Coin Flip option

[Robinette]

Robinette, what is 'streakiness'? just so we get our definitions out of the way.

good idea... especially since 'streakiness' is not a real word...


streak (strk)
n.
4. Informal

a. A brief run or stretch, as of luck.​

b. An unbroken series, as of wins or losses.​

So as it would apply here, it would be any sequence of consecutive wins or losses while attacking.
But since nobody complains about win streaks, lets just say that it would be any sequence of consecutive LOSSES while attacking

 

[Robinette]

Yes. Here again is the simple math.

The probability of flipping 3 heads is (.5x.5x.5) = 12.5%
The probability of rolling 3 sixes is (.167x.167x.167)=0.47%

And probability is DIFFERENT than streakiness.

wow...

Probability of rolling 3 sixes is not the same probability of winning, which makes that comparison worthless.

I am not sure if you are misunderstanding what probability is, or if you are misunderstanding what streakiness is...
Perhaps the definition of streakiness above will help sort this out for you...

 

[Robinette]

y'all can talk about this until pigs fly, but I think that everybody is overlooking the point that moving from random dice to a random coin flip significantly alters the time-tested play of the game. We're playing a game that has very specific odds, and those who have played and enjoyed the game do so with some understanding of the odds and play accordingly. Changing the odds fundamentally changes the game.

If it ain't broke don't fix it.

The streak problem isn't that streaks happen so much as it is that when streaks occur somebody occasionally gets upset about it - either the site is broken or the game engine is making them lose on purpose or the game sucks. Personally I like the game as-is, and i take it on faith that the server isn't out to get me. Somebody who thinks otherwise is going to do so regardless of the type of system we use. We might be able to slightly reduce the likelihood and frequency of streaks occurring, but we'll never be able to change User X's reactions to a run of bad luck.

I agree with much of what you said, however...

Many of the game options we already have fundamentally change the game, so as a game option, what's the harm?
Also, since the fixed force option will fundamentally change the game even more, are you even more opposed to that?

oh.. and User X should not be allowed to Vote or procreate, lol

 

[Robinette]

game settings: coin flip or dice:neo: everyone happy

 

[Robinette]

it is possible for other people to be good at math robinette. keep in mind that there is always someone who is better at a given task.

Bring it on...



lol, just kidding... actually i could not agree more...
BUT WHERE ARE THEY HIDING??? lol

 

[Jewmowrestler]

I'm getting a guy who's really good at math to come and try and convince you. And if that doesn't work I'll see if I can get a friend of mine who's gotten perfect scores on the SAT and ACT to weigh in.

 

[Jewmowrestler]

Hang on though. What I just realized is that isn't 'streakiness' a double edged sword so to speak? It can hurt us (15v3 becomes 1v1) or it can benefit us (the defender of aforementioned attack). Plus, if we eliminated the risk of the game entirely, winning would simply become a strength contest which would change the entire gameplay. I realize that you are proposing it as an option and not a requirement, but I thought I might as well approach the argument with logic rather than mathematics.

 

[WidowMakers]

I am not sure if you are misunderstanding what probability is, or if you are misunderstanding what streakiness is...
Perhaps the definition of streakiness above will help sort this out for you...

I know perfectly well what both probability and streakiness are.

But to keep with the discussion I will post the defs so we can continue with a common base.
--------------------------------------
DEFINITIONS
probability (prb-bl-t) http://www.thefreedictionary.com/probability
3.

a. The likelihood that a given event will occur: little probability of rain tonight.
b. Statistics A number expressing the likelihood that a specific event will occur, expressed as the ratio of the number of actual occurrences to the number of possible occurrences.​

streak (strk) http://www.thefreedictionary.com/streak
n.
4. Informal

a. A brief run or stretch, as of luck.
b. An unbroken series, as of wins or losses.​

I would also like to add this to the informal understanding of a STREAK
No one will ever be able to predict:

- When a streak will happen.
- How long the streak will last.
- When the streak will end.​

So I assume we can agree on these correct?
---------------------------------------------------
MATH
I will post this again because it is good math.

The probability of flipping 3 heads is (.5x.5x.5) = 12.5%
The probability of rolling 3 sixes is (.167x.167x.167)=0.47%​

I agree that probability is not the same as winning.
But probability is also not the same as streaks.

The math above just easily showed that rolling 3 sizes is harder to do that flipping 3 heads.
It is harder to get a streak of 3 sixes than a streak of 3 heads.

I can do more math for you if you wish but it will not really matter until we understand how we will use coins.
which brings me to the next section.
-----------------------------------------------------------
How will we use coins instead of dice. How will the Mechanics work?
One other thing we have not discussed (at least I am not aware) is how these coins will be used in place of dice.
Right now we have a possible 3:2 dice situation in the game.
Will we just have a 1:1 coin? Or will we allow 3:2 coins?
Any ideas on how this will work?

Please give me an attacker/defender coin example so we can all review and calculate the probabilities compared to dice.

 

[Robinette]

Yes yes yes...
Jewmowrestler, you are correct...

good idea... especially since 'streakiness' is not a real word...


streak (strk)
n.
4. Informal

a. A brief run or stretch, as of luck.​

b. An unbroken series, as of wins or losses.​

So as it would apply here, it would be any sequence of consecutive wins or losses while attacking.
But since nobody complains about win streaks, lets just say that it would be any sequence of consecutive LOSSES while attacking

But the "problem" apparently is that there are some who think the dice sometimes behave unfairly, that they are rigged, or the results are just not possible with real dice... I believe that this was the motivation for creating the "fixed force" option which will cause games to play out very differently indeed... while the game play with the coin flip would not be much different from the current dice MOST of the time, but it would significantly reduce the already rare occurrences of extreme streaks.

Although, it has been argued in this thread that since those same people would still complain when exposed to tiny and mild streaks, that there is really no need for this option at all... sort of an all-or-nothing perspective.

 

[Badorties]

The purpose of fixed force is to eliminate all luck from attacking. A side benefit is that it would eliminate bitching about unfair dice.

I am having a hard time wrappign my head around streakiness as a mathmatical concept. I get what streaks are, but streakiness seems to imply "the likelihood of streaks" which just doesn't make any sense.

A streak in Risk is constantly losing or constantly winning. Usually we recall the losing part. Whether we use dice or coins, there will be streaks of constant losing. With dice, the attack wins about 60% of the time. With a coin, its 50%. I just don't see how a coin would produce more streaks. A 1 in a 10 result will happen about every 10 attempts. This is true of dice, coins, or anything else. To say something something is less streaky is to say a 1 in 10 result would happen greater than every 10 attempts.

I have a lot of respect for you intelligence robinette, but I think you and others in this thread are arguing, albeit successfully, the wrong point.

 

[Robinette]

Widow... we might be more in agreement that either of us realizes...

But there are a couple things that you are saying that i still think are not relevant, and yet, i imagine that we are actually on the same page with these as well, but that we aren't saying it the same way...


Let me start with our common understanding of a STREAK... you say:

I would also like to add this to the informal understanding of a STREAK
No one will ever be able to predict:

- When a streak will happen.
- How long the streak will last.
- When the streak will end.​

Of course one cannot predict streaks...
THAT is what the probabilities define for us...
Probabilities are simply a mathematical measurement of the statistical likely hood of an event occurring.

So are we essentially saying the same thing here? Or I am missing something?

It is harder to get a streak of 3 sixes than a streak of 3 heads.

Yes indeed... I agree... but i will ask again, how is this relevant?

How will we use coins instead of dice. How will the Mechanics work?

All of the calculations I have done were based on the assumption that the battle would be a simple 1:1 coin toss.
The dice calculations were based on our standard 3:2 dice.

Please give me an attacker/defender coin example so we can all review and calculate the probabilities compared to dice.

okay...
One coin flip = 1 army removed from the board.
Auto-Flip = a whole bunch of single coin flips where 1 army per flip is removed from the board.

But perhaps there would be less confusion if we called these 2 sided dice, and instead of numbers, they said V for Victory, and D for Defeat.

 

[WidowMakers]

SO are you saying that with the coins I would attack 5 vs 3 I would flip 1 coin and if it was heads I would win and if it was tails I would lose?

5:3 (H) >>>5:2
5:3 (T) >>>4:2

correct?

 

[Jewmowrestler]

i have an idea that i put forth on a different thread. what if instead of the single die vs single die method, we did a sum of dice vs a sum of dice. For example, an army of 6 vs an army of 3: the 6 rolls 1,3,2,1,4,6 (17 total) and the 3 rolls 6,6,6 (18 total), then the 3 would win. the attacker would be penalized however many armies were put up (in this case 6). this would still include the element of risk, but would lessen the so called "streakiness"

 

[Robinette]

I am having a hard time wrappign my head around streakiness as a mathmatical concept. I get what streaks are, but streakiness seems to imply "the likelihood of streaks" which just doesn't make any sense.

A streak in Risk is constantly losing or constantly winning. Usually we recall the losing part. Whether we use dice or coins, there will be streaks of constant losing. With dice, the attack wins about 60% of the time. With a coin, its 50%. I just don't see how a coin would produce more streaks. A 1 in a 10 result will happen about every 10 attempts. This is true of dice, coins, or anything else. To say something something is less streaky is to say a 1 in 10 result would happen greater than every 10 attempts.

I have a lot of respect for you intelligence robinette, but I think you and others in this thread are arguing, albeit successfully, the wrong point.

Well first I must sincerely say thank you for trying to understand this,,, but there is a difference here...

Can i go back to the poker example please... (actually, i am very surprised nobody expanded on that yet)
In the example I gave, there was a 1 in 5 chance of hitting the Flush, and a 1 in 4 chance of hitting a Pair...

If we set this condition up and did it randomly 10 times, there is a higher probability of hitting the PAIR in sequence, than with hitting the FLUSH in sequence...

In other words, the Flush would be less "streaky" than the Pairs.

Does that help?

 

[Robinette]

SO are you saying that with the coins I would attack 5 vs 3 I would flip 1 coin and if it was heads I would win and if it was tails I would lose?

5:3 (H) >>>5:2
5:3 (T) >>>4:2

correct?

Correct

 

[WidowMakers]

Well first I must sincerely say thank you for trying to understand this,,, but there is a difference here...

Can i go back to the poker example please... (actually, i am very surprised nobody expanded on that yet)
In the example I gave, there was a 1 in 5 chance of hitting the Flush, and a 1 in 4 chance of hitting a Pair...

If we set this condition up and did it randomly 10 times, there is a higher probability of hitting the PAIR in sequence, than with hitting the FLUSH in sequence...

In other words, the Flush would be less "streaky" than the Pairs.

Does that help?

Probability DOES NOT mean streakiness.

Streakiness is what HAS already happened and what we have seen
Probability is what COULD happen.

There is a 20% chance of a flush and 25% of a pair.
But if we did that 10 times and each time I hit a flush I would be on a 10 flush streak but that would be WAY above the probability of such a streak.
I should in all probability only get a flush 2 times of those 10.

So again to my point, STREAKS can only be seen after they happen and regardless of the probabilty, a streaks will happen.

 

[WidowMakers]

i have an idea that i put forth on a different thread. what if instead of the single die vs single die method, we did a sum of dice vs a sum of dice. For example, an army of 6 vs an army of 3: the 6 rolls 1,3,2,1,4,6 (17 total) and the 3 rolls 6,6,6 (18 total), then the 3 would win. the attacker would be penalized however many armies were put up (in this case 6). this would still include the element of risk, but would lessen the so called "streakiness"

This is an interesting idea. I will think a bit more on how this could work.

 

[Robinette]

Probability DOES NOT mean streakiness.

Streakiness is what HAS already happened and what we have seen
Probability is what COULD happen.

There is a 20% chance of a flush and 25% of a pair.
But if we did that 10 times and each time I hit a flush I would be on a 10 flush streak but that would be WAY above the probability of such a streak.
I should in all probability only get a flush 2 times of those 10.

So again to my point, STREAKS can only be seen after they happen and regardless of the probabilty, a streaks will happen.

AH HA! Just as I thought... Semantics...

in the example, you are correct to say that the most likely occurrence would be getting 2 flushes...
but there IS a 1% chance (0.8%) that i will hit 3 flushes in a row... THAT is the predictable likeliness of that streak occurring.
You won't see them until they happen, but the probability is still there.

But aren't we really meaning the same thing here?

 

[zspBANNED]

This is an interesting idea. I will think a bit more on how this could work.

I fucking came up with this idea months ago and no one gave it it's proper consideration! lol

 

[giuppi]

i have an idea that i put forth on a different thread. what if instead of the single die vs single die method, we did a sum of dice vs a sum of dice. For example, an army of 6 vs an army of 3: the 6 rolls 1,3,2,1,4,6 (17 total) and the 3 rolls 6,6,6 (18 total), then the 3 would win. the attacker would be penalized however many armies were put up (in this case 6). this would still include the element of risk, but would lessen the so called "streakiness"

Among other things, this system would make it impossible to lose a 7vs1, or a 13vs2. That contribution to streakiness would be therefore eliminated. And streakiness in general would be greatly reduced (must do some math to see of how much).
One thought though: would this system basically make every attack a blitz? With the current system I can stop my attack at any time, but with the "sum" system I couldn't. Even if there was that option, it would never make sense to attack with less troops than I have available.