Deck "Characteristic Curves" (in preliminary development)

[Ekki]

EDIT: So statistically, is it possible to determine a single "Best deck ever"? One deck that consistently beats every other one (Or at the very least, beats more decks types than any other)? Or do you think we'll find many relationships between decks turn into a complex web of rock-paper-scissors?

I'd say the latter is more probable. Mostly since the game and its cards are really balanced and every now and then there are buffs and nerfs that balance it furthermore.
What is possible, though, is to find sort of an "ultimate" FG/AI4/AI3 farmer, as since they're rigged and you always face the same decks (although they are many), so you can figure out the faster counter to most of them. Unfortunately, I believe that is too much maths for a single punny human to proccess, but I guess that with a good maths tool by your side it could be possible.

[darkrobe]

Id agree with Ekki on that. but id give it a different spin.

If we can fine tune how different cards effect things like the damage potential curve, etc. we could for example, develop curves for all the false gods. and then, instead of designing a false god deck and then fighting the false god individually to see how effective the deck is against each one. We could just spend a few games getting an idea for that individual deck's curves and then predict which false gods (if any) it would be effective against and have an idea of how effective. without ever playing a single false god with that deck.

You could use that strategy to run through possible deck ideas much faster, and much more efficiently than we do currently. So maybe eventually we could figure out a "best deck" but my end goal here is the tool, not necessarily what people use that tool for later on.

I mean, (talking big picture, not expecting this to happen anytime soon) imagine a deck curve generator, that you could feed your deck code into, and it would tell you which false gods (or common deck types) your deck deck wins against as well as some statistical measure of how often or by how much it would win.

That could drastically change how decks are developed and make the evolution of the metagame occur much much faster.

[Ekki]

If we can fine tune how different cards effect things like the damage potential curve, etc. we could for example, develop curves for all the false gods. and then, instead of designing a false god deck and then fighting the false god individually to see how effective the deck is against each one. We could just spend a few games getting an idea for that individual deck's curves and then predict which false gods (if any) it would be effective against and have an idea of how effective. without ever playing a single false god with that deck.

I have this "apocaliptic" picture when, in the future, EtG players don't play the actual game, but instead they look to a set of curves to see what deck is their best choice, thus not playing at all because their opponents would know that they're gonna lose... But I guess you can go on with this project without feeling guilty for that

You could use that strategy to run through possible deck ideas much faster, and much more efficiently than we do currently. So maybe eventually we could figure out a "best deck" but my end goal here is the tool, not necessarily what people use that tool for later on.

I mean, (talking big picture, not expecting this to happen anytime soon) imagine a deck curve generator, that you could feed your deck code into, and it would tell you which false gods (or common deck types) your deck deck wins against as well as some statistical measure of how often or by how much it would win.

That could drastically change how decks are developed and make the evolution of the metagame occur much much faster.

^^This, it enlightens the future, and now I can see peace, equal games, bunnies and being last place in war.
Back to the real world, that'd be a badass tool, I'd like to see it!

[OldTrees]

If we can fine tune how different cards effect things like the damage potential curve, etc. we could for example, develop curves for all the false gods. and then, instead of designing a false god deck and then fighting the false god individually to see how effective the deck is against each one. We could just spend a few games getting an idea for that individual deck's curves and then predict which false gods (if any) it would be effective against and have an idea of how effective. without ever playing a single false god with that deck.

I have this "apocaliptic" picture when, in the future, EtG players don't play the actual game, but instead they look to a set of curves to see what deck is their best choice, thus not playing at all because their opponents would know that they're gonna lose... But I guess you can go on with this project without feeling guilty for that

Let me banish that fear. EtG will always be a very complex multidimensional web of RPS except R can still lose to S. These curves will not identify a best deck. Rather they will identify the ratio of decks to expect in a competition and help you decide what deck to pick. However if everyone when through the same throught process then they would all have the same deck right? That obviously changes what deck you (and thus they) would pick ...
Welcome to tier 1 of the metagame. You will discover that a metagame is a fairly resilient beast.

[Ekki]

If we can fine tune how different cards effect things like the damage potential curve, etc. we could for example, develop curves for all the false gods. and then, instead of designing a false god deck and then fighting the false god individually to see how effective the deck is against each one. We could just spend a few games getting an idea for that individual deck's curves and then predict which false gods (if any) it would be effective against and have an idea of how effective. without ever playing a single false god with that deck.

I have this "apocaliptic" picture when, in the future, EtG players don't play the actual game, but instead they look to a set of curves to see what deck is their best choice, thus not playing at all because their opponents would know that they're gonna lose... But I guess you can go on with this project without feeling guilty for that

Let me banish that fear. EtG will always be a very complex multidimensional web of RPS except R can still lose to S. These curves will not identify a best deck. Rather they will identify the ratio of decks to expect in a competition and help you decide what deck to pick. However if everyone when through the same throught process then they would all have the same deck right? That obviously changes what deck you (and thus they) would pick ...
Welcome to tier 1 of the metagame. You will discover that a metagame is a fairly resilient beast.

Ya, I know, I was just deliring because the tool sounded like awesomesauce
But it's true that this would certainly speed up the proccess (given that it works as expected), so it's even better...

[darkrobe]

right now I see alot of potential for straightforward strategies. like rushes. When you add in strategies that start affecting the other persons damage potential it gets much more difficult to calculate much more quickly. I already saw that something as simple as deciding when to play a heal in USEM is something that can differentiate a player winning 80% of matches compared to the computer winning 40%. The AI is not good at strategizing.

on a related side note. any thoughts on how to take into account quanta denial's effect on the opponent? there are various means of quanta denial and Ive been toying around with different ideas, but none seems too promising. I had a little trial with it in the pestal deck but the percents i was developing dont seem very predictive. I may have to try calculating quanta production curves and maybe relate that to damage or some other method.

thinking about calculating the effect of discord over the course of a match makes my head hurt.

[Elbirn]

... *raises hand* What is "meta-game"?

[OldTrees]

... *raises hand* What is "meta-game"?

It can mean 2 things. Either it is thinking about a specific type of question or it the collection of moves (decks, ...) people make based on those questions.

Ex:
Tier 1
You were sitting picking out a deck for pvp one day when you realize that your opponent also thinks about what deck to use. How can you take advantage of this knowledge?
Tier 2
You were thinking of how to exploit your opponent's rationality to gain an advantage in pvp when you realize that your opponent also thinks about how to exploit your rationality. How can you exploit your opponent's attempts to exploit your rationality?
...

People play pvp at various levels of metagaming. The frequency and power of a deck in a pvp format is also called the metagame.

[Elbirn]

So to simplify, it's predicting moves in rock-paper-scissors and acting accordingly?
'Tis an interesting psychology to gaming and something I think about a lot, didn't know there was a term for it.

[OldTrees]

So to simplify, it's predicting moves in rock-paper-scissors and acting accordingly?
'Tis an interesting psychology to gaming and something I think about a lot, didn't know there was a term for it.

Everyone's predictions and everyone's actions.
Each tier is playing a game using the previous tier as a board.

[Sevs]

Speed Poison

Code: [Select]

710 710 710 710 710 710 710 710 710 710 718 718 718 718 718 718 71a 71a 7gl 7gl 7gl 7gl 7gl 7gl 7gn 7gn 7gn 7gn 7gn 7gn
Napalm's Fractix

Code: [Select]

7dm 7dm 7dp 7dp 7dp 7dp 7dp 7ds 7ds 7ds 7ds 7ds 7ds 7f2 7f2 7f2 7f2 7f2 7f2 7f2 808 808 808 809 809 80i 80i 80i 80i 80i
Fire Stall

by Sevs

Hover over cards for details, click for permalink

Deck import code : [Select]

6rn 6rn 6rn 6rn 6rn 6rn 7dg 7dg 7dg 7dg 7dg 7dg 7dg 7dk 7dk 7dk 7dk 7dk 7dk 7dl 7dl 7dm 7dm 7dm 7dm 7dn 7dn 7dn 7do 7do 7dr 7dr 7dr 7f2 7f2 7f2 7f2 7f2 7f2 7f2 8po

Some pretty general decks

[mwaetht]

Is this still on?