#Another dice observation

49 messages · Page 1 of 1 (latest)

cobalt echo
#

For background, I'm a statistician. When I observer something, I TAKE DATA. I calculate probabilities.

In this case, I observed what I believed to be a tendency for a number to be rolled within 3 rolls of the robber being placed on it.

For example, an 8 has roughly a 14% chance of being rolled on a single turn. That translates to about a 36% chance of being rolled at least once in 3 rolls.

What I found, metaphorically, was an 55% chance of an 8 being rolled within 3 rolls if the robber is placed on an 8 and remains there for 3 turns.

I ran this analysis as a binomial with a set number of rolls...3. I also ran it as geometric. Similar results.

I'm not going to speculate on programming or motive or anything else. These are just the observations and calculations from a statistics professor.

long axle
#

How many games did you play to get these numbers?

cobalt echo
#

I don't recall the number of games.....it wasn't relevant to the calculation. It was 583 rolls with the robber in play.

muted girder
#

583 rolls? That would mean you have a data set of 194,33 since a set is 3 rolls. And you, as a "statistics professor", wouldn't think your result is biased?
As a "statistics professor", you wouldn't clearly know which analysis to run for this extremely easy statistical problem?

Other than you, I do speculate: I speculate, you are not a statistics professor

cobalt echo
#

I think you're confusing bias with error, perhaps because you think the data set is small? What role do you think the geometric plays here?

Enlighten us....."which analysis" would YOU run? Why NOT the other? (Both are legitimate...but please....elaborate.)

#

btw....not all binomial sets had 3. Perhaps you can figure out why.

dull jacinth
#

I’ve noticed this without being a statistician so very possible there’s some correlation

woeful nimbus
#

Would you be interested in sending the raw data?

muted girder
modest grove
worldly urchin
cobalt echo
#

I spent literally ONE SENTENCE decribing myself.....lol. The analysis takes into account that an 8 is more likely to be rolled. It's conditional probability. It's a comparison of expected value to realized results. I misjudged the crowd here. Be well.

woeful nimbus
#

What's your p value on your test?

#

What test are you using? A 1-sample proportion test?

#

How are you handling cases where another 7 was rolled within 3 rolls? Did you decide on 3 before or after you started recording data? Were you careful to record every roll or could there be bias if you remember to write it down whenever it happens (I would consider writing down every roll to avoid this)?

long axle
#

The rolls are determined by a seed at the beginning of the round. They cannot correlate with the robber, as they are predetermined.

muted girder
modest grove
muted girder
cobalt echo
# woeful nimbus How are you handling cases where another 7 was rolled within 3 rolls? Did you de...

You sound like a good AP Stats student :). The p value was 0.023. I used the binomial then ran it again as geometric. No need for normal approximations when the binomial/geometric are precise.

Great question about a 7 being rolled in less than 3. For those cases in the binomial, the number of trials was 2 or 1 and resulted in "failure" or "failure-failure".

Again....this is NOT a commentary on programming or motive. This is just numbers. Probability of Type I error is never 0. Randomness is a funny thing. I wondered based on what I was observing casually. What triggered me was seeing robber placed on a 3...and then a 3 being rolled immediately or soon after. Then with a 10 in the same game. I observed this over a few games. I was never triggered by the robber on 6 or 8. It's when the robber got placed elsewhere and I saw it happen that made me curious. So then I started recording data to see if what I thought I was observing was a biased illusion. I was surprised by the numbers too, given the "seeding" process of a predetermined set of random numbers.

jagged iron
#

Guys stop arguing. They fixed the dice with the update! It’s balanced now

IMG_9881.png
finite axle
#

Alright, let's roll it back to true random - I told you people would still be complaining, when they keep losing

jagged iron
muted girder
modest grove
modest grove
modest grove
muted girder
wispy schooner
#

I hate to say this because I was against changing the dice algorithm, but after about 10 games, I have to say it seems better. There is still plenty of seeming "randomness" but the same number hasn't been rolling over and over, many times in a row. See how it goes the next few games.

muted girder
#

Most people don't like randomness especially in competitions. Thing is that we now don't play the original game, but Catan with a different ruleset. Which in turn means that most of the game depends on your starting positions and numbers on ressources even more. Comeback potential is even lower than it has been so far

cobalt echo
muted girder
cobalt echo
# muted girder Feel free to not post unneccessary observations. You have that freedom, if not a...

But then you'd have nothing to do apparently. This is your life. And yet you have been absolutely useless to this conversation....I suspect this translates to everything else over your head you try to involve yourself in. Your life is predictable and sad that this is the best you can do with your time. You really don't know anything. It's clear to anyone who does. You are useless, know nothing, and expose yourself with every comment. Hell of an existence. I suspect you'll respond...because you can't NOT. You have nothing worthy to do with the dwindling moments of your sad life.

muted girder
cobalt echo
#

See....predictable....you're typing.....you are completely predicatble and boring.

#

You see people having an adult conversation....and have no idea how to participate. Keep replying. I'm happy to keep you occupied so you don't make the world worse somewhere else

#

keep typing...keep showing you CAN'T walk away...

#

lol

#

so predictable

muted girder
#

You try well to "win" this argument by either making me stop typing (so I don't prove your point) or making me look predictable when I don't stop.

I could make the exact same prediction with you

cobalt echo
#

But you didn't. Which makes you unoriginal too

muted girder
#

It is original to fabricate an argument which in itself is unchallengable? This seems rather stupid to me

cobalt echo
#

And yet here you are............lol......out matched, outwitted, completely exposed. Thank god for an anonymous screenname, eh?

muted girder
#

What exactly did you expose? That I answer to your comment?

cobalt echo
#

I exposed nothing......you exposed your lack of knowledge about anything from your first comment. It's been ad hominem on me and the argument the whole time.....with no actual analysis or alternative method suggested. BECAUSE YOU KNOW NOTHING. It's obvious to everyone. If you think it isn't, you're as delusional as you are ignorant.

Another difference between us......when I log off of this....I don't think about you at all. But you...you obsess over this.....I'll be on your mind all day...because you have nothing else to do....you have no value to anyone. This is your world. It's all you have. Sad.

muted girder
#

You make a lot of assumptions about me (and a lot of dots).

There was never a need for an alternative method, since you have 583 rolls (so around 10 games or even less). That's simply not a big enough sample to have a somewhat significant result.

cobalt echo
#

My god.....tell me you know nothing about statistics and sample sizes without saying you know nothing about statistics and sample sizes.......

You could not be more obviously ignorant......and anyone whose taken a basic stats class can see it. Keep frontin' for whoever you think your audience is though. lol

muted girder
#

You and your dots are too funny. But let us start again with your extremely specific observation. What was the purpose of you showing it to the whole discord channel?

jagged iron