yeah, but that’s not what variance means. Also, I love how you attempt to insult me:
When clearly it wasn’t what you were stating and you were using the term incorrectly lol.
Yeah. Moderately compelling. Not the slam dunk your brain is playing tricks on you to convince it is, but it is evidence. Unlike “I split players by MMR and they had the same MMR.” Which come on. I’m not accusing you of being remotely intelligent, but even you know that was, of the things you can say, probably the single stupidest.
Again, had you found this years ago, I’d be more willing to believe you’d been swayed. Instead, you mind was made up when you had literally zero evidence and you stuck to your guns for years despite having not one single shred of evidence.
Now, you’ve got an explanation (by the way, we actually had an argument where you tried to deny that practice had anything to do with results lol) and still a ton of missing MMR and a ton of missing extra games and seniority and no explanation for any of it.
But it means what i say it means. Its an indication that the standard deviation is high. Meaning that looking just at the average is flawed. There is a difference between mean and average and sometimes you rather want to use mean and sometimes you want to use average.
The standard deviation is an indicator if its actually useful or not to make generalizations.
Meaning if you say that terran has most time played because of seniority, it doesnt mean that terran CANT have most amount of new players.
Only because you cannot grasp what it means doesnt mean it was stupid. Many here understood what it means and dont think it is stupid. Its actually a pretty good observation point.
Nope. I just said it wasnt the only important thing. Which is true. I gave you the website saying there is more than just putting time into it.
go read learning curve.
But can you now finally accept that terran is flooding low level because they dont put the work in rather than terran being hard?
1 Like
You don’t know what the variance is.
Average isn’t a technical term. When I use average. I always use it to mean “mean.”
I know exactly what it means. It’s not incomprehensible. It’s just wrong. It wasn’t a difficult to follow argument. It was a moronic one.
No, you and Kelthar actually tried to state the was no causation lol. One of the most entertaining moments I’ve had on the forums.
Never said anything about them flooding the lower levels. I stated that the problem with TVP was the difficulty. Something along the lines of what Showtime said. In fact, he was the first pro I talked to on stream that explained to me that Terran was the hardest. BeastyQT was the other big name.
BTW I’m now absolutely loving how activity has become the be all end all metric, but only as applied to the lower levels lol. Fun.
OK Buddy.
We agree to disagree
But If you want to read further read:
medium .com/@khang. pham.exxact/overfitting-generalization-the-bias-variance-tradeoff-5800f8c2200
High variance means that you cant make generalized Statements that easily anymore. Obviously.
You certainly didnt understand it because 2 of your counter Arguments were plain wrong. Neither does Same BMI means that the bodies Look the Same nor would the height difference disappear between men and women. You simply dont understand the Argument. You simply dont understand the topic.
Nope we did Not. Please get evidence of that. We certainly Said its Not the only important Factor and ofc there is no linear Progression. Playing 2000 Matches wont Always get you x mmr
Ah yes Sure. Now Go Prove that.
But do you agree that terrans dont advance from Bronze and silver because they dont get the work in ?
Because you Love it so much.
Because learning curve ? Have you read the Site i linked ? Ofc Activity is only extremely important in lower areas because of diminishing Returns and because you wont learn a new Thing every Game. Or do you actually think Progression is Always linear in every Game in every Thing you do ? That would be contrary to scientific results but OK maybe you are built different. i brought Up the topic a few Times and you still act surprised when we Look at Higher mmr people
Its funny how you can Not comprehend learning curve. Why is someone with 30k Games Not instantly gm?
Yes exactly this i want to critizise.
I actually looked it up and in english there is actually no actual corresponding concept. For example we use “midpoint” to describe “average” pay:
Midpoint:
The reference point that is midway between a range minimum and maximum in a graded compensation plan. It represents the intended average pay for all employees at a certain compensation level.
We do that because avg is flawed when the variance is high. For example when if you live in a city thats divided between 1000 guys that make 30k a year and 100 guys that make 300k a year the average will be heavily flawed and you cant make good assumptions about either side of the city (avg would be 54k, almost double of the “poorer” side of the city and nearly a sixth of the “richer” side of the city; meaning you could say “hey what do you want?! avg pay in your city is great!”). So we rather use the midpoint and in this case the midpoint would be 30k a year because if you list up all the guys with their income 30k would obviously still be in the middle of the list because there are much much much more people who make 30k a year than 300k a year. hopefully i have explained that well enough for you to understand.
An important lesson here is that its very important to look at data and analyse it in many ways because one perspective is always flawed in some kind of way. Only to look at avg would lead to dumb generalizations like you saying that all terrans are just simply worse (and because that cant be, it must mean that terran is harder)
2 Likes
So, looking at it more closely, there’s no large variance in any league. Ergo, there’s likely slightly higher activity per league, but since it’s single digit integer reporting (not to the nearest 0.1), the margins are simply small enough that it’s undetected.
You realize i used variance for your Argument that terran has the seniority meaning they cant have newer Players since the average Playtime of terran must be higher ?
You called me a moron when i Said that there is no contradiction…its rather actually the logical conclusion when the Starting Race is terran then terran has the Most experienced Players but also the Most new Players (based on Playtime only)
That being Said.
Exactly. There is no great variance in Activity. But the difference in Activity in General is next to 0 and sometimes actually 0.
Still: because of learning curve you cannot State that Activity at Low League means the Same as Activity in high League. That should be Crystal clear. Your Problem is that you speculate how Things should be based on nothing. No Argument or scientific result or Proof of concept.
Again: you have 0 Baseline for your assumptions. All i say is that its plausible that we See those differences that we can observe. Nothing more nothing less.
“Variance” or greater standard deviation, is the basis of your entire argument. I’ve just shown that there’s no detectable variance between Terran and the other races. You’re going to have to try again.
RIGHT! Terran players are more or less the same as Protoss and Zerg. But they still have drastically lower MMR. It’s almost like:
twitter. com/IEM/STATUS/970101525166075904
We have explained it over and over to you. Just Stop crying already. No one is believing you that terran is the Harder Race only because of Bronze and silver Players Not getting their Work done to advance to the Higher leagues.
No one Cares about Scrubs.
Here you go removing facts , "lets remove half the population cause I can’t prove my argument "
1 Like
Average is absolutely a technical term and also this usage is generically incorrect. One should specify which kind of average is in question, especially when talking about data.
Median (actual middle of data set) and mode (most common) are vital, and many people will actually use a different one based on what’s in discussion - when someone says the average joe, they most typically are referring to a mode average human being.
The average player is absolutely a joe and if we look at the most common ranking of each race, they’re basically equal, because they’re all bronze.
If you think it is moronic, the only conclusion is that you failed to learn what it means.
This is incorrect?
Yes.
When you perform a mean average.
When you perform a median or mode average, they’re the same. When you clip the top or bottom off, they’re the same.
What this tells us is a thing that everyone already knows: Mean averages are sometimes not very useful to the point of being actively misleading.
If we have two cities, one with a population of 1,000 and another with a population of 2,000; and in the city with 2,000 people, the average pay is lower, does that mean that the population in the larger city are being underpaid?
Not necessarily.
If you dig and find that in both cities, there’s 100 CEOs, but 50% of the population are $20/hr clerk/waitstaffs type positions – so 500 in the small city but 1000 in the large one – would you continue to be surprised that the small city has a higher mean average income across its population?
2 Likes
Its Not removing facts. Its explaining Them. Avg mmr is lower for terran. Fact. The avg mmr is lower for terran because of lower League Players. Fact (because Higher League is even/terran is Leading). Low Level Players Play less than high Level Players. Fact. Low Level Players are Low Level because they are Low Level and deserve to be ranked in Low Level because they dont Put the Work in. Fact.
Summed Up: terran AVG mmr is lower because of Low Level Players. And this is nothing to worry about since they dont actually Play the Game/Play so few Games that its obvious that they cant advance to Higher leagues. And Bonus: why should the abilitiy/disability of very Low ranked Players Matter anyway ? They havent got a clue about the Game so why bother ?
1 Like
Alright, time to drop the fact-nuke mother-of-all-bombs to end this debate once and for all:
https://i.imgur.com/7xoWOwW.png
https://i.imgur.com/VCTevFF.png
Bronze is 100% a skill thing.
GG NO RE
But wait, there’s more:
https://i.imgur.com/duTKOnM.png
and
https://i.imgur.com/3gEIN6z.png
Lmao. Bronzies have a 100ms reaction speed on average, 50 apm on average, and they basically don’t use hotkeys. GMs by contrast are reacting to events in 40 ms, have 200 apm, use hotkeys every millisecond. GMs are usually 22 years old, have basically nobody over the age of 30. Bronzies have a drastically higher age group, sometimes as high as 40 to 45.
I knew from my studies on educational performance that mental performance correlates with age and that’s why, in school, tests are standardized by age. You don’t give a highschool calculus test to a 10 year old. This is common knowledge. So naturally it makes perfect sense that age will be one of the strongest predictors for SC2 performance. 45 year old gamers are busy in the real world and can’t spam apm 24/7 in ezcraft 2, and that shows up in their performance level.
It shows that they followed the precedent set by Marineking. Break your wrists playing bio in 2010? Switch to mech. Break your wrists playing Zerg and Terran in 2024? Switch to Protoss.
So I am trying to sleep and have this pounding headache, right. I decide I have nothing better to do than to watch the iem finals. I still hadn’t watched them. Game 1 maru doesn’t check to see if the third is being droned. His hellions aren’t checking for move outs. Very basic mistakes. He is out of position for the first two fights because he’s moving his army without knowing where the zerg army is. He’s doing this during a vulnerable period while takingan expo. Also a very basic mistake. Zzz closed the finals and probably won’t watch the rest of the games. Hots era goat terrans would never make these mistakes. Serral really isn’t that great. It’s just the talent pool in korea dove off a cliff. Innovation struggling against a foreigner at WESG? Yeah talent nose dived after the kespa scandal and the team houses disbanded. Fast forward to today and the goat terran can’t bother to keep tabs on zergs army positioning. Roflmao.
Innovation used to build a depot at the farthest off counter attack path so he’d know if lings were headed for his fourth or third. Tricks like that allowed him to move confidently on the map. He’d also scan to know where the zergs army is. It’s really the basics of tvz. At 4 base your scv count is good enough you start to scan for positioning rather than muling. If you mule, you definitely don’t move out unless you happen to know where the Army is by sheer luck.
Depots are better than scans because they cost half the minerals and last longer. They are also bait. The zerg will send a few lings to look for them then some marines and a medivac collect some zergling kills. But it genuinely looked like maru was f2’ing his army without even knowing where the zerg arma was at. Literal siege and forget micro in iem finals.
Yeah the performance out of Clem and Maru was disappointing tbh. Pig puts it best when he says that Serral doesn’t just go god mode his opponents also play worse. Historically in the Clem vs Serral match, Clem would win the drawn out games based on mechanics but lose to Serral playing wonky things like the cavebob style. Then the zerg nerfs came around and Clem was mixing up his builds which allowed him to win at masters winter. Clem knew when to play aggressive and knew when to play defensive ghost/tanks and won with relative ease. At IEM I was actually surprised that Serral didn’t do any of that against Clem but rather save them for the grand finals against Maru.
In the interview with Crank, Maru said that watching Clem get stomped hurt his own confidence and since Serral was able to save his builds you could see how it ended the way it did. This makes me more invested in the obvious rematch once Gamers8 2.0 comes along. Losing at IEM made Maru hungry for the title and choosing all Zergs in his GSL group makes it seem like he knows exactly what matchup he’s focusing on.
I have never seen spotter depots used in TvZ, they always just make sure to keep their hellions and banshees alive as the anti run-by squad. I think they just got so used to the last map pool that Terrans got lazy. I mean you had Byun putting orbitals at the fourth in most of his games and punishing Zerg players for thinking they could attack it.
You didn’t “explain” anything. You rationalized.
See how quickly you turned from evidence based to ad hominem attacks? The instant your last metric was proven irrelevant.
Why don’t you go cry at Showtime about how wrong he is and how you’re the one who really understands Starcraft.
1 Like
I put it in context.
The only thing you ever cared about how it is on average.
You ignore everything else. You ignore that terran is by far the most popular race (and why would you assume equal distribution in mmr when the popularity is not equally distributed), is the tutorial race, the campaign race, the most familiar race compared to other rts, the human race. The race with the most guys in bronze and silver that is dragging the average mmr down.
No one is suprised that bronze/silver is full of terran. Nor does anyone care. Righfully so.
How and why?
Oh yes the master of ad hominem attacks cries when i talk about hes crying about the statistic that bronze and silver is flooded with terran and therefore dragging the average mmr down.
Please get a recent quote from showtime where you specifically ask him about the state of bronze and silver and why there are so many terrans 
1 Like
https://i.imgur.com/osUj5l3.png
This is one of the builds that would regularly bomb vs 4.4k terrans while streaming. Magically it beats a 5.4k terran. Stream cheating just tanks your mmr. It’s legitimately a 1,000 mmr difference. Playing on a new smurf account, it’s at 5200, and I have a 65% zvt winrate. I thought maybe they were profiling the builds based on the name, not stream cheating, so I decided to play some more on my slammer account and, what-do-you-know, they aren’t profiling the name.
I flicked through a couple innovation games 8 years ago etc. He had depots spotting the counter attack paths and also kept the banshee alive for spotting as well. The banshee stayed alive all game. His medivac drops covered the other side so he has a pretty good chance at seeing what’s happening. Maru lost his banshee for free in game 1 and didn’t have anything spotting anywhere on the map for attacks.
Maybe it’s a home court advantage, or jet lag, or they are discouraged facing a “tough” opponent, or they find the tournament setting distracting with lots of commotion.
Except serral has been doing the queen drop builds for quite awhile. I am not even a pro player and I keep track of this stuff better than they do. You don’t even need to know that he’s going to do that build. You can infer there is trouble by the late third, roaches crossing the map, no droning at the third. Imagine a liberator instead of a banshee. Innovation loved liberator hellion openers. It was impossible to allin vs them. What is a queen drop going to do against that. Lack of scouting, bad positioning, moving out without knowing where the zerg army is, subtle build order differences. Imagine what you could do with the hellions if you realized there was no zergling speed.
Yeah hellions are great but they aren’t going to stop the ultra ling bane attack at maru’s fourth. You gotta have fore-warning for that one. Maru did more or less the same thing game 2 with turrets. Turret vs depot is a matter of detection or supply. Both are half the cost of a scan so it’s better to mule and build a depot. It’s the same thing with zerg and overlrods in muta vs muta. It sucks to lose an overlord but it’s much more important to know where his army is at. You can be out of position for 5 seconds and take game ending damage. Terran has the efficiency advantage and zerg has the mobility advantage; just build a depot or a sensor tower and be in position and you win, for crying out loud.
Serral gets unbelievably lucky with his build order choices. It’s pretty obvious he knows exactly what build they will do and they have no clue what build he’s going to do. If I had to guess what’s going on, Lambo probably scouts out the KR server and tells serral what builds are being used. You can’t practice on the ladder, but the teamhouses have been disbanded so who do you practice against.
My solution to that problem would be to use SC2’s AI interface to train bots and practice against them. Alphastar was beating serral with protoss for crying out loud. But these pro gamers don’t have the know-how to setup something like that.
Really the best option for them is to queue a different account every game, deliberately throw games to keep their mmr random so mmr value can’t be used to profile them, etc. Use a hotkey remapping tools so you can’t tell who is who based on hotkeys, etc. If they did that, I think Serral couldn’t be quite as prepared as he is, because it’s obvious he has some intel on what builds they are using. One of the big differences between game 1 of innovation vs serral (wesg) and game 1 of maru vs serral is that it was pretty obvious Serral was flying blind and didn’t know what innovation was doing. Maru is too easy to profile. That’s really what it comes down to.
Byun does a whole bunch of nonsensical stuff, like going 1 factory siege tank in TvT.