Hate Speech: Theory Fighter University: Remedial Math
- Thread starter Dr. Hates
- Start date
Signia
[13] Hero
Hates is my favorite poster ever
I've already thought about this Heaton, you simply read me like you read your own thoughts
The math is almost always on my side as a Sophie player, provided I'm playing on point. That's why I consider her top tier.
I've already thought about this Heaton, you simply read me like you read your own thoughts
The math is almost always on my side as a Sophie player, provided I'm playing on point. That's why I consider her top tier.
Signia
[13] Hero
I think the toughest part about this applying this concept is discerning your opponent's perception of risk/reward in order to predict what they will do or mix up with so you can take advantage of it. Players don't vary options in equal frequencies or play the optimal frequencies (nash equilibrium), so changing the frequencies of your mixup options to adapt to theirs can give you a mathematical advantage.
I feel like I know intuitively how to take advantage of a player that isn't choosing a good mix of options -- just keep trying to randomize your mixup options with nash equilibrium frequencies, but sometimes you'll have bad luck or you'll get read from mixup attempt to mixup attempt, and it won't be enough to win every time.
In order to do better, you have to discern their option choices and their frequencies (something that few players can adapt) and then divise a harder counter to it, so your risk of losing to random chance is smaller (nothing you can do about getting read over and over except try to think a step further or behind). Divising that hard counter is very difficult in the heat of a match, you really have to know your stuff.
Or you could just adapt intuitively, treating each mixup like all options have equal r/r, making it no different than RPS, but if you want to specifically counter your opponent's strategy, you'll have to do some more thinking.
I'm actually still a game theory newb, so if I'm using "nash equilibrium" incorrectly, please let me know.
Another thing about mixups: they're necessarily inconsistent. True consistency comes from avoiding mixups entirely, I've seen. Though I doubt you can just wait around hoping for free damage from unforced errors using non-committal defensive options in SCV due to the guard crush system and the throw break chip damage.
I feel like I know intuitively how to take advantage of a player that isn't choosing a good mix of options -- just keep trying to randomize your mixup options with nash equilibrium frequencies, but sometimes you'll have bad luck or you'll get read from mixup attempt to mixup attempt, and it won't be enough to win every time.
In order to do better, you have to discern their option choices and their frequencies (something that few players can adapt) and then divise a harder counter to it, so your risk of losing to random chance is smaller (nothing you can do about getting read over and over except try to think a step further or behind). Divising that hard counter is very difficult in the heat of a match, you really have to know your stuff.
Or you could just adapt intuitively, treating each mixup like all options have equal r/r, making it no different than RPS, but if you want to specifically counter your opponent's strategy, you'll have to do some more thinking.
I'm actually still a game theory newb, so if I'm using "nash equilibrium" incorrectly, please let me know.
Another thing about mixups: they're necessarily inconsistent. True consistency comes from avoiding mixups entirely, I've seen. Though I doubt you can just wait around hoping for free damage from unforced errors using non-committal defensive options in SCV due to the guard crush system and the throw break chip damage.
MONEYMUFFINS
[13] Hero
I feel like sharing my game on this topic. I’ll break it in steps. The last part is probably most interesting in regards to theory.
My first goal is to structure my move list until I feel I can stand a chance against any patterned attack. For this part I need to feel that my character is all-around solid and can truly counter any type of attack viably (TJ/TS/TC). The second part is classifying players’ habits, in a general and specific way. General being if they’re more mixup/punish/interrupt oriented. Specific being what particular moves they do in certain situations, often subconsciously. The final part is coming up with an algorithm in my fighting style that best suits a particular opponent (i.e. applying actions to my classifications of the person). This is usually decided on the fly while trusting my intuition, based on my observations from round to round.
My first goal is to structure my move list until I feel I can stand a chance against any patterned attack. For this part I need to feel that my character is all-around solid and can truly counter any type of attack viably (TJ/TS/TC). The second part is classifying players’ habits, in a general and specific way. General being if they’re more mixup/punish/interrupt oriented. Specific being what particular moves they do in certain situations, often subconsciously. The final part is coming up with an algorithm in my fighting style that best suits a particular opponent (i.e. applying actions to my classifications of the person). This is usually decided on the fly while trusting my intuition, based on my observations from round to round.
