RIP gsl you had a good run

To quote Harstem: “It’s not really a celebration, it’s more of a funeral”:

It looks like the GSL and ESL are being massively downsized. I guess Artosis saw the writing on the wall before he jumped ship to Canada. If it weren’t crystal clear already, it’s probably time for “pro” players to put their efforts into college. The golden goose is going to be laying its last eggs in the near future. They dropped a bucket into the well, needed a longer rope, and what water came back up was muddy – a drought is on the horizon. Get your rations before they run out. :man_shrugging:

I personally gave up on the SC2 pro scene at the start of LotV when tanks could fly and liberators could melt an infinite number of corruptors and unstoppable reapers could be massed on 1 base and zero-micro BCs could teleport into your main with zero warning. It was crystal clear to me that the pro scene had no substantial future as far as I was concerned. Frankly, I am amazed it made it to 2023.

After years of Protoss massing carriers and shield batteries, bad game design is finally taking its toll on the pro scene. One might even describe it as poetic justice. Who could’ve predicted that buffing the heck out of micro-less defensive mechanics, like batteries and siege tanks, would make the game so long and boring that nobody wants to play nor watch it? Oh wait, I did. :rofl:

Designing the game around efficiency mechanics was a huge mistake. Pro players now spend 40 minutes spamming snipe to get the peak value out of their ghosts. It used to be that the game was decided by making big strategic maneuvers, but they deleted all tactical maneuvers except efficiency plays and that makes the game really long and boring. Of course nobody wants to watch this game. Why on Earth would anyone play a 40 minute game of “dodge the snipes”?

1 Like

Man, BatZ managed to take a sad moment and still make it all about him somehow. SMH.


On the one hand, I’m sad to see GSL go. With blizzard pulling funding completely, I’m not surprised. There have been incredible games, brilliant moments and fun times that have been had.

But while GSL is seemingly on its last legs, ESL is still alive and still funding a not insubstantial ount of prize pool for the other regions, and looks to be putting in the effort to keep it that way.

For all our differences, the community does love this game, and despite the naysayers, SC2 is still unequaled in its gameplay, strategy and skill expression. It is unique and unrivaled, and had Activizard not shat the bed on multiple occasions, SC2 could still be one of the best competetive games to ever exist across any genre out there. As it is, it is still definitely and definitively the best RTS out there so fas as I am concerned - at least until stormgate comes along!

5 Likes

This is why I don’t read your posts. You are offended by the fact that I was right about how the hyper-emphasis on turtle strategies was going to ruin SC2 esports. You are so offended that you generalize a whole post by 1 sentence. Imagine being that mad. A man mentions one prediction he made and, bam, the entire thread is supposedly about him. Lmao.

The GSL is taking a huge funding cut, and you just can’t help but start flaming even when big news like this comes out. Ego is everything to you (clearly). Thanks for the reminder as to why I have you blocked. Would it hurt you to admit you were wrong every once and awhile? No it would not. Don’t bother replying because you are back on the blocked list.

1 Like

it still astounds me that Blizzard has simply given up the biggest RTS currently in existence. Other companies would pay to own Starcraft and Blizzard just abandons it because it doesn’t have as much turnover as their other IPs

3 Likes

let’s see how it develops.
Would be a pity now that the first foreigners have come to the Korean top lvl.
Maybe Maru now moves / lives in na, that he can play there in tournaments.

1 Like

Quite the opposite. Now is the perfect time for Batz to swing in and win a GSL like you’ve been promising. You could be the one to lead the korean Zergs.

:smiley: I still have a clip from 2016 playing with them for the good old times but was very good turtle breakers in TvT also very good (cheap) to use them against zerg lmao. This was an example of a series of such designs, Disruptor, mines, and TJ that followed.

I find it funny they even made them. Next they should have been flying tanks just with wings. Well campaign has such thing from Nova missions actually jump over cliffs.

GSL

But let’s not be like yellow newspaper reports, GSL still is there not removed so not that dramatic but obviously the interest and motivation will lower. Which is why events like IEM Katowice should be treated as global finals hype.

Cant say I didnt see it coming though. I stopped playing myself, sure Terran is better than 2018 but I still think ladder TvP games haven’t changed :roll_eyes:

So with that I am looking for Stormgate not sure if these graphics there are best but as long as they dont turn it into a milk machine and I do not mean just having cosmetics, well that game will be F2P and if they make the game as accessible as War3 is without such poor design (including the fast wiping and pace that pulls players away( it may be the next game to look forward to.

Either way I was also seeing it coming with all the Military Service, that killed the scene, yes it is needed life duty there but just saying, that when players go there, they return as lesser players basically Koreans after Military go FULL Foreigners skill level not more (see ForGG and Mvp streaming and will know what I mean) , that also for me was part of how the scene is not what it used to be.

or if it is how it should be, cant expect the same players to stay, again I think the game old plus how inaccessible it is to most players unlike League of Legends, made no fresh blood come and stay

Do you need to have a strict MOBA for it to be more popular? I dont think so look at war3, in fact some think war3 is a MOBA but it allows both strategy and not just hero control only.

To me the future of RTS is if the game that exists now like Stormgate is either like:

  1. War3 (balanced between hard to master, micro and intense) . SC2 was wipe everything fast pace not for every1 OR
  2. AOE4 - that style of games whether demons and tech that what Stormgate is about

I go for 1) because to be an ESPORT feels like working out - break your mind and hands to control all that you really do physical exercise too. Also war3 may have not been so to make Wrists pain like how SC2 did with pro Terran

Yea gonna be real sad when another round of kr pros retire.

I paid for GSL way back in the day when it was $20 a month, GSL means so much to me it’s sad to see Blizzard pull out I wish they wouldn’t.

1 Like

It’s never going to happen. SC2 is too simple and easy. They didn’t even try to make the game hard except for, maybe, the reaction speed aspect of the game. The game has zero strategical nor tactical depth. It’s all about reaction speed.

That’s why someone who is fast, like me, can get into GM with completely nonsense build orders. SC2 has a player base between 3-5 million. GM is a quasi international ranking due to cross-region play, so there are probably only 200 unique people in GM worldwide. It might be a little higher, maybe 300. The point is, if you can get into the top 300 out of 3 million, that’s a top 0.01% result. 99.99% of players are worse than you. I did that with random build orders. Strategy is totally irrelevant to the game. Tactics are totally irrelevant. As long as you make very simple decisions to pick very obvious answers, but can do this very fast, you will be good at the game.

That’s what killed the game, by the way. Strategically and tactically it is shallow. It’s about spamming APM. There is nothing interesting about watching Maru spam snipes for 40 minutes. That’s the polar opposite of interesting. Of course esports popularity was going to plummet. This isn’t an RTS game anymore, it’s a mechanics/multitasking game.

What makes a game interesting for me is depth. I am high in intelligence and openness, so I burn through hobbies really fast. I love trying new things and exploring and conquering new problems. Because I am smart, this process doesn’t take very long, and I burn through hobbies quickly. SC2 has been entirely figured out for at least 5 years, with maybe a 1-2% of the game to flesh out in some of the rare scenarios. That is not interesting nor entertaining at all. It’s literally a nexus of everything I hate. Why would I play a game that is simple, has no depth, is about spamming the same answers over and over again? The idea that I would go pro is pure lunacy.

The only strategic win that I can recall in recent memory came from Dark vs Clem. I think Dark is the only strategic player on the planet. He was losing the game, but then did a really clever maneuver that sectioned Clem’s army off from his main base using lurkers while lings went to town on the production. It reminded me of Innovation’s play where he’d be behind and yet still winning because he’d sacrifice things in order to get his opponent to take bad positions out on the map. It’s like dangling a free Rook in chess, only to get him to open up a weakness in his defense. That’s strategy. You can’t attack because attacking suicide, so you have to get them to attack you which means giving them a weakness to exploit. In exploiting your weakness, they create a weakness in that their future actions become predictable and exploitable. You make up for the loss in the following moves by superior strategic decisions. That’s how Dark won that game. You have to lose in order to win.

That kind of stuff is almost entirely absent from SC2.

:rofl:

I think I can understand your posts now. They must be satire and, if so, that’s actually really funny.

2 Likes

I am sorry for your ego, but there are no 3-5 millions of players. There are 141229 unique accounts, considering smurfs and multiple accounts ‹100k unique players.

1 Like

You are speaking of the 1v1 ladder. I am speaking of arcade, coop, commander, etc. These figures aren’t available to the public, but it’s been reported to be in the range of 3-5 million by people familiar with the matter. It’s been awhile since that figure was stated, but since the coop/arcade numbers are probably linearly correlated with the 1v1 ladder, it’s reasonable to estimate a proportional change. The 1v1 ladder activity has declined approximately 40% since then, which means the lower end of the 3-5 million is the more appropriate estimate since 5*0.6=3.

Furthermore, the standard deviation of MMR should be about 400 and the mean about 3000. 5000 mmr is where grandmaster starts at the moment, and 5000 is then a sigma (5000-3000)/400=5 outlier. A player >=5000 MMR would be the best player in roughly 3 million. So, the stratification of MMR values estimates roughly 3 million players as well.

Furthermore, the ideal distribution of MMR shows that accounts ranked >= 5000 occur too frequently, which is cut-and-dry proof of duplicate accounts / proof that the top 200 in each region is really just the top 200 worldwide with minor differences.

You have causation backwards. My conclusion has nothing to do with my ego. My conclusion is a product of the facts. What I find really interesting is that my SC2 performance is very similar to my standardized testing performance from school. That’s very intriguing. SC2 is probably very similar to a general cognitive test. The SC2 ladder is literally ranking cognitive performance on a bell curve. That’s been more or less understood for other games like chess for example. While I am just one data-point, it does beg the question of if video games do measure cognitive performance and if this correlates to real world performance and to what degree; it also begs the question of if video games can be used to increase your cognitive performance.

I know I learned a lot about strategic thinking, analytical thinking, abstract thinking, etc, as a product of learning to play SC2. Once you understand those elements, the game comes down to spamming APM and that’s extremely boring. But, one of the prevailing theories is that people who are predisposed to those modes of thought are drawn into arenas that stress those skills. So, even if you did a study and found that people who played the game more scored higher on a certain test, it wouldn’t prove that SC2 increases cognitive performance; it could simply be that people with the needed skills are more likely to stick around than those without. That theory is very plausible since without the skills for a particular specialization, it would be a frustrating experience to operate in that domain. If it’s frustrating, you’re more likely to drop out and seek success in another field.

In other words, the SC2 ladder measures performance against other people who play SC2, but people who play SC2 will be significantly better than people who don’t. The best player in 3 million is quite likely the best player in 30 million or more (if you were to select people at random from the general population and force them to play SC2). This also helps to explain some of the absurd MMR values like Serral’s 7k mmr which is a 10 sigma result. Serral’s is not the best player out of 100,000 – he’s the best player out of 7,888,000,000.

This means Elo ranking algorithms are more robust at estimating the distribution of the general population than would be expected. Elo measures a difference in skill, irrespective of how likely those differences in skill actually are. You could have bronze be 10 times more common on the ladder than in the general population, and the Elo rankings of those players would still converge on the proper values. That’s one of the brilliant attributes of ranking algorithms. If Serral is converging on a sigma-10, it means he’s probably the best player on the planet (even considering people who don’t play SC2).

1 Like

Batz is actually talking about the opposite of what the GSL has to offer and why it failed. People don’t want to watch their favourite players get blindsided by a build and get knocked out without even showing anything. There was so much hype for a player like Reynor to finally compete, and he was even thinking about playing Protoss but the tournament rules and format forbid him to do so. Imagine the draw that would have but instead he got knocked out 0:2 and subsequently a blow to viewership. This is also further exacerbated by only having a ro16 but that’s probably just a funding issue.

They were tinkering with ideas for example Code A being brought back, a roundrobin group stage 2 and a loser’s bracket in the playoffs but it was far too late. They also have a problem with time zones but that doesn’t have an easy fix.

The more I read your posts, the more sure I am that you are using ChatGTP to generate that nonsense. You take players from arcade and coop to calculate your level of skill for 1vs1? I would go even futher: you are better than 8 billions people on that planet, because most of them don’t even know that sc2 exists, but why not to include them too?

1 Like

That’s not an issue with the game, that’s an issue with the players. That’s a skill issue.

Reynor lost because he realized the harsh reality that Koreans are much better at SC2 (and video games in general, actually). Did you know the average IQ in Japan is 107? Taiwan is slightly lower at 106.7, South Korean is 103, China is 104.

There is a meme about developers adding a new difficulty mode to video games called “Asian mode” where they jack up the difficulty of the game to where it’s impossible to win:

I would actually support this, because I happen to agree that video games are way too easy. I played through Kingdom Rush on my phone. They have special items that you can buy with coins. I beat it without buying a single special item. There was one mission in particular that took about 50 tries to beat. I had to legit optimize a build order to beat it and, even with the best build order, there was an element of chance to the outcome that made a win rather rare. But, you only have to beat a level once so if you get past that level once you are good to go. Even with artificial difficulties that aren’t built into the game, such as restricting what options you can use, it was too easy.

There are things in life that I’ve done that take many more tries than 50 to get the answer right. How on Earth is someone supposed to tackle a challenging problem in real life, if a video game gives them issues? Video games cost nothing to try again. In the real world, trying again equals thousands of dollars down the drain.

I was working on designing something for a project about 2 years ago. It costs $20k to design, prototype, and test a single unit. You need probably 100 iterations to get a decent result because it’s a challenging design problem, which is the equivalent of 2 million. I solve that problem for about 50k. I designed simulation software to model it, and created a design that was paremetric meaning the program can alter parameters to change every facet of the design. An AI program ran 24/7 for 2 years on an RTX graphics card I had sitting in my basement. It did about 1,000 design iterations and drastically outperforms anything a human has ever designed. Soon, I will be prototyping it. It requires a 5 axis CNC capable of machining an aluminum billet block. Those are the kinds of problems that are fun to solve. “Do I start 1/1 upgrades before my third cc?” is not a hard problem in the slightest. SC2 is a very simple game, and Koreans are better at it than Europeans. If Reynor gets “blind sided” by a build, it’s because he sucks and he should probably go play some tic tac toe.

The frequency distribution of player rankings will approach a normal distribution with the same mean/standard deviation as the input elo parameters. That means the elo ranking algorithm is just a more advanced version of the bell curve that works under more extreme conditions. On a normal test, it’s you vs the test, but in SC2 it’s you vs the opponent. Ranking against other opponents gives more information than ranking against a static test. When you are pitted against a test, you are measured relative to the test,which is static. When you are pitted against an opponent, you are measured relative to the opponent, which is dynamic. The difficulty of the “test” varies with the difficulty of the opponent, and this encodes more information than a test with a static difficulty. This means more information is encoded in the match history between players than is available in the match history between players and an static test. This makes the elo algorithm much more robust at determining population parameters, and it works in conditions that other testing methods wouldn’t in.

If you standardized a test for, I dunno, reading comprehension, using people from the third grade, the results would only be valid for the third grade. But, if you standardized a elo ranking match history for the third grade, those ranking scores would be as valid as ones standardized against the entire schooling system.

Even though the SC2 ranking is standardized only on people who play SC2, it’s a valid estimation of the population parameters for all of humanity. If Serral is a sigma 10, he’s quite likely the best player on Earth.

ChatGPT is trained on internet data / digitized content such as ebooks. If this were generated by ChatGPT, it wouldn’t have a good understanding of how Elo works because nobody has done much research to figure out how it works. There is a lot of confusion about it even on the wikipedia article. Some of its properties aren’t even known. I know more about these systems than probably anyone else on the planet, and that’s evidenced by the fact that I can answer questions that aren’t available anywhere else.

I had a question at one point in time, and it wasn’t answered in the literature. Nobody cares about ranking algorithms in video games. I think they are more important than researchers realize because they are essentially an IQ test with more data points than any study ever done by researchers. That’s a treasure-trove of information about how humans process and solve problems. That’s interesting to me because I have a background in statistics & artificial intelligence. One of the questions I had is how the frequency distribution of elo rankings evolves by changing the population parameters. Nobody has answered that question before. You can google the heck out of that question and nobody can answer it except me. Do you think ChatGPT can answer it when even professors cannot? Give me a break.

I had to use a custom piece of software, which took me about 10 years to develop & test, and which isn’t available to anyone else on the planet, to solve that problem. Do you think ChatGPT can do that? :joy: :man_facepalming:

All this program really does in its basic form is find a set of equations which can map a set of inputs to a set of outputs. For example, I had some data about how various mixing ratios of concrete affects the strength of the concrete. Wouldn’t it be interesting to have an equation that can accurately predict the strength of concrete based on the ingredients? Why, yes, it would.

Another time, I sprayed a plastic beam of a certain size with speckled spray paint and then loaded it to varying degrees with a bending force. Programs can calculate the strain based on how the texture of the part deforms. I then ran this program on the strain data and it was able to figure out an equation that maps the original shape of an object to its deformed state given how it’s put under load. Do you know how long it took mankind to figure those equations out? We’re talking tensor calculus. This program did it in about an hour.

This program differs from mainstream AI programs like Deepmind by the fact that it outputs traditional mathematical equations. It doesn’t use nebulous neural networks. The outputs can be analyzed and manipulated using the normal tools you learned at school. That makes it infinitely more useful than neural networks, which are so nebulous that nobody knows how they work.

So, I developed that program, and one of the things I tested it on, along the way, was how Elo evolves based on population parameters. Do the elo rankings of players change based on the ranking or frequency of their opponents? That’s a very good question. Nobody has the answer, except me.

And what does your response has to do with the question: how do you manage to bring 3-5 millions of coop and arcade players, who play in totally different games in terms of gameplay, mechanics, skills in your calculation of how high is your rank in the world in 1vs1?

2 Likes

A) the mmr rankings are indicative of the population without including these players, and B) performance correlates with hours played; people with 0 hours played will be bad. For my assertion to be wrong, you’d have to posit that people playing the arcade are secretly Grandmaster in skill, despite not playing any 1v1, and that’s frankly absurd. Every single one of these players will score lower than GM in skill. GMs are going to be the best out of the whole population of players, and that includes people who play arcade and even people who don’t even play the game at all.

Do you really think a person who plays 2 arcade games a month is the next Serral, or that some random kid in from Somalia is about to win WCS even though he’s literally never used a PC before? Give me a break. Grandmaster is obviously the top 200, worldwide, out of the whole population of the entire Earth. There are enough players who play SC2 to get a decent estimation of the mean and variance of the population on Earth, and Serral’s ranking relative to that mean is so absurd it’s clear he’s #1 or maybe #2 (beside maru).

The major assumption here is that we ignore the density distribution within the sample of MMR under the assumption that a difference in MMR is the same no matter the actual density distribution. That definitely is a true because I’ve simulated a variety of population parameters and the difference in MMR remains the same. If you put 1000 bronze and 100 gold in the sample, it will be the same rankings as 1000 gold and 100 bronze. That means the z-score of the mmr value estimates the z-score of the population irrespective of factors like smurfing or activity/inactivity of certain skill brackets.

Mhm Koreans are better, WERE better because they had team houses and spent more time. After their services at the barracks you see Koreans going full foreigner, cant say they are better. The KR players that are still better are Dark, Maru and … I cant think of more names maybe Cure and Gumiho and Ragnarok, maybe Classic and the end. The rest are not any better

1 Like

If you break win-rates in the pro scene down by matchup, and then region, Koreans have a >65% win-rate against foreigners in all matchups, with a peak of 80% in one matchup. Koreans are definitely better.

Reynor is one of the best EU players. He went to compete on the home turf of the GSL long after the Korean scene peaked, and he was absolutely destroyed. Before the region lock, koreans won almost every major and premier tournament for several years straight. Blizzard had to ban Koreans from most tournaments using the region lock.

https://liquipedia.net/starcraft2/Premier_Tournaments/HotS

I spot 1 foreigner who won a premier tournament during HotS. One.