Four years ago, in another lifetime, I made this thread that analyzed the effect of matchup winrates on SL victories, and further on how those winrates tended to lead to bonjwas being mostly Terran players. Overall, besides the general controversy that comes with such a discussion, I think there were a lot of interesting points made and as a homage to the old BW pro scene, I'd like to make a sequel that better addresses the concerns a few veteran users brought up about my methods and conclusions. In the interest of trying to make a much fairer and more thorough analysis, I'd like to do some research on this topic. In that light, I'd like the help of the BW vets among you to offer me a few pointers on various historical BW topics.
General premise for the sequel: It's not controversial to say that Terran has been the historically dominant race in BW. Indeed, Terran enjoys the most favorable non-mirror winrates, the most MSL AND OSL titles, and the most bonjwas by far. Is this a result of the innovation of a few players, and something that could have gone a whole different way if, say, Boxer were a Protoss genius rather than a Terran one? Or are players mostly a product of their environment, including their race and the maps? To answer this, we'll try to recreate (through simulation) the environment that BW has had in the past, and we'll see if under similar circumstances, we can replicate the results of BW's professional history.
What I'd like from you folks: If you could give an answer to one or two of the following questions based on your own BW knowledge, that would be a great help in setting up the process. 1. What is the format of the OSL and MSL prelims? How many players, how many advancing players, seeding, etc? I have some trouble finding a very detailed study of this. I have what I need for this. 2. Broadly, what are the eras of BW by date range? If you feel comfortable enough giving specific cutoffs, month/year cutoffs would do a lot of good. Also, are there any major innovations you would say are associated with those era cutoffs? 3. For anyone who works on TLPD or knows someone who does: is there any way for me to get the Korean BW TLPD data that doesn't involve the manual way? If not I can definitely get it manually, but it would be a huge help if I could just get the data set. 4. Have there been any major changes in rules of the MSL and OSL that would significantly change the results of the tournaments? I have what I need for this. 5. What would you say is a fair way to account for the variance in "coin-flippiness" of the different mirror matchups? They're obviously not all 50-50 in reality, since some have a stronger coinflip element to them than others. Never mind, I got this. 6. Historically, would you say that there's a systematic difference between results in SLs and results in Proleague? Or are most players generally good at either both or neither? I know there are some "SL beasts" and "PL beasts" but I'm not sure if those are exceptions or the norm. 7. Are there any other points about the general premise, or the previous discussion, that haven't been made in the previous thread, that you would care to bring up?
a very ambitious Journey, i wish you best of luck and i like your idea!
The first thing that comes to my mind is this blog i read recently, a discussion on the status/terminology bonjwa and an attempt to calculate the success of the players in question by assigning points for OSL /MSL/PL but also long forgotten tours like KPGA Tours.
The op also tries to define the different era's of the game (in a sound way, i think)
so, check it out, it might give you some orientation.
As for #3: I have TLPD access rights and there is no way to dump data easily. You may want to PM Nimdil about that, I think he had some related project at some point.
6. is hard to answer. Teams placed different values on PL and SL and thus some players had a coach preparing them extensively and some didn't have that. PL changed format from known opponents to unknown opponents and what followed were players suffering and players flourishing. Also, you can't forget 2v2 which is another big point which has to be considered.
Eras of BW...wow that is a very tough question to answer.
2004-2006: Boxer, GARIMTO, [nc]Yellow, Nal_ra, JulyZerg, GoRush, Oov, and Nada were the first pioneers of today's matchups. Aside from Boxer's lasting impression on Terran micro, I think July deserves a lot of credit for discovering how to stack mutalisks easily and effectively with the overlord trick. However, I would arguably say that Oov, Nada, July, and GoRush (basically the second generation of upcoming players) changed the game from a micro-oriented game to a more macro. map-control management oriented game wrong dates
2006-2008 was a very intense stretch, and probably where the modern game of BW was defined and recreated. This is when Savior started perfecting the 3hatch mutalisk and deflier control late game and basically was running over all the formally dominant Terrans like Nada, Oov, and Midas. I think the highlight of this type of play would be between GGplay and Iris[gm] in the 2007 OSL finals. At the same time, this was also when Bisu created the DT/sair build and 3-0'd savior, bringing the protoss resurrection. And then at the same time as that, Jaedong took the reigns and found that DT/Sair isn't invincible as long as you used scourge and simcity properly. And then in 2008, Flash showed off his 2-2 160/200 push that beat all the fast carrier and arbiter usages on the huge 4 player maps. This was basically the highest point of BW
2008 onwards is where the game started becoming more...two sided. Jaedong and Flash were basically gods from here on out, facing little competiton from anyone else other than fantasy, bisu, and best to name a few. I believe the only other innovation from here on out was the mech transition in TvZ on 2 and 3 player maps, where the terran would fight a war of attrition until there were no more minerals on the map.
Sorry I can't be more specific...i don't know the names or the exact dates of a lot of these games.
endy: Thanks, I sent him a message. We'll see if it's possible to get that data.
Lucumo: That's definitely a good point. On the one hand, the players are playing the same game on the same maps. On the other hand, the changing of format structures is something that could fundamentally alter the winrates. I hope not, since it would really complicate things if I couldn't use PL data. And yet I'm not altogether convinced that you can really say that they're not equivalent...
imBLIND: Great high-level outline, and I think that I can get a lot out of the data by those cutoffs. Do you have any thoughts on the 01-04 era, or on the seasonal trends within eras (e.g. Six Dragons, Swarm Season, etc.)? Admittedly my own knowledge of the older pro-scene is decidedly second-hand since I only got here in 2010.
On May 20 2016 14:17 LegalLord wrote: imBLIND: Great high-level outline, and I think that I can get a lot out of the data by those cutoffs. Do you have any thoughts on the 01-04 era, or on the seasonal trends within eras (e.g. Six Dragons, Swarm Season, etc.)? Admittedly my own knowledge of the older pro-scene is decidedly second-hand since I only got here in 2010.
My dates are pretty messed up according to Liquipedia....the correct eras would probably be 00-02, 02-04, 04-06, 06-08, and then 08 onwards. There are so many players and so many memorable games, and it's getting way too confusing for me to list them all out by dates. I mistakenly lumped GARIMTO into the 2004+ era when he actually in the 2001-2004 era...brain fart there since GARIMTO is basically the godfather protoss.
I think it would be best if I tried to separate the eras out as best I can...
2000-2001. Boxer, GARMITO, and [nc]Yellow were the first of their respective races to show good results. Micro here was king, as most battles rarely consisted of more than 2 full control groups of units. Lots of dropship and shuttle play. Don't remember much about zergs and protoss (i was a terran player), but I believe muta/ling/lurker armies were used to secure a 3rd gas, rather than 3hatch muta harass play. Terran used to be thought of as the weakest of the 3 races, and the Boxer came in and smashed everything. His TvZ was absolutely entertaining to watch, but his TvP was noticeably weaker. I think he has a losing record against GARIMTO and Reach, two the most strongest Protoss players at the time.
2001-2004. Terran reign of terror. Only a few top-notch players like Yellow, Reach, Nal_ra, and GoRush, to name a few, could stand up to the massive amount of new and old terrans. Names of note: Nada, Oov, Boxer, Midas, Xellos. Macro and timing management started becoming more important than micro.
2004-2006....i actually got this mixed up with the 2000-2001 era. I think Midas and Xellos, OOv inspired a new wave of terrans peaking in 2006. Lots of terrans flooded the pro BW scene like Sea, Light, Hwasin, fOrGG, firebathero, Upmagic, and Iris, to name a few. Anytime was the only notable protoss on the scene, as he was one of the few to beat both OOv and Boxer.
I think this is the era where timing started to become extremely important. Personally, I think the terran reign of terror was mostly due to Terrans having the first timing attack in all matchups (the "FD" attack in TvP, the 2rax academy pressure in TvZ, TvT was a rock/paper/scissors kind of thing with goliaths>wraiths>tanks), which allowed them very good control over the flow of the game. Zergs still have a solid all-around game involving a better macro system (hatcheries versus barracks/starport/factory), but needed to maintain an economic control in order to do so. Basically, the terrans were stuck on 2base for a really long time denying the zerg a 3rd/4th gas, while the zerg was trying to secure the 3rd by constantly repelling the marine/medic raiding parties. I don't know much about the PvZ matchup as I didn't watch many of these games, but I know that lurker contains and securing the natural for protoss was a huge pain in the ass because of the fraility of observers and speedlings coming into play before speed zealots. TvP was skewed towards Terran I believe because 3rds were difficult to defend and a lot of naturals were very open and difficult to defend against vultures. Ride of Valkyries was probably one of the few maps this advantage was negated.
2006 -- Julyzerg vs Hwasin on Rush Hour 3. First documented game using the stacking of mutalisks. The beginning of the Swarm season in my opinion (peaked around 2008-09 until Flash figured out how to beat it with 1raxFE). Mutalisk control started separating the good zergs from the bad ones. Eventually, it becomes a necessary skill to survive at the top level. Terrans had to adapt to this new way of fighting, and they weren't very good at it since the 2rax academy expansion timing lined up nearly perfectly with the 3hatch mutalisk pressure timing.
2007 -- Bisu vs Savior. All the information you need here. The beginning of the Protoss resurrection. The biggest part of the PvZ matchup leading til now was securing the natural expansion while denying the zerg a 3rd expansion. Using a forge fast expand got rid of the securing the natural problem, but the zerg was free to outexpand the protoss. DT/sair punished overly greedy zergs, and balanced out the PvZ matchup. Terrans also started toning back their early game aggression against zergs with a 2rax expansion, followed by a 2base timing attack rather than a 2rax academy timing attack...i believe larger maps also had something to do with this. I think Iris vs GGplay was probably one of the most entertaining matches I've watched. The season of the swarm, was basically a period in time where terrans couldn't figure out how to deal with the mutalisk harass, and then they had no money/ways to beat defliers late game.
2008-2010 -- The famous 2-2 160/200 3base Flash build was probably the first 3 base timing attack that aimed for the arbiter/carrier/4th base timing rather than the 3rd nexus timing. Zergs also got overly confident in their mutalisk micro in the ZvT matchup (started cutting sunkens, faster 3rd), which probably led to the downfall of the swarm season. Flash showed terrans how to deal with 2hatch and 3hatch mutalisk harass efficiently using a 1rax FE and punishing greedy zergs. He was able to get a decent marine/medic army outside the zerg natural and getting an expansion up before mutalisks came out, which means that he retains most of the advantages of the originally standard 2rax academy in 2004 without the problem of securing the natural expansion. Fantasy came up with a different answer to this, using a valkyrie inspired build order. TvP became a problem for the Terrans, mostly due to the maps (in my opinion) screwing with the FD and 2base timing attacks that were traditionally aimed at the Protoss expansions. The protosses learned from the Zergs and started mixing in reaver drops/dt drops and a variety of strategies that screwed with the Terran's 2base timing attack aimed at their expansions...in short, Terrans were mostly 1 trick ponies that relied on a midgame advantage to win lategame versus protoss, and none of them were good at beating Protoss by pure macro. As soon as the protoss changed the timing, the terrans started losing. Flash was the only Terran that could stand up to them, due to his superior game sense, macro, and scouting abilities.
2010 onwards was basically just Jaedong and Flash exchanging blows, with Stork, Bisu, and Fantasy occasionally making the headlines. TvZ also got kind of stagnant since Terrans discovered that meching in TvZ is possible late game with a mech transition/floating raxes + shitton of tanks.
So long story short, Terrans had a better early/mid game (1-2 bases), and protoss/zerg had comparable mid game (3+ bases), but better late game. And then Zergs figured out how to get to mid game quicker, shortly followed by terrans, then shortly followed by protoss. And then Protosses realized their late game is just their mid game + more gateways, so if they can stall their opponents mid game, then they win. Terrans couldn't figure out why they were losing, but then Flash comes along to save the day, except all the other Terrans are still not as good as him. In the meantime, Zerg figures out how to skip mid game and go straight to late game versus Terrans, and they figure out that the forge FE can be countered by a 3hatch muta into a massive midgame.
.....yeah I got bored and started watching some late night vods >.>
Oh wow, that's really helpful and definitely a lot to work with. If I had to notice a trend, it would be that Terran seems to have the largest metagame advantage throughout the years, concentrated mostly in the early years but certainly extending fairly far up to the modern era. Protoss get a pretty short-lived reign of greatness, and the matchups don't tend to favor P. Also seems like TvZ is at the core of what defines the meta at any given time. My hypothesis that I've been working with is that Terran seems to rise to the top of the world so often because of three main factors: 1. TvZ is historically Terran favored and a fairly strong race-based advantage. 2. PvT is historically the weakest race-based advantage, and ZvP is historically the strongest, so Terrans get to minimize their weakness. This is corroborated by the historical dearth of Protoss qualifying for individual leagues. 3. TvT is the most skill-based mirror, which tends to lead to the very best players of T being the ones who win games and eventually titles, aggregating those achievements into individuals who are more likely to reach the "bonjwa" bar of achievement. That might explain why it is Flash and Flash alone who leads the field and gets the lion's share of the Terran individual titles in the modern era.
My final question would be, do you have any thoughts on the metagame evolution of the mirror matchups? That meta seems to be somewhat more stale, with pretty consistent historical trends. No one really focuses too much on talking about those and I wonder if that has a lot to do with the fact that historically they mostly follow a consistent pattern. TvT is the only one I'm quite poorly familiar with, since few people play it and even fewer really care to talk about it.
On May 21 2016 00:16 LegalLord wrote: Oh wow, that's really helpful and definitely a lot to work with. If I had to notice a trend, it would be that Terran seems to have the largest metagame advantage throughout the years, concentrated mostly in the early years but certainly extending fairly far up to the modern era. Protoss get a pretty short-lived reign of greatness, and the matchups don't tend to favor P. Also seems like TvZ is at the core of what defines the meta at any given time. My hypothesis that I've been working with is that Terran seems to rise to the top of the world so often because of three main factors: 1. TvZ is historically Terran favored and a fairly strong race-based advantage. 2. PvT is historically the weakest race-based advantage, and ZvP is historically the strongest, so Terrans get to minimize their weakness. This is corroborated by the historical dearth of Protoss qualifying for individual leagues. 3. TvT is the most skill-based mirror, which tends to lead to the very best players of T being the ones who win games and eventually titles, aggregating those achievements into individuals who are more likely to reach the "bonjwa" bar of achievement. That might explain why it is Flash and Flash alone who leads the field and gets the lion's share of the Terran individual titles in the modern era.
My final question would be, do you have any thoughts on the metagame evolution of the mirror matchups? That meta seems to be somewhat more stale, with pretty consistent historical trends. No one really focuses too much on talking about those and I wonder if that has a lot to do with the fact that historically they mostly follow a consistent pattern. TvT is the only one I'm quite poorly familiar with, since few people play it and even fewer really care to talk about it.
My mirror matchup knowledge of PvP and ZvZ are not very good, as I was predominately a Terran player. As far as I know, Mutalisk and scourge control was 90% of the ZvZ matchup, and there was a very pronounced build order rock/paper/scissor with 9 pool > 12 hatch > 12 pool > 9pool. PvP was something similar, but it was more of a matter of speedlots/templar tech vs reavers I think.
TvT I can actually help you with though. I did a lot of research on this matchup when I was playing BW, and I drew a lot of inspiration from a 2007 MSL game between firebathero and iloveOOv. This was the during the whole Fast Expand meta, where both sides would take their natural expansion before their 2nd factory. Probably one of the most entertaining TvT's that I can remember (personally, i don't like klazart's commentary, but its the only one left). Long story short, OOv proved that spider mines could be used to secure a fast 3rd expansion, even when faster dropships would probably be the more logical choice.
The meta before the 2006/2007 fast expand trend was to go 2fac goliath drop, 1fac1port tank/wraith push, or 2port wraiths, and then to expand. Kinda like how the ZvZ rock paper scissors BO went, but this was not as pronounced (for instance, going 1fac1port allowed you to get cloak faster, whereas 2port wraiths forced you to use scans as a counter to cloak since the gas went into the 2nd starport, not cloak research). 1facFE was considered weak since it put you in a more passive position and you could just get contained and it sucks breaking out of contains. The timing revolution and larger maps allowed for 1facFE to adapt quicker to counter any sort of attack before they could set up a contain.
On top of that, large maps with a lot of poor vision (hills) made mass vultures and spidermines a more appealing choice, since they can zoom around the map much quicker than dropships/tanks/goliaths. Eventually, all late game compositions involved a tech switch, whether to mass tank/gol/dropships to get over the mines or a battlecruiser techswitch to break through tank lines, or mass wraiths to just overrun everything. To get there though, you needed the money and time to do it, so I wouldn't consider this a meta per say. I think the idea of expanding behind the minimum amount of units and maximum amount of information is what OOv pioneered and was what Flash perfected.
I personally think FantaSy is a better TvT player, but Fantasy's TvZ was pretty weak so the two of them never had to face off that often. Flash was just an all-around macro powerhouse, while Fantasy's multitasking and decision making allowed him to use his units to the fullest extent.
I mostly agree with your hypothesis, but I would urge caution when you say historically. A lot of the matchup imbalances are only prevalent prior to 2008-2009 when the game was still being refined.
I think it would be better if you focused your hypothesis In terms of micro, macro, and game management (timing, strategic decisions, unit placement) rather than in terms of matchups. The balance between those 3 pillars was strongly skewed towards micro in the early years, which is probably why Terrans had such a great advantage (all ranged units, everything has a spell, everything has to be used properly). Once micromangement started evening out and macro started being developed, Zergs would start taking all the wins (Savior, Jaedong, Calm, Effort). And then when micro and macro started evening out and game management was being developed, we see that Protosses have advantages (each build order has different timings they are weak at, and protoss having the most amount of tech choices, they could mix up their BO's every game). And then when all three come together at the end, it comes down to who is the better player. Why is it Flash and Jaedong, but no Bisu? I'd prolly venture a guess and say observers are the weakest form of scouting amongst the 3 races.
You know, we could talk about metagame balance and how it's changed in history in qualitative terms (when micro was more important, when strategy really took over) for another decade, and I don't think we'd get anywhere. It's just one of those things that we can't really know for certain, and that is to a large degree a matter of interpretation. It's all a very gradual process, and even if we all watched every VOD a few times over (if they still all exist, which is doubtful) I think we'd still have a pretty hard time really understanding when everything fell into place. That's probably why discussions on the effect of balance tend to go nowhere and why they generally aren't taken too well.
And yet I think it's altogether too simplistic to say that it's just the meta that made things the way they are. Players are, ultimately, a product of the race they play, the maps they play on, and the environment they play in (tournament formats, schedules, life in general). The data also does show that Terran generally does have matchups that are favorable for them in the long run, and it is pretty clear that Terran is the most accomplished race over BW's history. Yet in the modern post-2008 era, it seems that T and Z are in a virtual deadlock in terms of titles, with Protoss just being sidelined like usual. The differences between the OSL and MSL, and the fact that Z seems to have the most successful mid-range players (most players below the top seem to be Zerg), is a real head scratcher. I'm not sure what to make of it and I'd probably have to look into it in more depth to really get useful results. I think this is shaping up to be a multi-parter.
There's also another interesting paradox with Bisu, Jaedong, and Flash. Overall, it's pretty clear that Bisu is less accomplished than Jaedong or Flash in the modern era. And yet he's pretty much universally acknowledged to be the best Protoss player in BW history. In terms of achievements, Nal_rA and Jangbi are the only ones who have enough titles, and Stork (and I guess you could make an argument for Reach and Nal_rA, but it's a longshot) is the only one who is close enough on skill level, to even be in the position to contest that dominance. Jangbi was decidedly inconsistent throughout his years and just managed to put together two titles at the very end, so I'd say he's out. Stork is hard to argue for since he only has a single title, so he's also out. Nal_rA is probably the closest, but Bisu had a more dominant run and has more titles to his name, and I think it would be a hard sell to say that Bisu isn't the better player. For Jaedong, he has Savior to contend with and that's a tough one to evaluate (personally I can't really choose between the two). Flash, he has NaDa and iloveoov as pretty solid challengers for the title of best Terran. Curious, to be sure.
Yeah trying to figure out why Terran is the most successful is mostly a matter of opinion...maps, talent pool, the race itself, take your pick. What I do know for certain though, is that each build order that the players used was refined and perfected to the second. How the game played out afterwards was a matter of game sense and skill.
And even though discussions on balance never go anywhere, I still enjoy talking about talking about them. It brings me some sort of closure knowing that I could be right =p
From my point of view, Terran is the most successful because of two(three) things:
- most of their units are ranged (gives them a natural advantage) - they have scan which, when properly used, is way better than an overlord or an observer (good scans are of vital importance -> they can win you a game, just as a bad scan can lose you the game) - the most important factor in my opinion (- they have repair (the most useful in the early game))
On May 21 2016 20:04 Lucumo wrote: From my point of view, Terran is the most successful because of two(three) things:
- most of their units are ranged (gives them a natural advantage) - they have scan which, when properly used, is way better than an overlord or an observer (good scans are of vital importance -> they can win you a game, just as a bad scan can lose you the game) - the most important factor in my opinion (- they have repair (the most useful in the early game))
I think the things you've mentioned are true in the sense that if you generate a random map with just a few parameters, terran is most likely to win. Need to take into account that the maps decide the balance a lot in brood war. Anyone can go into the campaign editor and make a map that favours any of the three races in a few minutes, and that race will be 'imbalanced'. Imbalance is dependent on multiple variables, and the 'map' variable, possibly the most significant of all, kept changing over time.
Maps always go back and forth in balance, and sometimes a single map can be so ridiculously favorable for one race that it makes a mockery of the notion of general racial balance. Yet I'm certain that there is something internal to each race that predicts a tendency on the aggregate for balance. At least for T>Z>P>T, that much isn't really controversial and it's pretty well-acknowledged. The question is, is data about winrates over time (which does, in a way, aggregate balance concerns on all maps), and about tournament performance, predictive of the tendencies that we've seen in BW history, which is decidedly Terran-favored and Protoss-disfavored? I suppose we'll find out.
The thing is, BW balance is treated too sacredly. I personally I see it like chess. Terran is white and Zerg/Protoss is black. Playing with white pieces has an inherent advantage and is really difficult to play against when in the highest levels of the game but is not unbeatable.
PS. You don't smell of bisu fandom, are you sure you're Lightwhip?
I think it's fair to say that if someone played white every game, they would be ridiculously OP. IIRC the spread is something like 55-45 but that's a pretty solid advantage for white. I was considering making that analogy last time around, but I felt it would be too provocative, since chess doesn't have map imbalance to deal with.
You know, the modern post-Kespa scene doesn't really appeal to me the way the progaming scene did. It's a shell of what I liked about the old scene, and even the players are not really at their best without the old structured environment of progamer houses. A lot of that old enthusiasm does fade when the game environment is not that interesting.
EPT
