Void Grubs Analysis vs a Twitter Data ‘Analyst’

Hello.

Uh, and not picking up

now.

Now.

>> Oh,

now.

>> Hello.

>> Wait, does that work?

>> Yes. Yeah. Good.

>> Wait. No, no, no, no, no, no. Hold on. I

I was only able to pick you up on

my That's weird. I was only able to pick

you up on my speaker or my my

microphone. Not Yeah, that's weird. How

about now?

Now. Now.

This is so scuffed, man. I'm so sorry.

[ __ ] my life.

Uh,

wait. Properties default

L. Now

>> you

>> you sound perfect to me.

>> Perfect to me.

>> Okay, good. All right. All right. I got

it. I got it. Okay. Um,

>> hello. Okay.

>> Hello.

>> Hello. Okay. So

obviously I mean I I asked uh the uh 10

or 1010 on Discord right to ask for

Discord um because of replying to Twitch

chat. Okay.

>> So what do you feel like is the issue

>> with

the main issue is

okay you can test in practice tool and

that sets a perfectly good premise.

Okay, that gives you some data

>> to make it concrete or at least close to

the truth. [snorts] My thoughts are that

you need to look at the data. You can

either use the pro player data or you

can use the solo Q data. Obviously, the

solo Q data having the most information

available and that gives a practical

context in how void grubs actually

operate in the game. That's the premise.

>> Okay. But all right, so I'm I'm

responding right to your tweet where I

made the the the giant tweet about void

grubs, right? And then uh yes,

>> you replied uh why bother doing this?

All right, here I'll just

>> um you replied why bother doing this?

>> Cuz you've replied to a lot of my tweets

uh in general, right? So it's a

recognizable name and so it's not

difficult uh as well as your profile

picture. Um so you said why bother doing

this uh in head calculation rather than

looking at the data directly. Uh nothing

you say here is backed by real

experiences. Go through the data and

confirm these things for real. Okay. So

I mean using the same logic that you're

using, how do you know that it's in head

calculations?

>> It's it's an assumption given that we

don't see any kind of data backing it up

as in like what is your source data

backing up this information. So let's

say you gave those numbers and you

actually said okay I calculated this

based on end games from masters players

or these games from 2026 and pro data

that's perfectly fine

>> but just saying the numbers has very

little value in itself

>> why

>> because where's your evidence right if

someone just came up to you

>> the absence the absence of evidence is

not the the sorry the the Like

>> just just because something hasn't been

proven true doesn't mean that it's

wrong.

>> But also saying that the absence well

saying that that there's an absence of

evidence does not mean that there's not

evidence.

>> Yes. Correct.

>> Okay.

>> And you kind of have the onus to provide

that evidence if you're giving some out

giving out public information,

>> right? Because this is what people

expect when they read it.

What?

I don't I don't

>> No, no, no. Well, absolutely it's

annoying to do, which means that I don't

have to do it

>> because I mean, you can already concede

that the absence of evidence is not

>> It doesn't mean evidence doesn't exist.

It's facious to to claim otherwise.

>> No, I I I know cuz you can't really say

the absence of something means it

doesn't exist. That's fine. But when

it's not there and you don't provide the

evidence, it just makes it impossible to

believe what you're saying.

unless you actually go through the

evidence in the first place. So, it

comes with negativity to it. I'm not

saying you have to, but if you want a

more positive perspective to whatever

you say and more people to adopt what

you're saying, evidence is key. And if

you have someone that can do that,

that's fantastic. That helps

substantially.

>> But evidence isn't key because why do

you think I don't do anything the last

couple of years?

>> Why do you think more and more Twitter

analysts and stuff?

>> Absolutely. So, why does Virgo V2, why

does Molecule, why do all of the former

publicly uh forward- facing analysts and

people that used to do the things that

you're claiming, right? They would open

up videos, they'd post videos, etc. It's

not that they don't do it, right?

Sometimes even I I do it myself

sometimes, right? It's just it's it's

way less frequent than it used to be. Um

it's because it doesn't matter. Like the

the the freezing video, right? That that

didn't matter.

>> The um I understand that. I I I actually

don't agree with just showing video. I

don't think that's a substantial form of

evidence. It's again would be what n

equals 1 and if you do multiple VODs

maybe n equals 3. What I'm saying is

addressing kind of the large sample pro

games all pro games maybe from 2026

maybe from a specific patch. Same thing

for solo Q. It's a significant

difference from just using your pro VODs

or experiences from your games that you

watch. It's a massive difference.

But it has nothing to do with proplay

because that that's an appeal to

authority to to claim that pro play

would matter more than the objective

facts of reality.

>> It's difficult to the objective facts

are kind of like the average of what

happens within a game. That's that's my

understanding. Right. So what is the

average of what happens?

>> I'm saying I'm saying the the the

objective facts of how much damage a

void grub deals to the turret. How much

golden exp a void grub gives. the

objective fact of how much gold a minion

wave gives or how much exp a minion wave

gives.

>> That's fine, but then how many times do

people attack those turrets across the

entire game from 15 minutes plus given

any amount of time during the game?

That's quite difficult for you to

quantify in your head. In fact, I'd say

rather impossible,

>> right? So, so you don't need to quantify

it in your head. can make an equation

which is what I outlined in the tweet of

you know X multiplied by Y equals Z

which is then you insert the champion

variable into X and into Y and then you

get Z. You you don't need to go champion

by champion because ultimately what

we're doing is is we're we're taking the

champion as a variable, right? We're

inserting it as a a rough approximate

because some champions are going to do

roughly the same amount of damage to

turrets. And obviously there are going

to be outliers or there's going to be

extremes like Fiora or you know uh you

know Jinx maybe or Tristana uh Ziggs uh

these types of champions etc. There will

be outliers. Absolutely. Sure.

>> And then you can insert that champion

into that equation as a formula because

the way that I've always operated is

formulaic in in my approach of how to do

things. It's why you don't need to

depend on everything because most things

are evergreen. Um, so if you insert the

champions and then you calculate Y for

number of auto attacks, then you get an

output. And you don't need to go game by

game by game by game or ask every single

champion. You can get a rough estimate.

And I'm using estimate not as the

English word estimate, but as the

mathematical word estimate. So sort of

like when people hear the word theory,

they imagine they immediately think an

idea rather than the scientific uh

definition of theory or the way that you

know it can be applied. So so you can

get an estimate in that regard and then

the estimate allows you to then apply

that and then factor in more discernment

etc to then reach a more quantifiable

conclusion.

>> I appreciate that. Uh I think the the

the fact is I can't I can't say your

formulaic approach is completely

incorrect cuz I haven't got an example

via my approach to state otherwise.

>> Okay. So I just outlined for you

>> X Y right X Y equals Z. Okay.

>> What is erroneous in what I just said? X

being champion placement time Y auto

attacks done equals Z damage dealt to

turret.

the the formula would be incorrect and

it's missing variables. That's that's

what my response would be. How

>> how

>> so first of all, you don't know how many

times people are able to hit the tower,

right? You don't know how much space

they're being given by the enemy team,

>> right?

>> You don't know how long the game has

lasted.

>> There's a lot of variables in play that

we can't even quantify because obviously

right doesn't always provide that

information. So, Caitlyn,

>> most of the time, don't I'm sorry to

like cut you off. It's it's it's because

I I feel like I'm having to go step by

step because maybe something that I feel

is inferred is not being understood. If

X* Y, if we replace this as Caitlyn

times

eight auto attacks equals Z damage, how

is that lacking context?

>> How do you know how many auto attacks in

the first place? because you've

simplified it down to a formula that

allows you to then go through and check.

>> No, but I'm saying that you're not if

you're not aware of the number of auto

attacks and let's say she does 200 per

game, that massively changes your like

kind of like the outcome from that

formula. And not knowing that is

actually is very detrimental to actually

the accuracy of that formula.

>> No, no, because you're you're you're

taking 200 as a as a holistic number.

you're taking TDD a as as a more

relevant statistic or measurement than

than DPS or or or or damage dealt in X

instance. TDD is a is a fake stat. If

Riot adds 10 damage to Lucian's Q and it

doesn't alter his ability to clear a

minion wave or do anything, right? And

his win rate goes up by one, but he

casts 100 extra Q's per game and all of

them land and they go through armor

resistances. And because armor is

currently an evergreen thing that Riot

hasn't altered the stats of, then you're

able to conclude that Lucian on average

is going to get X more TDD, total damage

dealt, which means that his his endgame

scree obviously goes up. So for Caitlyn,

when you say 200 auto attacks against a

turret inside of a game, that doesn't

matter because what does matter is what

in a practical sense, if we're if we're

using pragmaticism, we're able to look

at the game and then ask with an

estimate. Again, using the mathematical

word to determine what is likely or what

is probable because again, we're trying

to be pragmatic. We're not trying to be

AI. We're not trying to be robots and I

don't expect them to be. So then you can

go through a wide sample size of of

games and then you can reach an

estimate.

>> Yes. But that that is what I'm

suggesting to take a wild sample size of

games

>> from the existing pool in solo Q or pro

games.

>> Yes.

>> So we can confirm this.

>> But but the the pro or solo Q would not

matter more than the objective reality

of of just the number of hits. I mean

yes. Okay. So I would say that pro play

pro play on average. Yes. It's actually

slower, right, than than solo Q. It so

generally when I do a lot of my tests, I

actually try to skew it for the opposite

side. Um I I try to skew it for against

my initial hypothesis and I try to see

if even that is reasonable.

>> Mhm.

>> What

>> So yes, go ahead.

>> Sorry, there's a bit of a delay. Um, so

all I'm suggesting is let's say your

formula and your approach for the

formula actually matches what happens

when we evaluate it from the view of the

massive sample size from pro or solo QA

games.

>> Then that validates what your assumption

is and you don't need to reference the

evidence anymore because we've already

proved that actually your formula

matches what the large sample says.

That's perfect and you're done.

>> No. No. That that accepts the current

status quo as being true and that's not

what I'm trying to do.

You're you're

>> I don't know what you mean by status

quo.

>> So so so you're you're saying that if

the data matches my formula that it

proves it true and I disagree

>> for correlated and therefore you don't

need the extra evidencing steps from the

going through the data

>> because it it accepts something as as

being true that might not necessarily be

true and that's not how you break the

ceiling.

But I'm very happy for you to go through

the evidence like we just mentioned. But

then I would I would just say you're

asserting that I haven't

>> because first you refer to them as in

calculations.

>> Yes, but that we've already established

that that is a fallacy.

>> The absence of the evidence being there

is not the absence of evidence entirely.

It doesn't mean the evidence doesn't

exist. It just means it's not in that

tweet.

Yes, but that it needs to be there for

it to not be received negatively. You're

reading it.

>> Why

>> don't don't you feel like you've been

received negatively a lot?

>> Everyone has received me negatively

because most people would be RECEIVED

THE THE GUY THAT DISCOVERED that we have

germs on our hand killed himself and was

mocked and scorned by his peers.

>> Do you know about that story?

>> Yep. I agree. I agree.

>> Okay. So what does it matter if I've

been received negatively?

>> That doesn't matter. What matters more

to me is

>> the occurrence of that.

>> Very likely. Everyone's that you cannot

have 100% uh favor. That doesn't make

that doesn't even exist.

>> No, I don't I don't assume I don't even

assume that my assertion is necessarily

true. It's just what I have been able to

verify.

>> It doesn't make me right.

>> Yes. But the the whole point is that the

lack [clears throat] of evidence to a

majority of the population comes across

as negative. That that is why it will

get received.

>> They can't discern the evidence anyway.

>> Yes. But the the lack of discernment

even if it's there, they don't need to

discern it to know that you've provided

the effort to provide, you know, that

that evidence. That's what's important.

>> They might not understand it, but it's

there. But it's there.

>> Do you know how many pro players I

interact with?

I'm going to guess 30.

>> 30.

>> Much more than that. What?

>> It's a guess. [laughter]

>> Okay. Sure. So, so much more than that.

So, what what seems better for aspiring

pros that I might not interact with or

or pros that maybe follow stuff that I

do uh but don't directly interact with

me as well as ones that I do interact

with. What would be better? manually

sending a Discord me sending off dozens

and dozens and dozens and dozens of

Discord messages um or just making a

general tweet for maybe even people that

I don't know could value it or entertain

it. What would be a better usage of my

time as well as what is a better way to

reach a wider audience?

I mean, if you're trying to interact

with just specifically the Okay, so

first of all, if you wanted to interact

specifically with those individuals and

only have a positive um outcome, you

would want to directly DM them. If

you're if you do want to if you care

more about the time investment, yes, you

just want to tweet and ignore everyone

that replies. But I don't feel like I

feel like you do want the positive

affirmation from the larger public

because having a 20% favor whenever you

tweet is annoying. At least that's my

assumption.

>> Of course I'm going to have a 20% favor.

People don't have discernment.

>> But but they do. Some of them.

>> They don't. They don't. No, no, no, no,

no, no, no, no, no. Let us not be Let us

not be misguided. The mass majority of

the League of Legends player base exists

in bronze, silver, gold, and platinum.

That is the overwhelming majority. Okay.

Do you think a gold, bronze, silver, or

platinum player can open up a VOD of

their own games? Determine what the

mistakes actually are, then skip forward

to a diamond game, then a master, then a

challenger, and also discern mistakes.

No, they can't.

>> They cannot discern, but they can

appreciate the evidence when it's there.

They might not understand it, but they

can appreciate.

>> But [laughter] appreciating it doesn't

But you said they can discern it.

Appreciating it and discerning it are

two separate things.

>> No, I didn't say discern. But if I did,

then I disagree with what I just said. I

don't think they can discern. But I do

think they appreciate the fact that you

went through the evidence and it's very

clear and the evidence is strong. That's

what's important.

>> Do you follow chess? Do you follow

chess?

>> I do. Yes.

>> Okay. So, I follow chess, too. Right. To

a certain extent. If I go on a Gotham

chess channel and I watch a video of him

discussing a chess game, etc., and he's

doing he's drawing all these lines on

the board, and he's explaining why some

things, you know, he might do things

that are actually really good, but maybe

Magnus or uh Hikaru or Fabiano or

someone like that would actually have a

better line, right?

>> But he could also lie in his video and I

wouldn't know. Why? Because I just I've

chosen that I like Gotham chess. So I

just choose to believe him as long as

he's presenting the data or the

evidence.

>> Yeah.

>> Is [laughter] that your logic?

>> I wouldn't draw the same conclusion.

It's the for the video format is

different, right? He's on a live he's

kind of presenting the game live

providing his own opinions on the lines.

>> No, no, no, no. Let's say let's say

let's say it's a video. Let's say it's a

video.

>> Like it's it's a game that has already

happened.

What prevent how how what majority of of

of his viewers without using Stockfish

can tell if he's lying or not?

>> Ah, I got you. I got you. Um, so very

low, but they I do believe they

appreciate the fact that he does use the

Stockfish lines to confirm whether or

not what his suggestions are

appropriate. He doesn't just use He

doesn't just use Stockfish,

>> right? But that that that isn't my

argument. The main argument is that they

can't discern anyway.

You're you're going back to

appreciation. You keep saying the word

appreciate. That's an emotional appeal.

>> Yes. Because a lot of people respond to

these kind of things emotionally, which

is why you do want to appeal to that

emotional

>> perspective. The way to do that though

is by the

>> I don't care about that. You seem to

care about that. And then because you

do, you want to impose that view on me.

But it has it has no basic.

I prefer I I don't want to impose

>> but I do want to see

>> you have a basis in the logic

>> have you ever tried this perspective

like like even once I think if you even

tried it once and confirmed that it had

no value that's completely fine but I

think if to to deny it completely you

should at least try it once use petu pu

has the data right you can ask him

>> why are you making the assertion that I

I evidently as I already stated earlier.

I have done posts that have a lot of

data.

>> Mhm.

>> Okay. Obviously, it doesn't do anything.

>> When I when I manually when I manually

there was a world guy in chat probably

knows this like various people. There

was a world where I manually counted the

amount of auto attacks Kai'Sa was doing

in team fights across several games. I

manually went and counted every single

Kai'Sa team fight and then I made a

document on it. Okay. And then I

calculated the average amount of damage

that she was doing and then I compared

that to AP Kaisa. I posted that. Did it

make people start playing AP Kaisa?

>> Sorry. Was that the average amount of

damage from AP Kaiser from also

professional games within that pack?

>> Calculating the amount of autos that an

attack speed Kai'Sa did.

>> Yep. And you were comparing that to AP

Kaiser from the same professional

>> tournament. Several games. So, I

manually went through several replays,

games, VODs, whatever, and manually

counted the amount of auto attacks as a

means to argue in favor of AP Kaisa.

>> Yes.

>> Okay. And then I confirmed that not only

would AP Kaisa have dealt more damage

in those fights, she also would have

provided different options, a different

different ranges of mode, and all these

other things. Did that make people play

AP Kaisa? No.

I I I think that's a massive amount of

effort to go through. So,

congratulations on that.

>> So, yes, it's called Brandini's law.

Yes. Brandodini's law. Do you know what

that is?

>> It's not It's not No, I don't have

>> Okay. It is the [ __ ] It's the

[ __ ] asymmetry principle that the

amount of effort and exertion that one

must expend in order to refute something

that is not true is significantly more.

Uh, so the only thing I'm going to want

to clarify is that even seven games is

not much. Like even a 100 games is not

much. Like what what I was suggesting is

that you need to be doing it across tens

of thousands, hundreds of thousands of

games to get confidence in what you're

saying. And that's what provides other

people confidence as well. Now

naturally, if you're using solo Q data,

people will say, well, it's not pro. So

if you'd have to use both. But like if

if you ever want people to have

confidence in these kind of numbers,

they need significant sample sizes and

strong evidence.

>> Some people some people do have

confidence in these numbers.

>> 100%. 100%.

>> Yes, some some people do. Yes.

>> But why do they have confidence if there

isn't strong evidence that I assume they

>> because I've done it for them enough

times.

But you shouldn't be so so you shouldn't

be so confident in these kind of

approaches. Like they might be

confident, but maybe they're wrong in

their assertions that they're confident.

>> Maybe I'm wrong.

>> I don't think it makes sense.

>> Just Okay. So, because you don't think

it makes sense, it doesn't make sense.

>> No.

No. I'm saying that I think you need the

stronger evidence to form good

conclusions. I think you can make these

statements as much as you like but the

outcomes is always clear which you've

already felt throughout a very long time

many years which is always negative

negative

>> and that might just be because of your

personality or it might be because of

the evidence and I'm coming it from the

perspective of the evidence

that's it I might be wrong but this is

what would make me a lot more

>> I don't know a good word for it pleased

with kind of those facts being thrown

around

that's

But but you're not the the target for

what I'm trying to do.

>> It's not for you. You're not an aspiring

pro. You're not in a pro that I don't

interact with.

>> I know. But um so obviously you won't

know anything about me. But

>> sure,

>> I put a lot of time into the data

analytics perspective of League of

Legends. So I spend a lot of time around

that. And what matters to me in terms of

providing sufficient evidence would be

that kind of large sample sizes across

many games. And just to make it clear,

that's actually not as hard as you

think. Once you have it set up, it

becomes very quick to make these uh

assertions and deductions without just

having to go from your formulaic

approach that you have now. So once you

set it up, if you were ever inclined,

that might be quite easy, but it's just

about going through the effort of

setting up in the first place. And from

my perspective, that's important, but it

doesn't have to be from yours. There

there's there's nothing easy about a lot

of various different things that need to

be discussed because some of them

gonna I'm gonna I'm gonna copy paste you

uh can you read quick? Are you a fast

reader?

>> I'm on my phone so I really highly doubt

that I can do very well right now. Maybe

you can give me a summary.

>> Okay, I will I will read it for you.

Okay,

>> thank you. Okay. So, I I I've referenced

this a lot because this was something

that I thought, okay, this this this

made a lot of sense um in terms of the

difficulty with some of the stuff that

even you're describing right now,

especially when talking to uh the

public, right, who lack discernment. So,

>> the need for technical language does not

result from only ideas being

communicable in this or that particular

word. Particular words possess no such

power as this. Rather, the need for

technical language results from the need

for an economy of thought and writing

that is the condition of referring to

and thinking about complex ideas.

Imagine if you had to explain partial

differentials to someone who under

whatever example of math you ended your

math education with only and if only

every single sentence and indeed every

single word you had to begin the entire

process of thinking about mathematics a

new as if your reader could not even

accomplish the most basic addition but

rather needed the very meaning and means

of such processes explained to them at

length. And if you could never trust

your reader to refer back to any such

explanation once given in a previous

section, but to repeat them in each

section of the text that is needed, as

if your reader's memory was wiped after

each section, how well could you explain

advanced math under these conditions?

>> Yes, uh I understand that. Um, so

advanced math,

>> so essentially you're saying that

difficult topics require too much nuance

for people without sufficient

understanding to appreciate what they're

saying and then you'd have to backtrack

too much to be able to provide a

thoughtful conversation on the topic.

No, it's it's the inability to have the

discussion without the need for

technical language because the readers

or the person that you're attempting to

communicate it to is incapable of doing

it because they're not at a level that

can understand it.

>> Sure. But I don't think what we're

discussing right now about the

evidencing, at least the initial point

comes under that, you know, supremely

difficult to understand advanced

mathematics topic. I think it's fairly

easily appreciated by a lot of people.

>> How do Okay. So, so how do how do you

how do you explain win rate probability

to a gold player?

>> Uh the the average outcome based on

assess.

>> Okay. So then how do you as an expansion

of that, how do you explain the

fluctuation of a win rate probability or

a hypothesis of one in a real game in

real time based on micro things that are

happening all across the map and how

that ultimately bleeds into something

larger or you explain like deltas that

factor into this.

I

>> I don't even think I would have that

conversation. I think that's becoming

too deep now.

>> Yes. Thank you. Now we've arrived at my

tweet.

>> No, because the tweet exists, right? In

which case the tweet wouldn't exist with

that suggestion, right? But

>> but but the tweet can exist as me

essentially saying here's where the

water is. It's not on me to make you

drink it.

>> No, I correct. But I I appreciate that.

It's more like the response from the

public is a result from that lack of

evidencing

from the tweet and it

>> existing,

>> right? But ju just less than 60 seconds

ago, you already agreed that there would

not be discernment. So what does it

matter?

>> Because I think people can appreciate

the fact that the evidence exists,

right?

>> Sometimes that has nothing to do with

what I'm doing. That has nothing to do

with what I'm doing.

>> It has nothing to do with it. It's an

appeal. You're appealing to emotion.

>> It has nothing to do with what I'm

doing.

>> I'm not appealing to it. It's more like

it's a factor that I I think is is

important. [snorts]

>> Why is it important?

>> So, I think the evidence is important,

but the fact that people can't discern

it doesn't mean it shouldn't exist. It's

like if you wanted to prove something is

true, you wouldn't provide the proof at

all because you know it, but anyone

reading it wouldn't know it. Why would

So, you think the proof wouldn't need to

exist? That doesn't make sense to me.

The proof exists for peers.

>> Yeah, but what if the peers couldn't

understand it either?

>> I know they do from what you say.

>> So then you discuss it with the peers.

So So now now I put out the tweet. So if

a pro player sees it and DMs me and asks

about it or wants to have a follow-up

about it, now we can.

>> But then there's no proof provided,

right? In my opinion, what do you mean

sufficient proofed?

No, I I mean sufficient. I mean

sufficient in my opinion. I don't think

that the amount of evidencing provided

was sufficient to have confidence in the

outcome. That's why I was saying the

tens of thousands, hundreds of thousands

of outcomes from Proplay and Solo Q

games. That's my that's my approach and

that's why I view the tweet negatively.

That's the idea.

>> So So why why does the calculation of X

Y equals Z why does that not matter to

you? What is so difficult about X Y

equals Z? Why can't you go through a

proplay VOD and and do X * Y equals Z?

>> Because it becomes one. It becomes N

equals 1 evidencing, right? Xod

X * Y equals Z. Sure, that's a formula.

But how do we know the formula is true?

Because you could have just made that

up, right? How do we know that captures

everything associated with

>> this isn't some madeup formula. I if if

you don't think that it's true, tell me

what is erroneous in its construction.

What I mean when I say it's not true, I

mean it doesn't capture the entire value

of what void grubs provides to the team.

That's what I'm saying. And I think

that's very difficult to do in a simple

formula. Almost impossible.

>> Correct. Which is if you read my tweet,

>> I said that it's important to quantify

why we do things.

>> Yeah, I agree. But my quantification is

different from yours, right? That's that

that that's that's exactly why I came

out with that response because you're

not providing the level that I think is

strong evidencing. And frankly, like if

you look at evidencing within scientific

studies, it's actually similar as well,

right? You need a massive general sample

size to confirm the the existence of

certain things. And I think short small

sample sizes, 10 to 50, are just

insufficient, especially when larger

sample sizes exist.

Oh,

>> how do you get sample sizes of not how

do how do you get sample sizes that go

against the idea of prioritizing grubs

if people don't do it?

>> Um I think a ba baseline would be okay

what is the win percentage of teams

right one second

>> that doesn't matter

>> what is the win percentage what is the

win percentage of teams that take void

grubs okay and then you see a

differentiation from 50%. So in LCK at

63% for the uh 63% of the time the team

that takes Rift Herald wins. Okay. What

does that tell you?

>> Now we need to now we need to understand

okay what is the context behind this 13%

difference from 50% which would be

random.

>> And then we need to look in the data

okay how much damage are they doing to

the towers?

>> Does the gold contribute to this

difference?

>> Maybe it's because the teams naturally

have a fighting ability earlier on and

therefore that's what's causing the

difference and the void gobs aren't

adding any difference. But there's a lot

of factors that go into how much value

do void grubs actually have in a game.

Maybe more attacks, more auto attacks on

the champions. Maybe if had action mid

instead, you'd get more value from void

grubs. Who knows?

>> But the fact is,

>> but but void void grubs have a static

number of damage that they do. Are you

familiar with it?

>> Yes. On on each auto attack, right? And

then on auto attack, it burns. How many

seconds does void grub burn?

No, I I I don't know that. I don't know

that. I know it deals damage. Okay. How

much damage does void grub do if you're

a melee champion and you have one void

grub?

>> I'm not I'm not aware. I'm not aware.

>> What if you're a ranged champion?

>> I'm also not aware. I don't have the uh

the knowledge of what the the damage

does. I just know the kind of practical

application of it. It burns the town.

Why are you operating from the

hypothesis that what exists

>> is good or better? Why don't you

quantify it?

>> As in like why don't I go through out

through with my approach and then prove

you wrong in that way. Is that what

you're saying?

>> Evidently, yes. Because that that's also

what the exposure of of of peer-reviewed

papers opens it up to is that if I am

wrong, someone would be able to come

along and just exert that I'm wrong,

right? or like or exclaim that I'm

wrong. Sorry.

>> That's that's fine. That's fine. So like

if if you need someone to disagree with

you, they have to provide the evidence,

then that's fair. But I suppose I'm

coming from it from a more basic

perspective that I literally looked at

the tweet and I wanted to provide my

view on what is strong evidencing and

why I didn't like the the assertions you

were making. I didn't have time to go

through all the void grubs information,

but yes, that's a natural follow-up to

prove that I'm I would be correct or you

would be correct.

So if if if I see a chess tweet and I

don't know anything about something on

some opening and I just don't like how

the tweet reads, but I don't know

anything about that opening, should I

reply to that tweet?

Probably not. No, but the general

principle of strong evidencing is

carried out throughout research

practices, right? So it's not really

super niche like that, right? Having a

strong general large sample size is I

think one of the highest forms of

evidencing within the pyramid and n

equals one uh personal experiences are

near the bottom.

Okay. I I want to I want to ask you a

question as because I'm going to go back

to something that I said earlier which

is that I often try to skew the tests to

be against my initial hypothesis right

as a as a baseline measurement to try to

Okay. So when I say that, what I mean is

I go into games and I try to create

scenarios that are most favorable for

void grubs to kill the turret.

>> I try to auto attack more than any

normal amount of auto attacks you'd be

able to find in a in a in a proplay

game. So if if on average, let's say

that an AD carry attacks eight times the

turret, they they throw nine auto

attacks or 10 auto attacks. in testing I

will go and do 15 or 20

because it it is it is on the other end

of the extreme. Okay. If I can then

determine that even in the extreme

example something that is beyond human

capabilities or something that is

inhuman is still falling short to my

other hypothesis.

What? What?

>> I'd be quite

>> I'd be quite surprised if you could

approximate the number of auto attacks

someone would do, not in a single

instance, but across the entire game on

average. Like if you if you could skew

that significantly significantly in your

favor,

I would be surprised. But it's not in a

single instance when you're attacking

the tower. It's more like across the

entire game, across all the different

champions, what does this look like?

That's really difficult to do. I I I

don't really believe in someone's

ability to do that on their own

successfully every time. Maybe you do it

once correctly, but every experiment you

do is always

done judiciously and appropriately.

That's very unlikely.

>> Why is it unlikely? Because you wouldn't

do it or you you don't interact with

people that do it. So you think that

it's not possible. That's also facious.

>> No, because it's just human error. You

know, everyone makes mistakes as normal.

Yes, people make mistakes, which is why

I used the word estimate earlier.

>> Yes, that's that's fine. But um forgot

to

>> turn that off. Sorry. Um it's it's

mostly about being sure all the time and

having that general case.

>> But I'm not sure all the time. I put it

out in the public because I'm not sure.

>> But that's that's where you get the

negativity from. That's my part doesn't

matter.

It does it not?

>> No, it it it doesn't it doesn't matter

because if I get the DM from the pro

player and I'm then able to help the pro

player, that is what matters.

>> I think that's a different conversation

we're having now. But that's fair if

that's that's how you feel about the

tweet in an emotional sense and what

like practically you can provide to pros

versus normal players. That's that's

okay. But I'm mostly coming at this from

the evidencing uh perspective. No, but

I've already cited to you instances in

which I've done the thing that you're

claiming to and it doesn't actually have

anything meaningful. And then I can

logically explain to you how the absence

of discernment also just means that it

doesn't mean anything either. Because

ultimately what needs to happen is that

the person that has the ability to do

discernment, one that possesses

discernment must follow up with me

anyway.

They're going to do that anyway.

previous example.

Your previous example on the AISA for me

wouldn't have met the threshold for

confidence. It was a lot of games and

it's a lot of effort, but doing maybe

10, 20, 30 games isn't quite at the

limit of a general population survey

tens of thousands.

>> So, so then go and skew it in favor of

the auto attacks, which you can do. Then

I need to be confident in in your skew.

>> Like has your skew been applied

correctly?

>> It's not about meense.

>> It's you can do it.

>> Yeah. I mean I never

about who it's never about who says

something. It's what's demonstrable.

>> I'm not right or wrong because I'm LS.

I'm right or wrong for this reason

>> or I am more right or more wrong.

But interactions with your messages on

Twitter or X is never going to come with

someone else coming with a full page of

evidencing from their own notes. Always

very unlike it doesn't make sense.

>> You'd only get sometimes it actually

does. No. No. It's just not true. That's

just not true because sometimes it does.

Yes.

>> But very rare.

>> It's rare. But it's true. It It's rare,

but yes, it's uh it is uh I mean it

happens.

>> Right. Right. But um I I mean that's

fine. That's fine. I guess it does

happen, but it's very rare as as we both

know. But I feel like

>> I feel like we're going in circles a

bit. I kind of appreciate your

perspective that you've provided

evidencing from your own form and you're

confident in your approach

>> and [snorts] you care about the people

that receive it well as it is.

>> Brother, you can you can go and do these

things yourself. Why why is there such

an aversion to it?

>> You you realize that Kai'Sa is not

Kai'Sa,

>> right? Kaisa is not Kai'Sa. Kai'Sa is a

a string of numbers and letters that

produces the figure in Summoner's Rift

that you associate and you name as

Kai'Sa.

Okay. Kai'Sa comes with a set static

amount of damage that if you then equip

her with certain items changes to

different numbers. You can open a math

formula and skew the test to be

favorable for auto attacks that you can

determine a human would not

realistically be able to do. And then

being pragmatic about it, you can then

try to quantify what it is that you're

trying to figure out in terms of what is

better, what provides more options, and

then you can you can expand upon that.

>> Look, it it makes sense in principle

what you're saying. Sure, it's possible.

The problem is is that these things

always come with error and I believe the

error provided by a n equals one human

example doing it by themselves

interpreting these formulas on their own

is too high. That's where the stronger

evidencing from large population was

coming.

>> Okay, so how did how did lane swap

start?

>> I I have no idea. No idea.

>> JESUS CHRIST, MAN. BECAUSE SOMEONE did

what I'm talking about and concluded

that the advantage was better. So it

starts

>> you're handwaving away n equals 1 but

then n equals 2 then it equals 3 then it

equals 4. Someone has to start it.

>> You start it with a scientific based

approach.

>> Include.

>> Yeah. I mean I don't I'm not saying you

shouldn't do it at all. I'm saying to

support it effectively and have a strong

conclusion, you need the general

population sample size. I'm not saying

to not do your approach on your own

experimentation.

>> THE GENERAL POPULATION DIDN'T LANE SWAP.

SO, HOW do you need the general

population sample size? It did not

exist. So, what are you saying?

>> Mhm. But this is this is a very niche

example, right? You're saying

>> I can give you more examples. The

absence of evidence is not the evidence

of Oh my god, man. I can give you more

examples. You not knowing about them

doesn't mean that I don't possess them

or you not having them readily available

or something doesn't mean that they

don't exist. This is this is so

ignorant. You're appealing to ignorance.

>> But we're kind of going off topic from

the original point.

>> How is it off topic? I'm trying to

explain to you how logical things are

followed or that how scientific

approaches would be done uh following a

hypothesis that something might be true

because some evidence for it seems to

exist in the data and then you go out

and you test it and then you keep doing

it. But your logic examples like lane

swapping I wouldn't have responded in

the same way because it's extremely

niche. You're not going to see it come

out in the data. However, things like

void grub value is not extremely niche

and this was absolutely come out in the

data and it's very easily represented.

Well, not very easily represented.

>> Why? Why? Why? Why? Why? Just because

everyone does it.

>> No, because for example, what is the

value of void grubs, right? You have the

XP, the gold, and the damage on towers,

>> right? So, okay. So, we have we have

some variables that we can quantify.

Okay.

>> Okay. So, yes. So, how much how much

gold does void grubs give?

>> No idea. 30.

>> Okay. So, you didn't even read my tweet.

You didn't even read my tweet.

>> It's It's not the the purpose isn't

knowing the specific value.

>> You didn't read the tweet and you

responded to it.

>> Well, I can I [snorts] didn't remember

the tweet, but yeah, I did read the

tweet. That's the difference.

>> Okay. How How much exp does void grub

give?

>> 65.

>> Okay. Nice. Great. Yes. Okay. So, it

gives 30 gold and 65 exper. Okay. So 90

if you get all three and then 195 exp.

Okay. So how much exp does a champion

need to level up if they are level

seven?

>> No idea.

>> 880.

>> Okay. How much gold does a non-cannon

minion wave give?

>> No idea.

>> I thought that you interact with league

data.

>> I do, but obviously I didn't go through

the league data before answering. Right.

That that's the main point. But I could

if you'd like.

>> So the fundamentals of the game when

discussing a topic that surrounds the

fundamentals. You don't think that's

relevant?

>> Yeah. But now comp compare the cannon

wave with the persistent effect of void

grubs across the entire game on average

for every game. What does that look

like? How would you even know?

>> By running thought experiments and then

also evaluating it.

>> But again, this is a human experiment.

Lots of error. You don't have an

appreciation of

>> most things are human experiments.

>> We don't have AI. We don't have

[laughter]

machines running things. We're not

scriptors. If we were scriptors, I would

say that [ __ ] B1 Carus every game.

Like

how do we do anything?

>> Well, mostly what I'm saying is, yeah,

you need to do human experiment, but you

need the sample size to support your

views. That's the

>> No, you don't. Because sometimes n

equals 1 ends up being correct.

Yes, sometimes. But usually when you do

n equals 100,000, it's it's wrong. It's

right all the time. That's the

difference.

>> But you can determine if n equals 1 was

correct for other reasons. It's I if

you're if you're going and you're you're

attempting to experiment or test

something that is quantifiable using

numbers or other types of things, then

you can Yes. Like handwaving it away

just because it's a low sample doesn't

mean anything.

because then that you shouldn't do those

thought experiments. I'm not sure if

that's what you if that's what you think

I'm getting across,

>> but I'm saying in order to conclude

these thought experiments and take them

further, you really should do a large

population sample size to confirm.

That's that's really it.

>> Yes. And how do you get a larger sample

size of the population to do something

if you're claiming that it shouldn't

start because n equals 1?

Uh it dep I mean like for the void grubs

example it exists right for the lane

swapping example I have no idea that's

something I'd need to think about much

more.

>> So how do you think I if if if I

believed that void grubs are like this

what is the best way for me to influence

pro play?

>> Um [snorts] I don't know I think it's

highly networking based right. Uh, if

you mean in terms of like proper

evidencing, I I provided that view

already, but I can't give you like a

really good approach to influencing

proplay. Frankly, I don't know if anyone

does.

>> Okay, sure. So, you you mentioned

networking. Okay. So, what did I say

earlier? Who did I say that the tweet is

for?

>> No idea. I don't remember that. I was

focusing on the numbers, brother. It was

like it it was like less it was 10

minutes ago.

Well,

>> I think longer than that, right? We've

been talking for a while.

But yes, I don't remember.

>> I said that the tweet was for pro

players

[snorts]

>> or aspiring pros.

>> I'm the player, right?

>> Okay. But you just used the word

networking. So is pro players and

aspiring pros not falling under the

umbrella term of networking?

>> Yes. Yes, it is.

>> Okay.

So

then what?

>> I I I I don't I'm not following the

point entirely. As in

>> So I ask what if if if if N equals 1 on

lane swap, right? And that's how it

starts. How do I if I think void grubs

are bad? What is the best way for me to

influence pro play? And then you said

networking.

>> Okay. So obviously setting up scrims to

set up the scenario, I suppose if I had

to guess, but this is really like off

the top of my head.

>> Okay, sure. So set set up a scenario.

Okay, in a scrim, but then you would

handwave that away as n equals 1.

>> No, no, no, no. In this case, it's too

niche for you to be able to provide a

population example. That's that's the

difference. boy does not

>> what

>> that's that's my problem with

>> that

>> I can't how are you going to look at

like you said before you're not going to

find lane swapping examples when it

doesn't exist in the first place but you

are going to find examples of void grubs

>> uh value because

>> everyone takes void grubs every game

almost

>> that's the difference

>> everyone does take void grubs but almost

no one intentionally foregoes void grubs

because of other

factor in variables such as like win

rate probability and stuff.

>> Sure. Sure. That makes sense.

>> Okay.

>> But uh it's just looking into that using

the data, not just an N equals one

experiment for void grubs. Perhaps for

lane swaps that's fine, but not for void

grubs because that is easily

quantifiable information that's

available in the public view. Like they

all take void grubs. That's available.

>> Yes,

>> that's the difference.

>> So does that make it does that make it

true that they all take void grubs? So

it's good.

No, that that would be a horrible

extrapolation.

>> Okay, correct. Nice. So, okay, we agree

on that. So, okay. So, now how do we

Okay, so now teams have void grubs. So,

now how do we determine the amount of

value on average using an estimate that

void grubs is providing to the teams?

[snorts]

>> Well, they influence XP, gold, and

towers, right? So you'd look at tower

damage, you look at towers taken, find

an average, and see how much gold they

extract from those towers across the

entire game.

>> Yes. Okay.

>> Okay. So, how are you going to calculate

how much damage on average the void

grubs are doing to the turret?

>> Oo, that's a tough one. Um, I think

you'd have to average across the games

where one team has it and one team

doesn't. That would make sense, most

sense to me. But I'm not sure if there's

a more specific example in the data.

>> Sure. Okay. So let let's say that you

take you take the example of the team

that has the void grubs. How do you

determine how much damage they did?

>> Um can you look at the autos? Maybe

>> um

>> what do you mean? Why would you not be

able to look at the auto attack?

>> It's very difficult. Like if you're

looking in a single game, how can I

determine how many times they've auto

attacked? I need to account for their

items and therefore the amount of

physical attack damage.

>> The items don't alter. The items do not

the items do not alter. Touch of the

void.

>> Yes, but I need to calculate how many

auto attacks they're doing.

>> Sure. Okay. So, we just defined X in my

previously mentioned formula. Now, what?

>> And then we What do you mean? And then

we aggregate the data from the

population samples. That That's pretty

much

>> No, no, no, no. Follow up. Follow up. I

I No, no, no. I asked how do we kill a

turret?

So you you check their items to see what

they're

multiplied by Y, right? Equals

>> what do you mean equals the total damage

from

Z. Correct. Perfect. We got there.

>> No, but

but the difference is you don't know how

many auto attacks happen on average

within a game, right? You need So we're

going to get an estimate. Yeah, we're

going to get an estimate. Yeah.

>> Yes. But make sure your N estimate is

confident. That's why you need the large

population sample difference from your

mind,

>> right? We But we don't need it to be 100

confident. We need it to be able to show

that it likely produces higher positive

results.

I I don't I don't understand what you

said there. That didn't make sense to

me.

I mean literally I I genuinely do not

mean disrespect with this. Okay. It

feels like to me that you are in out of

your depth because it seems like a lot

of things that I can either break down

or explain to you about the game or

processes that someone should use to do

these types of things you either seem

unaware of or you're not used to doing

them. Nor do you understand the the the

layers that go into it. and then you're

telling me that I don't understand it or

that I haven't done the evidence or like

whatever else. But then this is just the

tip of the iceberg.

>> You never said you don't understand

anything.

>> It was actually a very simple point from

the beginning

>> that in order to have confidence in

these values, you really do need to

provide evidencing from the population

size that that from the large population

size. That is really it.

>> But I just walked

at the start of the call.

>> Yes. But it it it doesn't it doesn't

affect it doesn't affect like the

reality of that evidence requirement.

There's a reason why it's required in

research.

>> Like you can do n equals one experiments

and they're fine but it doesn't I go

game by game.

>> So okay so you open up

>> how many games do you need to do? That's

why I also mentioned to you earlier that

you can skew the test in a nonhuman

manner to also get an estimate.

So, how many

>> times does a champion attack during the

game when they have void drops? On

average? Like, do you know that

>> off the top of my head? For what? For

for what game? On average, yes, it's

about eight. Going off of MSI games,

>> a champion attacks the tower on average

for the whole game. Whole game.

>> In Midame.

>> No, but I want the whole game because

void goss applies for the whole game.

does apply for the whole game but god

man you need to be pragmatic in your

approach what is pragmaticism

>> I don't have a definition but it's

something like you know take the best

available information

>> whether or not it's right or wrong

>> being pragmatic

by the way I have a bust

>> sorry being pragmatic means dealing with

situations sensibly, realistically, and

in a way that prioritizes practical

outcomes over abstract theories or

ideals.

>> It focuses on finding workable solutions

based on existing conditions and real

world consequences rather than sticking

blindly to fixed rules. Okay? So if you

go and you take a test and you skew it,

which is not pragmatic, and you skew it

to be grossly in favor of the opposite

of the hypothesis uh thing, and even

that

>> can't provide enough value, then what?

>> Uh I'm not sure. I think your skewing

idea is actually appropriate. However, I

would like you to approach it from my

perspective of the evidence rather than

the way you were approaching it already,

which is like the 15 attacks on on a

tower. Like, I'm not looking at mid

game. I'm looking value from the whole

game with void grubs. And it's very

difficult for you to know how many

damage how many times each champion is

attacking per game on average to get

value from void grubs and then the gold

from the towers as a result. That's

difficult.

>> I don't know. [laughter] I don't know

how we even

>> because that's not it's not pragmatic.

Okay. When Wukong got nerfed a couple of

years ago.

>> Yes, you do. Yes, you do. When got years

ago,

>> literally the data is available.

>> The data doesn't matter because the data

can have its variables altered or an

introduction of new information which

would then demand that you apply

variable change.

That's that's a little bit

it's what is untrue about what I just

said.

>> I don't even know what you said that

doesn't it's a bit too too many long

words like just make it simplify what

you're trying to say. the the the data

can be okay. If you have the data from

2014 and everything that they believe to

be true and you take Faker from 2026 and

you time jump him back to 2014 and he

starts doing everything opposite to what

other people are doing because he has

new information.

>> Sure.

>> What does the data say about Faker?

>> Uh he's well if he went back from 2026

to 2014 it'd be insane, right? You'd be

the goat. But but but but what does the

data in 2014 say about things faker is

doing?

>> I I have no idea. I'd need to check.

>> But I thought you had the data. So 2014

should just be able to to to say by that

logic.

>> I genuinely don't have data before 2015.

I only care about

>> more recent events. I understand what

you're saying.

>> It's not literal. It's not literal.

>> Mhm.

>> Baker would be new information. He would

be injecting new information and new

variables.

>> But there's not many

I I don't understand.

>> There isn't so many niche variables that

wouldn't come out within the general

population sample size, right? It

provides a lot of data. I think it's

it's in 2014

if Baker if you went back in time and

told people about minion wave

manipulation in 2013 or 2014 they'd say

you're [ __ ] because they have all

the data.

>> Um I would say that they haven't been

communicated with properly to establish

the facts which is mostly the case.

>> So even if you [laughter]

prove something that's obviously already

right

>> you miscommunicated.

So, but how do you prove it if they just

say that n equals 1?

>> Well, because don't use n equals one. My

point would be use n equals more.

>> If it's completely absent,

>> then yeah, you have to use n equals one,

but you need

>> how do you get n equals more if you

never start n equals 1.

>> This is getting a little abstract, but

it's

>> it's not getting abstract.

You're you're talking about entirely

absent information from the game. That

is pretty abstract. Like this is not

normal normal state. But

>> one second. I need to check if my boss

is coming. My lord

in a difficult position here.

[snorts]

Uh actually I feel like we hit a little

bit of a

of a block. Would you like me to go away

do the analysis and then come back to

you with a comment and then we'll see

what happens from there? Or would you

rather just

>> Well, no. I would I would just say I

would just say your analysis is n equals

one, right?

>> Well, no. This will be a general

population sample size from like solo Q.

It would be solo Q. Is that fine?

>> So, you're you're appealing to

population size now.

>> Yes. That that is that is actually my

whole point really. It would be

initially

>> Okay. So, so lane swaps never should

have existed. Lane swap should never

exist.

No, that's a that's a niche that's a

niche thing. I think

Nivia should never be picked top lane.

>> No, that'll be in the data for sure.

There's a lot of

>> How would it be in the data before

someone did How would it be in the data

before someone did it?

>> Uh, a lot of people started it anyway,

even if it's [ __ ]

>> Yeah, it always

Yes, people do it anyway, even if it's

[ __ ] That's the Yes. And people win

without void grubs. Of

course,

>> people win with void grubs and it has

nothing to do with the void grubs

>> naturally. Yeah, that happens all the

time.

>> Okay.

>> But what is the value? Right. That's

what we care about. What is the value of

>> Exactly. Exactly. What is the value

exactly? So let's quantify it.

>> Yes. So my approach is different to

yours. That's that's the difference we

have here. You like n equals one. You

think your approaches are

>> my god. you the the the best approach to

try to quantify it is to do the thing

that the human can't and then start

there as a baseline. The reason being a

couple of games ago on EWC stream today,

Yike Yik did not kite his wolves when

killing it.

>> He just visibly didn't. Okay. As Nefiri,

he did not kite the wolves.

>> So when he gets to the gromp, he's

missing damage on the gromp that he

would have had he kited the wolves.

Yes.

>> Okay. So, do you want me to go away and

do this do this analysis and then we can

have this conversation again which which

would be fine by me but I understand if

it's not fine by you.

>> I don't believe based on things that we

have discussed that the types of

questions that you would ask yourself to

confirm validity of things would be

appropriate.

Yep.

>> Uh, are you there?

>> Yes. Hello.

>> Let let me do the let me do the analysis

and then if you're not happy with it,

just let me know. That's fine. You can

just discard it. That's I I don't really

mind. It doesn't take me too long. Maybe

a couple of days.

>> You just discard it and it makes it

easy.

>> Definitely possible. Now, how are you

saying you're going to go do it?

Uh,

>> general, I've got I've got the data.

>> What data?

One second.

>> Uh, soloq data.

>> What does solo Q have to do with pro

play?

>> Uh, hello.

>> What does solo Q have to do with pro

play?

>> No, no, it would be a proxy because I

don't have access to the pro play.

I don't have grid, so it has to be solo

Q. Do you know what I mean?

>> So, right. So, so, so go into

>> I'll give it as a

>> proxy.

>> skew the data to be in favor of void

grubs.

>> Let me use my my approach and then you

can confirm my approach. I think that's

the best way forward else we're just

going to be going in circles.

>> No, no, wait.

>> I'm just I'm just so brain blasted. I'm

just so brain blasted. All right. Yeah.

Have a nice day. Yep.

>> You too. Bye.

That was wild, by the way. That was

absolutely wild. This whole conversation

is just wild. Well, I could tell that he

got upset when when he realized that he

did this. I I could I could tell he got

upset when he realized that he actually

did X multiplied by Y equals Z. Dude, I

don't know. Like I I I I wish I was more

articulate in in being versed at

handling this because this was this was

frustrating. I mean, I was being mega

condescending to him. Hey, Fraggy. It's

because we're we're an hour deep into

the call, brother. We're an hour deep

into the call and he's coming to me at

Imagine you go up to essentially a

professor or someone that like has like

a [ __ ] uh like like their entire life

is in something and then they can tell

that you're not getting certain things

and then you tell them that something

doesn't make sense. Of course, it was

straw manning. He was also gish

galloping. It was just fallacy after

fallacy after fallacy. And I could tell

that most of it was rooted in his

emotional conviction. Well, that was the

other wild thing, by the way, is that he

was forgetting things that I had just

said moments prior in the conversation.

But then also, he essentially didn't

read my tweet. It's It's just so crazy.

But he also he doesn't even Oh my god,

man. The data would tell him NEVER TO

PICK RISE AND PROPLAY because Rise is

solo Q win rate. SO IF IF RISE'S SOLO OR

LIKE [ __ ] Casante or Mel

This is so crazy, man. He It's like It's

like he can't figure out or factor how

to do something if it hasn't been done

before. The t the Faker time jump bit

was soon infuriating. So infuriating for

him to not understand

because Faker would do stuff that's not

in the data. That's the whole point.

But also, it was so weird how he was uh

so concerned with

he like the way that he would view how

he would want to do things and that I

should want to do it from that angle.

That's so weird, dude. It's so It's so

wild. It's also wild to how he can't

understand that you can skew a test to

be in favor of void grubs and then use

that as a baseline to argue against. If

you can skew a test, what is so hard

about this to understand? If you can

skew a test, if you can, if you can go

into proplay games, if you can go into

proplay games and you can determine that

void grubs, um, you know, on average,

like you, you know, you go across a

sample size of, you know, various

different, um, games, whatever, blah,

blah, blah, and you get like an

approximate, right, of how many auto

attacks and how much damage the void

grubs are doing to the turret, or you go

into the replays, etc. Then uh then you

go into practice school and you skew it

even more. The the the thing that he

wasn't understanding when I gave the the

the Nefiri example or like it reminds me

of Max Waldo on Wukong, right? In 2023

when Wukong got nerfed and pro players

said that Wukong was going to be dead um

because it was going to affect his clear

speed. Max manually went through some of

the best junglers in the world playing

Wukong and he jotted down their clear

times, right? He calculated the speed of

their their clears. Then Max went on the

PBE with the nerf and he performed

faster clears than the best junglers in

the world because he was he was paying

attention to the clear and he was

maximizing patience. And ultimately what

that shows is that Max did an extreme

example, right? He did something that if

someone were to actually just be paying

attention or, you know, be doing these

types of things on this, you'd still be

performing better than these pro

players, etc.

Well, the the other thing is that his

logic falls short for so many things in

life. And how could you ever live this

way?

It's so weird.

Like, like, okay, by his own logic, if

only Geng T1, and Hanwa Life pick uh

Garen top lane, and then Garen has a

positive win rate in all of LCK, they

play it for an entire year, right? Does

that mean Garen's good? because Garren's

win rate. No, it doesn't mean Garen's

good. That doesn't make Garen a good

champion. The the other thing is is that

on a sometimes you you can go and you

can you can auto attack a turret once or

two times in Midame, right? Or three

times maybe. And then what did that what

did the void grubs actually do? Well, if

you had three and you're a melee

champion, it did 48. It did 48 damage.

If you were a melee and and and if if

you weren't a melee, let's say you're

you're a range champion. Let's say

you're an AD carry or an AP carry,

right? And you auto attack the turret

three times. Wow. So amazing.

So amazing. You did, you know, you did

24 damage and then maybe you refreshed

one of the stacks. Um and then you made

the void grubs. Uh you know, may maybe

it did 48 damage. What is 48 damage?

What is 48 damage to a turret?

What is that? It's less than an auto

attack.

It's less than a quarter of an auto

attack.

What are we actually [ __ ] saying? How

do you also how do you even quantify

that value? And some things are

extremely difficult to quantify, but at

least we can get these numbers. We can

get these approximates. We can get these

estimates. The the other funny thing is

like there's so many variables he's not

even realizing or are being entertained,

right? Uh the ability to juggle or like

paint multiple uh you can quantify it by

comparing it to similar actions.

I'm not exactly sure what that means. I

would need to follow up on that. So

maybe something. Anyways, so the the the

other thing is like, okay, there are

literally some games where minions just

live longer or aggro the turret for

longer and deal more damage to the

turret, right? Like there there's so

many things that are happening at once.

What if you didn't rotate to the void

grubs and you instead stayed in the

lanes and crashed a wave into the

turret? is the wave that you crashed

into the turret as well as whatever

amount of auto attacks that you did to

the turret, is it doing more on average

than the void grubs that you would have

pursued and thus had to delay hitting

that turret? Like there's there's so

many factors and variables that are are

compounded.

Oh my lord.