[17:51:05] <Blaise> OH BABY YEWWWWWW
[17:51:13] <Blaise> YOU GOT IRCEEEEEEEHEEEE