#1: Maru - Greatest Players of All Time - Page 67

On July 26 2024 18:50 Mizenhauer wrote:


There aren't enough people playing anymore to get a rating that high-at least that's the case on the KR server.

On July 26 2024 18:58 WombaT wrote:

Is potential high MMR reliant on a certain player pool size?

On July 26 2024 20:36 WombaT wrote:
I’d always assumed it was some relativistic measure and not one’s share of a finite amount of sweet, sweet MMR

I’d always assumed you’d have say two 7K players in different eras, but one would be better than the other if the playerbase declined. Not that one couldn’t physically get to 7K

Or as per your example, there’s always been a pretty chunky gap between the low end of GM and the middle/top (when you start to see it be mostly pros)

So my intuition is that it’s easier to sneak in at the bottom end just with player drift for sure, but I’d attributed that to there being fewer pure amateurs playing seriously, but the semi-pro/pro cohort remained at least somewhat stable, they’re the ones who have the incentive to grind.

Like just to pick entirely arbitrary numbers out of my arse let’s say 10 years ago you had 50 pro-adjacent players and like a thousand ‘GM capable’ amateurs. Not all necessarily active at the same time 100%, but floating around. Now let’s say today it’s 40 of the former, and like 300 of the latter. Yeah it’ll be easier to get into GM at a lower MMR, but IDK if it’s due to MMR itself deflating if you get me. Most pros are still way above 6K after all.

Interesting this should come up as Happy was recently embroiled in some WC3 controversy for leaving games on start. One of the points made was as his MMR was so high that he’d (in a limited sense) be seeing MMR actually leave the system. I.e he leaves versus player he thinks is an asshole, there’s a big gap, player gets a bunch of extra MMR. Then that player marches someone else high on ladder, but they’re not nearly as high as Happy, so the reward on offer for his opponent in terms of grabbing the MMR that was previously gifted by happy will be lower. Etc etc

The argument was that (in extremis) you’d see MMR leak

In combination figured I’d ask as I was curious how these systems actually work! You’d think I’d know by now haha

On July 26 2024 08:57 rwala wrote:


I just want to really nail down that Mvp is as low as 16th on your list and that it bothers you that Life cannot be a GOAT contender because in your view morality and greatness are completely unrelated concepts. I think you’re pretty smart and understand the perils of taking these positions, but I admire that you don’t seem to care!

If GOAT assessments were like contracts, perhaps you’d be on to something, but alas they are not. But even with contracts, there is often what we lawyers like to call “strategic ambiguity” in which the contract does not itself definitively resolve all matters but instead prescribes the process by which resolution happens (e.g. arbitration or litigation in X court under Y law). But even when you try to button everything down, the standards by which alleged violations are adjudicated are not so clear and leave plenty of room for discretion. Ask your girlfriend about the “totality of circumstances” or “preponderance of the evidence” standards and I’m sure she’ll smile because we lawyers joke that these standards are so subjective and flexible that you can basically argue whatever you want! Of course if these assessments were so objective and deterministic you could simply have algorithms render them (as you’re trying to do with the SCII GOAT). There would be no need for lawyers, judges, or juries.

Regarding your flat earth/electricity analogy, it’s a great point, though perhaps not for the reasons you think. These are scientific questions with pretty clear answers that can be determined with a high degree of certainty using the scientific method. I do not want anyone’s subjective conception of “common sense” involved in determining the answers to these questions. Especially given the consequences.

The issue you’re having is in primarily viewing questions of greatness as a scientific or mathematical exercise. I again go back to Muhammad Ali, the undisputed boxing GOAT, for many the greatest athlete of all time. His stats alone cannot justify this, because “greatness” both in our hearts and our minds is about so much more than the numerical value of what one achieved. It’s about how they achieved it, what they overcame to achieve it, the moments they achieved it, and what impact their achievements had. These are over time ultimately questions of legacy and even legend that transcend any understanding our silicon overlords could ever have.

There is good news for you as a Serral fan tho! I believe Serral over time will quite likely be regarded as SCII’s GOAT, at least to some degree of consensus. Not because of the math, but in spite of it. An algorithm that has Serral as #1 and Mvp as #16 will have no impact (and may even be counterproductive) in this regard. But his achievements and what they mean for the e-sport have the makings of legend for sure.

New post
I didn’t mean it as a jab, but an honest observation of what is happening here. From a purely “scientific” perspective you are buffing Serral’s stats while you claim to nerf them. If you really want to be objective, you need to discount the value of tournament wins by the average player rating of the competition (or some other method that captures competitiveness rather than nomenclature like “world championship”).

When you start factoring in “prestige” and prize pool, you become victim to the very subjectivity you seem to loathe. It is not reasonable to think that players try less hard in GSL. For many players, their career goal is to win a GSL.

I don’t disagree on your GSL “soft lock” point but my bigger point is that it is the only tournament that does not artificially limit the player pool.

You say that world championships were the highest priority for players but you ignored my observation on how the pro tour/WCS circuit was specifically designed to systematically give players like Serral more chances at a world championship than Korean players of a similar championship caliber. It was designed to buff Serral, and it did.

The problem is that it’s very hard to determine how much Serral was buffed because it’s possible that he may have qualified for all, some, or very few of the international tournaments he played in had he been subjected to the Korean regional allotment limits instead of the more generous European ones. uThermal off-the-top speculated that if this were the case maybe he would have won 10 tourneys by now instead of the 25 or whatever it is, but personally I think that’s too low and not really fair to Serral.

I don’t believe in such alternative history nerfing and personally I think Serral deserves full credit for all his tournament wins despite the buff. But maybe you don’t need to double-buff by also over-valuing those wins over other tournament wins that had more competitive player pools?

Again, I personally approach these things a bit more subjectively. But I’m just trying to hold you to your own standards here.

p.s. I don’t agree with lots of other things about your criteria, especially the Aligulac analysis. But as I’ve said before I can’t take it seriously until you actually reveal the results of your list and show where guys like Mvp end up.

On July 26 2024 15:13 Starcloud wrote:
Very good and respectful discussion on the last pages, thank you all




I do not agree on all points you make, but you write very well and have many good points in your posts.

However this one I quoted is 100% what I think about greatness too. But if we take this into consideration, isnt Serrals career by already have legendary elements which rise to him that status too ?

- Before he started winning tournaments, he was mentioned by lot of other pros by a "ladder monster", who would become something great one day. Created some inside hype before he rose to the top

- First player outside Korea to win multiple international tournaments.

- First player outside Korea to become World Champion
- Until that, many people thought that Koreans are kinda invincible and the best of them simply cant lose to foreingers in a big tournament. (f.e. Dark, who "never" lost to foreingers)
- At different interviews, people at that Blizzcon were buzzing about this some dude that finally won against the long regime of Korean dynasty
- Came from Finland, from little village in middle of nowhere, without any team-houses, coaches or practicing with other pro-players(Koreans) to the number one spot
- Basically started to carry the hope of foreingers to actually fight back and even possibility to win against Koreans. Created movement in all Starcraft community to watch what he does next and will he ever meet Maru in Premier Lan tournament. This narrative sadly kinda fell flat since it didnt happen for too long time.

- Practically all the other pros say that he is the GOAT

- Still, at 2024 is the man to beat after 6 years of the initial success.

- Maru vs. Serral GOAT debate has been going on for almost that 6 years too. And it seems never to stop either.

Etc. etc. In the end few other questions that imo should be given thought when thinking about this issue, that I keep important:

- How can Maru be the GOAT, if he cant beat his rivalry partner ?
- How can he be the GOAT if he never won the World Championship ?

Cheers.

On July 26 2024 21:17 MJG wrote:

It is relative, but if there are too few players above your MMR then you need to maintain a higher win rate against players below your MMR (as you'll start encountering them more often) in order to stay at the same MMR.

I'm speaking anecdotally again, but it used to be the case that I could maintain my ~5.6k MMR with a 50/50 win rate because I would play people above/below me fairly equally, whereas it now feels like I need to be winning 60/40 to maintain my ~5.2k MMR because I play a lot more people who are below my MMR than above my MMR.

On July 26 2024 21:20 PremoBeats wrote:

Where did you read that Mvp is as low as 16th? I said that I didn't rank the pre-contenders. So I simply do not know if he places 5th , 8th, 11th or 16th, but he is likely in the top ten.

Yes, I am aware of that, but my point, going into contracts, was that subjectivity can only take you so far. If an employment contract says you have to work 8 hours a day, you can argue about if the break is 20 or 40 minutes if it wasn't specified, but it surely won't net you a 4 hour break. The same is true for Mvp... you can put weighting on the metric that makes him GOAT in your opinion, but it becomes similarly obvious that the weighting becomes ever more unreasonable the more you have to emphasize the weighting.

Why do you think greatness is not something that can be - at least to a certain extent - measured by the scientific method?
You would need to...
1. define greatness ("Being the best at something.", "Doing things better than the rest.", "Achieving unheard of results", "Outstanding success")
2. define metrics that are able to give weight to your definition
3. define the ratio between metrics
4. measure the data that is relevant for the metrics
5. compare the gathered data

Point 3 is irrelevant if one player is the best at every metric as a different weighting wouldn't change the overall result. It would only affect the distance place 1 and 2 have in relation to each other as well as determining lower placements.

I've got no idea about boxing. No clue if there are statistically better boxers out there and if Ali would come out on top with era adjusted weightings. That is something I simply can say nothing about. I only collected data about StarCraft 2. And unless someone is making a consistent, rational case (which can include more immeasurable feats), I simply see no point as to revise my result.

Seriously, I have no idea if you read the article, as I did exactly what you asked me to do. I made average player ratings (this was the most work of the whole thing) for tournaments to measure their worth. Prize money (which isn't subjective) and subjective prestige were used too, but the biggest factor were tournament structure and average player rank.
All the subjectiveness in the article was negative to Serral. Are you even aware how little World Championship tournaments play a role in this scoring effort? Little to none as there haven't been a whole lot 1st or 2nd places by the final contenders and second that lessening the value from these doesn't change the end result by much. The impact, my decisions to lower Serral's tournaments even more than Miz did heavily outweigh a perceived favoring of Serral here. I even didn't do what Miz did: Devalue World Championships pre-2018 in relation to post-2018, which would have helped Serral.
So even if one can count my categorization of WCs as favoring Serral (which can be argued about), the impact is slim to none. Plus, the insane era-multiplier favors all other contenders except Serral a lot more than any WC-weighting would even be able to do could favor Serral.

If tournament structure is counted as artificial limitation, of course GSL is complicit in doing this, by simply stretching qualifiers to finals over several months, which artificially limits foreign participation (weekenders in contrast are much easier to take part in, although they can have other limitations). You will probably act as the defendant's attorney

I would disagree with the notion that it was designed to buff Serral. It was designed to buff every region outside of Korea, which Serral was part of. This of course was done to make other region grow and be able to develop. But the idea that any region would be able to compete in another region's qualifier for a world championship in and of itself is absurd. You wouldn't have Brasil play in the Asian qualifier in football, because it wasn't able to qualify through the American one. This was a phenomenon which helped Koreans beforehand immensely, although it made sense to have the best of the world compete, as these mostly hailed from Korea, but - as Miz pointed out - since 2018 has changed.

As I said further up above (even if we count the weighting of WC post 2018 as buffs to Serral): The impact was minimal and outmatched by many other decisions I made - but I can give a full list if you want me to do so, although I mostly pointed them out in the article.

And again: I never made a list of the pre-contenders nor did I analyze their careers in depth, because it already was months of work to go through all tournaments and placements of 4 players. The pre-screenings can be ranked but will utterly lack any relevant depth. My guess is that Mvp will be ranked in the top 10 as his tournament-, efficiency-, Aligulac rank 1- and HoF-score are among the best. But I cannot say for sure as I never went deeper into the pre-contenders when seeing, that 4 players massively stood out.

And what do you disagree with exactly? Perhaps the things aren't as they perceive them to be?

On July 26 2024 22:33 rwala wrote:


My friend, if you think determining the GOAT of anything is like reading a contract and determining if someone if someone took a 20 minute versus 4 hour break, we’re quite simply at conceptual odds that are so deep and profound that they cannot be reconciled.

What I am trying to do though is to encourage you to hold yourself to your own standards. You claim your calculations are objective, but admit they include subjective assessments like “prestige.” You say prize pool is “objective” but fail to demonstrate any statistically significant correlation between competitiveness of the tournament and its prize pool. The result is that you value in many cases more “prestigious”, big money tournaments higher than tournaments with more competitive player pools. Which is totally fine, I do too! But please stop pretending this is an objective assessment.

You say you have no idea about boxing and are only focused on SCII. But I really encourage you to broaden your horizon and take a look at other sports and games, think about the GOAT convos there, etc. Your football example is actually quite ironic, because Messi’s GOAT candidacy is possible almost exclusively because he is an Argentinian that was nonetheless permitted to play on a Spanish team and in European leagues. If he was required to stay and play in the Argentinian league, maybe he’d still be the GOAT, but it would be a harder case because he would not have results playing in the most competitive leagues and tournaments. Sound familiar? The World Cup, football’s “world championship,” is important to fans and players as a matter of national pride, but is not considered the pinnacle of competitive football because of regional allotments and because players spend all their time and energy playing for their club teams. World Cup results factor into GOAT convos, but much less so because most of the die hard fans and commentators understand that (European) club leagues and tournaments are where the best players and best teams duke it out.

Come to think of it, you don’t need to run the “Messi stayed in Argentina” hypo, since Pele is a good example of a really compelling GOAT contender who played in the (less competitive) Brazilian league for most of his career and derived most of his accolades from international completion. The thing is tho both the Argentinian and Brazilian domestic leagues are still among the best domestic leagues in the world even if some European leagues are more competitive. So there are not perfect analogies.

I could talk about chess and other sports and games, but again I strongly encourage you to dig in yourself because I think it’ll help give you some perspective. In the end it’ll help situate what you’re trying to do in a context that resonates more broadly. It’ll almost certainly challenge some of your assumptions, as new perspectives tend to do.

p.s. regarding GSL’s “soft” lock, while it has been practically difficult for players outside of Korea to compete in it, that hasn’t actually stopped many of the world’s top players from doing so. Special, Neeb, Idra, Naniwa, TLO, Stephanie, Scarlett, Reynor, Astrea, etc.

On July 26 2024 22:58 WombaT wrote:

This broadly isn’t true any more, amongst the best perhaps but so far beyond the top Euro leagues they aren’t really even in the same ballpark.

Whereas in Pele’s day that was far less the case, so one has to assess him through that lens. A modern GOAT candidate basically has to play in Europe, but one cant go backwards and apply it to Pele.

One could perhaps apply that same lens to the GSL in say 2014 to 2024

rwala wrote:

I totally agree! What I am taking issue with is this (silly, in my opinion) claim that these stats guys are “nerfing” Serral and he’s still clearly #1 by some objective assessment of the numbers. They’re doing the opposite (and Premo pretty much admitted that’s what they’re doing in his last response to me).

Because once you actual get into an “objective” assessment of the numbers, it’s super duper clear that 1) he has literally zero results in SCII’s most competitive era and tournaments; and 2) the pro circuit system was specifically designed to buff players like him (and did). uThermal (who thinks Serral is the GOAT) estimates Serral has about 2.5x the tournament wins than he would have had he competed in Korea this whole time (i.e. subjected to the nerfs Korean championship contenders were subjected to). I think that’s not really fair or even a super useful hypothetical exercise. But at least it’s an honest way to think about the reality of the situation. It’s much more objective than anything you’re reading here, and it’s from a Serral fan who also actually understands from first-hand experience how much more competitive the scene was in South Korea.

Basically my only thing here is that if you’re claiming to be objective, then do your best to be objective. Don’t say you’re taking some neutral numbers-based approach that you say is designed to “nerf” Serral and then give extra points to tournaments with less competitive player pools that Serral happened to have won.

rwala wrote:

My friend, if you think determining the GOAT of anything is like reading a contract and determining if someone if someone took a 20 minute versus 4 hour break, we’re quite simply at conceptual odds that are so deep and profound that they cannot be reconciled.

What I am trying to do though is to encourage you to hold yourself to your own standards. You claim your calculations are objective, but admit they include subjective assessments like “prestige.” You say prize pool is “objective” but fail to demonstrate any statistically significant correlation between competitiveness of the tournament and its prize pool. The result is that you value in many cases more “prestigious”, big money tournaments higher than tournaments with more competitive player pools. Which is totally fine, I do too! But please stop pretending this is an objective assessment.

You say you have no idea about boxing and are only focused on SCII. But I really encourage you to broaden your horizon and take a look at other sports and games, think about the GOAT convos there, etc. Your football example is actually quite ironic, because Messi’s GOAT candidacy is possible almost exclusively because he is an Argentinian that was nonetheless permitted to play on a Spanish team and in European leagues. If he was required to stay and play in the Argentinian league, maybe he’d still be the GOAT, but it would be a harder case because he would not have results playing in the most competitive leagues and tournaments. Sound familiar? The World Cup, football’s “world championship,” is important to fans and players as a matter of national pride, but is not considered the pinnacle of competitive football because of regional allotments and because players spend all their time and energy playing for their club teams. World Cup results factor into GOAT convos, but much less so because most of the die hard fans and commentators understand that (European) club leagues and tournaments are where the best players and best teams duke it out.

Come to think of it, you don’t need to run the “Messi stayed in Argentina” hypo, since Pele is a good example of a really compelling GOAT contender who played in the (less competitive) Brazilian league for most of his career and derived most of his accolades from international completion. The thing is tho both the Argentinian and Brazilian domestic leagues are still among the best domestic leagues in the world even if some European leagues are more competitive. So there are not perfect analogies.

I could talk about chess and other sports and games, but again I strongly encourage you to dig in yourself because I think it’ll help give you some perspective. In the end it’ll help situate what you’re trying to do in a context that resonates more broadly. It’ll almost certainly challenge some of your assumptions, as new perspectives tend to do.

p.s. regarding GSL’s “soft” lock, while it has been practically difficult for players outside of Korea to compete in it, that hasn’t actually stopped many of the world’s top players from doing so. Special, Neeb, Idra, Naniwa, TLO, Stephanie, Scarlett, Reynor, Astrea, etc.

On July 26 2024 23:47 PremoBeats wrote:


I can only say it again: I was nerfing Serral all the way.
The following nerfs/buffs have been applied:
1. Serral would have gotten a match win rate buff as the Koreans also play lower tier Koreans and qualifiers and tournaments. Serral only plays the top of the top which inflates the Korean's scores. There were inflation corrections of up to 3,96% that were not used for Serral
2. The idea to only look at match win rates versus Koreans makes sense, but is there to mostly nerf Serral (for obvious reasons, but a nerf nevertheless).
3. Looking only at prime years in this regard would have also been something to make mostly Serral's (And Maru a little bit less) success be more apparent as he simply crushed it post-2018.
4. For the tournament-win-rate I only counted Serral's tournaments with top Korean participation; again a major nerf, which made sense for the comparison, but still is a nerf, nevertheless.
5. The 50% buff for pre-2018 is of course a nerf to Serral.
6. Average place achieved was also a nerf to Serral (and Rogue) as I counted only prime years which would have made INno's and Maru's results much, much worse
7. Also a 50% buff for pre-2018.
8.Tournament score: In relation to Miz I penalized Serral's ESL Masters and DH more and made a separate category for HomeStoryCups which also meant mostly a penalty for Serral. The upward-corrections for WESG helped Maru and INno, at the same time penalizing all other players.
9. Another 50% for pre-2018.
10. Efficiency-score: As a dividend of tournament score this was also affected by the 50%-buff for pre-2018 which only was a penalty to Serral.

And I can only say it again: In most areas ridiculous buffs would have been required to make INnoVation get on par with Serral (buff of 300% for pre-2018 in tournament-score for example). That is simply not fair.

Funny that you highlight the subjective guess of uThermal and not the hard facts he lists: "If you compare Serral to Maru and Rogue in particular, I think it's definitely pretty clear. If you look at the most important metrics you can think of, Serral wins them all. Most tournaments won, most prize money, highest win rate, most dominant. Even stuff like Aligulac - I know Aligulac is not something to go of - but he's like 400 rating above Maru, I think it is an insane difference you know."
When talking about competitiveness he says that the players we see currently are the best players we have ever seen but there is a difference if you compete with 100 or 10 guys. How I heard it is, that he continued to say that Serral - if you put him in 2015 the way he is now - he only wins 10 and not 26 tournaments. Which means to put 10 wins of Serral in today's era, you would need to multiply 10 with 2,6. But Innovation would need a multiplier of over 3 to get on par with Serral within my setup (which I wrote either here or the other thread before).
And I don't know how much uThermal thought about this number before putting it out, because in my opinion it doesn't make sense for Serral to only accumulate as many tournaments as INnoVation in 2015 when INno had 14 participations from 2018 - 2019 which netted 1 win, while Serral had 12 participations and 5 wins (counting only tournaments with top Korean participation). Now to anyone saying that INno was way worse at that time is simply being delusional, as his win rates versus elite players and overall win rates only marginally dropped from 2013-2015 to 2018-2019, he still was top 5 and sometimes even rank 2 in that time frame and as uThermal points out, players mostly get better. The competition got less, but a lot better.

I never said my numbers are neutral. I pointed out several times in the article and this thread how I advantage or disadvantage players. What you are saying is simply false.



I can only say it again: This 1 of 7 metrics includes a tiny part that is subjective, which was not even the most important factor of that metric, average player rank and structure of tournament was (both objective). So can you acknowledge this or not? Because I am getting tired of explaining the same thing over and over.

And all of the talk about football to simply ignore the way I described how greatness can be measured in StarCraft 2?

Regarding soft lock. Ok, so 9 or perhaps 15 out of... how many? In how many GSL tournaments over the years? Come on...

"To that end I do need to say that if what you are really going for is something scientific, there needs to be a lot more work because the number of potentially confounding variables you’ll need to control for is…staggering. Proper statistical modeling and regression analysis would 100% be needed. I think that’s overkill and unnecessary, but just want to note some methodological requirements that would need to be met if you want to credibly call what you’re doing “scientific.” - rwala

On July 27 2024 01:00 UnLarva wrote:


You're right. Exactly. It looks to me that 3-body problem would be piece of cake to solve compared to the enormously complex StarCraft 2 phenomenon. However, you also very well realize that in the scientific process - particularly when operating with insane complexity - one must start from somewhere and get a overall picture on the matter of the study. As far as I can recon Premobeat's work this far is the deepest and most serious effort ever with this topic.

You know, everyone can nearly always use 'the god of holes' qualitative argumentation when ever things have no measurable quanta, or that has not been yet succesfully done even if it is theoretically possible.

Guys, take it easy. It's just a game. Entertainment.

On July 27 2024 01:10 rwala wrote:


For me the place to start is: is this a scientific question?

However one answers that and putting aside differences in perspective, Premo’s work is certainly impressive. They may not believe me, but considering and factoring in some other ways of thinking will only make their work more powerful and credible.