FoLermon: Ultimate Showdown - GRAND FINALES - 2/8

[Ash]

I need to actually run this on a larger sample size.
60,000 games means that this is only reflective of about 850 games per pairing, and I want about 10,000 games per pairing, so I need roughly 700,000 games to get as much data as I want.

[orangeandblack5]

how are you even simming this anyways

is it just a bunch of 1v1s

are actions randomly taken

how do

[Ash]

Simulation of games with fairly basic logic.

There’re some exceptions, but the fundamental logic is:

• If an FoLermon can be knocked out by both opposing FoLermon, they will Protect.
• If a team can take out an opposing FoLermon, they will attempt to do so, preferably with an FoLermon that outspeeds that FoLermon.
• If an FoLermon will die to an opposing FoLermon before attacking, it will attempt to switch to a member on the bench that takes the least amount of damage from that attack. Failing this, it will Protect; failing this, it will attack.
• If no kills are threatened, each team will seek to maximize damage.

[orangeandblack5]

also fun fact but by smogon rules every single folermon would be no lower than OU

[orangeandblack5]

oh so like I should outperform this by a lot

[Ash]

Tried to implement a minimax algorithm, but that was disastrous and broke everything.

[orangeandblack5]

idk dude if ur algorithm can’t do my double protect opener from our battle it might just suck

[Ash]

Double Protect isn’t necessarily a great strategy?
Like, sure, you lock your opponent into a move, but you also lose your opportunity for momentum.

[Ash]

There are tradeoffs, and given that we have no ways to heal, I chose to reward more aggressive play.

[orangeandblack5]

I’m acutely aware

if I thought it was always a great strategy I would have done it in my first game too but I didn’t

both openers accomplished the goals I had for them

i was just meming tho

[orangeandblack5]

actually though my first game was baller I had basically locked in my move ahead of time because I knew exactly what folermon would be picked, which ones would lead, and what they would do turn 1

v happy with how accurate i was on that one

[Ash]

There’re exceptions to this - Amelia will activate their ninja turns if that’s down and if they don’t see a kill, Tutuu will attempt to switch out if they’re at low health and know there’s an Orange with higher health on the opposing team’s bench, a Scaler gets priority on a kill, etc - but this is the basic logic.

[orangeandblack5]

from my action submission dm

@orangeandblack5

I am going to lead Wazza and Zone

my back line will be lol and Kiiruma

@orangeandblack5

I’m going to double into sulit with both smug attacks

@orangeandblack5

im anticipating a protect on arete and attacking into wazza

but i’d rather take the hit on wazza for 90 damage (putting me at 10 HP and guaranteeing I move first next turn) than protect wazza and end up giving sulit faster speed next turn

@orangeandblack5

I can threaten to reverse KO her and guarantee hits onto Arete’s slot in one fell swoop

@orangeandblack5

for the record I knew these would be the leads and I was already going to double into sulit here from the start lol

hence sending out both my smug mons first

@orangeandblack5

it practically forces arete to protect since they wouldn’t do supereffective damage anyways

@orangeandblack5

and in fact are resisted by zone

@orangeandblack5

while wazza is already weak to sulit anyways

[orangeandblack5]

and after

@orangeandblack5

wow thats crazy arete protected

@orangeandblack5

anyways im going to double into arete

[Ash]

Updated Version

[orangeandblack5]

can u give me the summary statistics again lol

[Ash]

This is basically what my channel with Tutuu looked like, by the way.
Amazed she put up with me doing this nonsense.

[Ash]

@Ash

[orangeandblack5]

tfw amelia and achro both have <50% into each other

I guess the remaining percent is god smiting both of them?

[Ash]

Failure rate of about 0.85%.
Defined as “you’ve gone six turns without anyone dealing damage”.