Litotes wrote:I can tell you Thingol that 18 vs 20 dice is very similar to 6 vs 6. Only very marginally worse for the attacker.
Very similar, yes.
I'm not so sure about the second bit. I believe it is very marginally better for the attacker than 6v6 (assuming 3 dice attack only).
Litotes wrote:I can tell you Thingol that 18 vs 20 dice is very similar to 6 vs 6. Only very marginally worse for the attacker.
Right, but much closer to 6v6 than 5v6, correct? Atleast that's what I was shooting for. I was considering d17 vs d20, but stayed with the d18.
Thingol wrote:Litotes wrote:I can tell you Thingol that 18 vs 20 dice is very similar to 6 vs 6. Only very marginally worse for the attacker.
Right, but much closer to 6v6 than 5v6, correct? Atleast that's what I was shooting for. I was considering d17 vs d20, but stayed with the d18.
Definitely closer than 5v6.
If you want attacker to be handicapped 18v20 is virtually pointless. If you don't might as well stay with 6v6.
Pratik wrote:Litotes wrote:I can tell you Thingol that 18 vs 20 dice is very similar to 6 vs 6. Only very marginally worse for the attacker.
Very similar, yes.
I'm not so sure about the second bit. I believe it is very marginally better for the attacker than 6v6 (assuming 3 dice attack only).
You might be right. I ran ten big simulations and attacker came out about 0,02% worse with 18v20 than 6v6. Not big enough margin to declare with confidence.
Litotes wrote:Pratik wrote:Litotes wrote:I can tell you Thingol that 18 vs 20 dice is very similar to 6 vs 6. Only very marginally worse for the attacker.
Very similar, yes.
I'm not so sure about the second bit. I believe it is very marginally better for the attacker than 6v6 (assuming 3 dice attack only).
You might be right. I ran ten big simulations and attacker came out about 0,02% worse with 18v20 than 6v6. Not big enough margin to declare with confidence.
Can you try with d17 v d20?
if i did it correctly, it's roughly 60%
Thingol wrote:Litotes wrote:Pratik wrote:Litotes wrote:I can tell you Thingol that 18 vs 20 dice is very similar to 6 vs 6. Only very marginally worse for the attacker.
Very similar, yes.
I'm not so sure about the second bit. I believe it is very marginally better for the attacker than 6v6 (assuming 3 dice attack only).
You might be right. I ran ten big simulations and attacker came out about 0,02% worse with 18v20 than 6v6. Not big enough margin to declare with confidence.
Can you try with d17 v d20?
d17vd20 is as close to neutral as I've seen. So much better for attacker than d5vd6 (or d8vd9) but much worse than d6vd6.
Thanks guys. I think I will go to the d17 vs d20 now, after all that analysis.
This conversation has got me more interested in the topic. Need to study it a bit more.
But I need a question answered first: What is the maximum number of sides a dice can have on WarGear?
d20 is max per turn-based games.
d100 for simulgear games.
Litotes wrote:I can tell you Thingol that 18 vs 20 dice is very similar to 6 vs 6. Only very marginally worse for the attacker.
Interesting. The splits are close to identical too. For all practical playing purposes, they are equal as far as I'm concerned.
Finally got some time to address this "fairness of dice" topic. I'll share some preliminary results.
Short Version
Most Fair (Up to d100) : d17 v d23
Most Fair (Up to d20) : d10 v d13
d17 v d23 has a perfect 0.5 odds! Whereas d10 v d13 has a very marginal defender advantage: at 0.4991.
d17 v d20, which was being spoken about before, is far from fair. It is at a whopping 0.6 . i.e. Decent attacker advantage.
Long Version
I wrote my own version of the prepostnik odds calculator ( calculating odds by running a whole bunch of simulations). Ran a brute over all combinations, from d1 to d100 for attacker, and d1 to d100 to defender. Results above are based on 10000 simulations each, which is why the results are preliminary.
The fairest odds are determined by the probability of a d3 v d2 attack succeeding ( So, if you were to use the prepostnik calculator, you would have to give the input: [ NbSimulations, 3, dAttack, 0, 2, dDefend ] ). The odds of success for this can be seen here:
Up to d100: https://uploadir.com/u/w0qbdhrh
Up to d20: https://uploadir.com/u/7pn5sj01
If you look closely, you will see that the odds don't increase as you would think they would. There is a lack of smoothness in the surface, which is due to the "small" number of simulations. As you would expect, this becomes more evident in the larger dice sides, and thus, in the d100 figure. The smoothness should increase as the number of simulations are increased.
Soon enough, when I have available resources, I will run the same thing with a whole bunch of simulations more (aiming for 10^7 - 10 million- for each combination - which gives a total of 10^11 simulations).
Also not sure what else to share from the data set produced. If anybody has an idea, I can share more information about this.
d17 v d23 most fair? Who would have thought that? I guess we can throw this one http://prestopnik.com/wargear/index.html away, then. It gives from 1,9% to 3,8% chance of success for a 100 vs 100 attack d17 vs d23.
What?
Maybe I was not very clear in my last post. I am comparing a 3v2 armies attack with dX v dY, till one person wins. Not 100v100. The ranges for X and Y (number of sides of the dice) was from 1:100, since Thingol mentioned that's the maximum possible dice sides on WG.
As the number of units attacking and defending changes, obviously the fairest dice for that attack combination will be different. I just chose 3v2 since that IMO resembles a base attack (although the comparison is till one person reaches 0, which need not happen in the base attack)
OK, I thought Thingol wanted dice neutrality. As far as I know, d17 vs d20 will, in the long run, lead to attacker losing about as many units as defender, and d17 vs d23 will not.
Btw, since his Monopoly board is not SimulGear, d20 is max.
Yep, you're correct Litotes, that's what I was shooting for. Thanks all for the analysis.
btw, here is the beta version of the map, for anyone interested in opening a game. I've got several games open myself and I'm maxed on how much I can keep track of at once.
Thingol wrote:That is an interesting calculater Pratik. Thanks for the link. I was thinking along the line of chance of success of winning one dice roll, assuming the attacker has 3 dice to attack with and the defender has 2 dice to defend with. Success being that the defender loses 2 units.
That's on the right side of prestopnik.com/wargear. i.e. use the calculator not the simulator.
d18vd20:
Offensive Advantage: 118%
Avg. Needed to kill 10 defenders: 8.51
Avg. attack loss/roll: 0.92
Avg. defense loss/roll: 1.08
Avg. # attackers/defender: 0.85
Expected splits
Defense Lost 2 | 0.37 |
Attack Lost 2 | 0.29 |
A/D Split | 0.34 |
We were going d17 vs d20 Ozy.
Litotes wrote:OK, I thought Thingol wanted dice neutrality. As far as I know, d17 vs d20 will, in the long run, lead to attacker losing about as many units as defender, and d17 vs d23 will not.
Btw, since his Monopoly board is not SimulGear, d20 is max.
You are certainly right that my approach was too limited in it's scope to answer that question. Problem for me is that I'm not really sure how to measure dice neutrality in the long run.
Thingol wrote:That is an interesting calculater Pratik. Thanks for the link. I was thinking along the line of chance of success of winning one dice roll, assuming the attacker has 3 dice to attack with and the defender has 2 dice to defend with. Success being that the defender loses 2 units.
Thingol wrote:Yep, you're correct Litotes, that's what I was shooting for. Thanks all for the analysis.
btw, here is the beta version of the map, for anyone interested in opening a game. I've got several games open myself and I'm maxed on how much I can keep track of at once.
These two are two different questions, and have different answers in terms of most "fair" dice.
I was trying to answer the question in the first post quoted above (see the stuff I bolded). The way I see it, the dice combinations in my earlier post answers this question: Attacker has 3 units to attack with, defender has 2 units. Successful attack is defender losing all units. Failed attack is attacker losing all attack-able units (i.e. keep attacking till one of the two players is out of units). And for this problem (and this problem only), I wanted to see which dice combination led to the closest to 50% chance of success.
However, if you are aiming for dice neutrality in the sense that luck 0.0 implies attacker and defender lose the same number of units, my above simulations are absolutely useless.