Mixed reviews...

@midyin said in Mixed reviews...:

Er..no, it's exactly the same - 50%.

Except that it's not....
If you set down and roll a D10 20 times & give yourself a point every time it's a 6 or better, you'll end up with more points that if you sat down with a D6 and gave yourself a point every time it was a 4 or better.

I'm sure there's someone that can explain it better, but it does happen that way most of the time.

It's because on a D6 there's only 3 sides that can result in a point for you. On a D10 there's 5. And for some strange reason that makes it easier despite there being 5 sides that won't get you a point.

I'll try and get in here before VoodooMike rips you a new one, but this is basic maths.

Take a d6, and you want a 4+.
1, 2 and 3 are fails (3 results out of 6 = 1 out of 2)
4, 5 and 6 pass (3 results out of 6 = 1 out of 2)

Take a D10 and you want a 6+
1-5 are fails (5 results out of 10 = 1 out of 2)
6-10 are passes (5 out of 10 = 1 out of 2)

They are exactly the same - 50% chance.

You could even take a coin (D2)
Tails fail (1 result out of 2)
Heads pass (1 result of out 2)

The only "strange reason" is your failure to grasp basic probability.

last edited by Darkson

@midyin said in Mixed reviews...:

Except that it's not....
If you set down and roll a D10 20 times & give yourself a point every time it's a 6 or better, you'll end up with more points that if you sat down with a D6 and gave yourself a point every time it was a 4 or better.

I'm sure there's someone that can explain it better, but it does happen that way most of the time.

It's because on a D6 there's only 3 sides that can result in a point for you. On a D10 there's 5. And for some strange reason that makes it easier despite there being 5 sides that won't get you a point.

Oh wow. That's so wrong it's not funny.
http://www.wolframalpha.com/input/?i=>%3D6+on+d10
http://www.wolframalpha.com/input/?i=>%3D4+on+d6

Note that the "probability of occurrence" in both cases is 0.5, or 50%.

I don't want Madden. I don't even like Madden. I want a better BB.

No, you want a different BB. Whether or not it would be better is clearly a matter of opinion, and yours appears to be different to mine.

@Darkson said in Mixed reviews...:

They are exactly the same - 50% chance.

If we take a certain actions with 50% probability for both dices, like "rolling 4+ successfully" and "rolling 6+ successfully", and start rolling D10 and D6, registering results as "Success" and "Failure", I still think there will be significant difference in how chaotic this endless line of "Sucess" and "Failure" results will be. Sorry, I don't know the correct term for this, but the more edges dice has, the more results of a certain roll will seem closer to the expected probability, on a short range of measurements. Thus reducing the feel that "everything is about dices" a bit.

last edited by Mori-Mori

No, if you're only rolling 1 dice (assuming it's unweighted of course) then 50% is 50% it doesn't matter if you want a 4+ on a D6, 6+ on a D10, 15+ on a D28 or 51+ on a D100 - it's still 50%.

http://anydice.com/program/cdc3
D2, D6, D10, D12, so you want at least a 2, 4+, 6+ and 7+ respectively - check the %.

last edited by Darkson

@Mori-Mori said in Mixed reviews...:

If we take a certain actions with 50% probability for both dices, like "rolling 4+ successfully" and "rolling 6+ successfully", and start rolling D10 and D6, registering results as "Success" and "Failure", I still think there will be significant difference in how chaotic this endless line of "Sucess" and "Failure" results will be. Sorry, I don't know the correct term for this, but the more edges dice has, the more results of a certain roll will seem closer to the expected probability, on a short range of measurements. Thus reducing the feel that "everything is about dices" a bit.

What? That makes no sense. Seems to me it's more about feelings than facts.

@Darkson said in Mixed reviews...:

No, if you're only rolling 1 dice (assuming it's unweighted of course) then 50% is 50% it doesn't matter if you want a 4+ on a D6, 6+ on a D10, 15+ on a D28 or 51+ on a D100 - it's still 50%.

But when you roll D6 and D10 dices 50 times each, trying to succeed a certain roll, there will be less extreme cases for D10 like "5 Successes/Failures in a row", than for D6, in general. I believe it's like that. It's about distributions of results of some (not too big) amount of rolls, not the probability of one single roll's result.

last edited by Mori-Mori

@Mori-Mori said in Mixed reviews...:

But when you roll D6 and D10 dices 50 times each, trying to succeed a certain roll, there will be less extreme cases for D10 like "5 Successes/Failures in a row", than for D6, in general. I believe it's like that. It's about distributions of results of some (not too big) amount of rolls, not the probability of one single roll's result.

No there won't. 100 50-50 chances is 100 50-50 chances regardless of whether you use a coin or an RNG with 2^19937 -1 (for example) possible results.

Just.... no.

@dode74 said in Mixed reviews...:

@Mori-Mori said in Mixed reviews...:

But when you roll D6 and D10 dices 50 times each, trying to succeed a certain roll, there will be less extreme cases for D10 like "5 Successes/Failures in a row", than for D6, in general. I believe it's like that. It's about distributions of results of some (not too big) amount of rolls, not the probability of one single roll's result.

No there won't. 100 50-50 chances is 100 50-50 chances regardless of whether you use a coin or an RNG with 2^19937 -1 (for example) possible results.

Okay.. Here is what I meant (very simplified example, of course, and I know that such short ranges show nothing; Success is rolling 4+ D6 or 6+D10):

1. Success, Success, Success, Success, Success, Failure, Failure, Failure,Failure, Failure - probability of Success on this short range is 5/10=50%

2. Success, Failure, Success, Failure, Success, Success, Failure, Failure, Success, Failure - probability of Success is still 5/10, yet distribution is more "random"/"even"/"closer to expected 50% of a chance"/whatever

Won't D10 gravitate to the latter case more than D6?

last edited by Mori-Mori

No, because it has just as many chances to fail as it does to pass. Each roll is completely independent of the last - just because the last roll was a "fail" it doesn't make a "pass" any more likely. Each roll will be a 50/50 chance of a fail/pass, and the number so faces a dice has has no bearing.

Now if you were arguing for a 4+ on both a D6 and a D10, then yes, there would be a difference as a 4+ on a D10 is more likely than a 4+ on a D6 (7 faces of 10 against 3 on 6, or 70% to 50%), but you're not, you're asking for a 50/50 roll. A 50/50 roll is 50/50 regardless of how many sides it has (even numbers only of course, as you can't get 50/50 on an odd numbered dice).

I'm convinced! The system needs to be changed to d100 so 50% chance rolls almost never fail!

Obvious trolls are obvious...

@Mori-Mori said in Mixed reviews...:

@dode74 said in Mixed reviews...:

@Mori-Mori said in Mixed reviews...:

But when you roll D6 and D10 dices 50 times each, trying to succeed a certain roll, there will be less extreme cases for D10 like "5 Successes/Failures in a row", than for D6, in general. I believe it's like that. It's about distributions of results of some (not too big) amount of rolls, not the probability of one single roll's result.

No there won't. 100 50-50 chances is 100 50-50 chances regardless of whether you use a coin or an RNG with 2^19937 -1 (for example) possible results.

Okay.. Here is what I meant (very simplified example, of course, and I know that such short ranges show nothing; Success is rolling 4+ D6 or 6+D10):

1. Success, Success, Success, Success, Success, Failure, Failure, Failure,Failure, Failure - probability of Success on this short range is 5/10=50%

2. Success, Failure, Success, Failure, Success, Success, Failure, Failure, Success, Failure - probability of Success is still 5/10, yet distribution is more "random"/"even"/"closer to expected 50% of a chance"/whatever

Won't D10 gravitate to the latter case more than D6?

Oooooh, I think that I finally understand what you're thinking...

It's not the range of numbers in a 50/50 roll that get you a more "normal"/"expected" result after a series of rolls, it's the number of times that you roll the dice...

If you roll a d6 6 times (several different times), you'll probably experience more variance from the expected 50/50 result than you would if you rolled a d10 10 times... The more rolls that you make should (over time, but obviously there are no forces at play that make this happen) get you a more predictable average result...

@SHunterUSA said in Mixed reviews...:

If you roll a d6 6 times (several different times), you'll probably experience more variance from the expected 50/50 result than you would if you rolled a d10 10 times... The more rolls that you make should (over time, but obviously there are no forces at play that make this happen) get you a more predictable average result...

Yes, the word "variance" is what I think I've seen in some discussions of the matter which I cursory checked in the past. And yes, that's what I had in mind: whether or not D10 can produce more... smooth experience than D6 on a short range, say, in a single match. I still struggle to wrap my head around it, so please ignore my blabbering before

last edited by Mori-Mori

Once again, though, if the percentages are the same, that's not going to change from d6 to d10. Making more rolls of whatever dice should (in theory) get you closer to that 50:50 result...

Now, on a d10, you'll get less 1s (since 1s are an automatic failure), but that's still not going to change the 50:50 probability... would just mean less auto-failure

last edited by SHunterUSA

@Mori-Mori said in Mixed reviews...:

Okay.. Here is what I meant (very simplified example, of course, and I know that such short ranges show nothing; Success is rolling 4+ D6 or 6+D10):

Success, Success, Success, Success, Success, Failure, Failure, Failure,Failure, Failure - probability of Success on this short range is 5/10=50%

Success, Failure, Success, Failure, Success, Success, Failure, Failure, Success, Failure - probability of Success is still 5/10, yet distribution is more "random"/"even"/"closer to expected 50% of a chance"/whatever

Won't D10 gravitate to the latter case more than D6?

Heh, no. They are uniform distributions.

Here's a fun game called the Monty Hall problem:

You're on a game show and in the final round you are shown 3 doors... behind one is a car. The host asks you to select a door, which you do. Then, he opens one of the doors you did NOT choose and shows that it does not have a car behind it.... finally, he gives you the option to stay with the door you chose, or switch to the other closed door....

Should you switch doors, stick with the door you chose, or does it not matter? What gives you the best chance of winning the car?

@VoodooMike said in Mixed reviews...:

You're on a game show and in the final round you are shown 3 doors... behind one is a car. The host asks you to select a door, which you do. Then, he opens one of the doors you did NOT choose and shows that it does not have a car behind it.... finally, he gives you the option to stay with the door you chose, or switch to the other closed door....
Should you switch doors, stick with the door you chose, or does it not matter? What gives you the best chance of winning the car?

I just happen to know the answer from before, which is a bit counter-intuitive, yes, though can't say I remember exactly why it is like that. I'll better do some googling, probably, I most likely misinterpreted the point of some article/discussion I saw before.

@Mori-Mori said in Mixed reviews...:

Yes, the word "variance" is what I think I've seen in some discussions of the matter which I cursory checked in the past. And yes, that's what I had in mind: whether or not D10 can produce more... smooth experience than D6 on a short range, say, in a single match. I still struggle to wrap my head around it, so please ignore my blabbering before

What you described is not variance, it is something... nonsensical if I am honest. In the example given you roll both dice the same number of times with the same odds (4+ on d6 and 6+ on d10), so there's no change to the expected variance from them.

What you do get with D10 instead of D6 is a decreased propability for an automatic success/failure (given you still only get those on the extreme end of the scale). You won't get a smoother propability by using just another die. What the problem seems to be though is psychological, not mathematical. Every distribution of the same amount of successes and failures on a given set of throws is equally propable: rolling all successes and then all failures has exactly the same propability as rolling them alternatingly. The problem here is that we intuitively tend to see one as "less random", which just is not true.

Yep. If you used 1 as a fail on d6 and 1-2 as a fail on d12 you'd have exactly the same results.

It's amazing how all these people can roll 10 or 20 times and end up with a perfect 50/50 wins/loses. It's almost like they're assuming rather than actually rolling the dice...

• 63
Posts
• 6298
Views

Looks like your connection to Focus Home Interactive - Official Forums was lost, please wait while we try to reconnect.