Recently, the need to figure out the defining characteristics of various players has been on my mind. What were the strengths that allowed certain players to excel, and under which circumstances did these players best thrive?
One variable that is relatively easy to examine are the maps. What were the maps that brought out the best in legendary players? What kind of characteristics did these maps have? If a player was the best performing individual in multiple maps, did those maps have anything in common?
I decided to see if the maps that were the favourite battlegrounds of Taek-Beng-LeeSsang had any common characteristics, and if those characteristics said anything about the player. I had several rules to follow:
1) The map in question needed to have at least one hundred games played on it, otherwise the sample size would be too small for me.
2) The only maps that would be examined would be the maps that had at least one of Stork, Bisu, Jaedong, or Flash having the highest ELO out of everyone. The reason for this was because I figured these players would top the charts for multiple maps, making the comparison between maps easier for me, and because these players played in an era where the sample size for the maps in question was plentiful.
I would examine the following things:
1) Map size 2) Number of starting positions 3) Mineral and gas count at main base 4) Gas count per player if divided evenly 5) Total number of games played 6) Whether it was used in the ProLeague 7) Most played mirror-match-up 8) Least played mirror-match-up 9) Match-ups with the losing race having a win rate of less than 40% 10) Any noteworthy points about the map I felt like adding
Now that I've outlined my criteria, I'll start my analysis for each of the players.
Stork: Highest ELO in two maps
Return of the King
1) Map size: 128 x 128 2) Number of starting positions: 4 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 6 5) Total number of games played: 364 6) Whether it was used in the ProLeague: No 7) Most played mirror-match-up: ZvZ 8) Least played mirror-match-up: PvP 9) Match-ups with the losing race having a win rate of less than 40%: ZvP 10) A bad map for protoss
Baekmagoji
1) Map size: 128 x 96 2) Number of starting positions: 2 3) Mineral and gas count at main base: 13M 2G 4) Gas count per player if divided evenly: 4 5) Total number of games played: 126 6) Whether it was used in the ProLeague: No 7) Most played mirror-match-up: PvP 8) Least played mirror-match-up: ZvZ 9) Match-ups with the losing race having a win rate of less than 40%: PvT, TvZ 10) A good map for protoss
Conclusion: Stork wasn't good enough at multiple maps for me to draw any half-decent conclusions. Maybe Stork likes imbalanced maps with relatively smaller number of games played on it, even if they happen to be bad for protoss.
Bisu: Highest ELO in three maps
Destination
1) Map size: 96 x 128 2) Number of starting positions: 2 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 5 5) Total number of games played: 1195 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: TvT 8) Least played mirror-match-up: PvP 9) Match-ups with the losing race having a win rate of less than 40%: None 10) A good map (in terms of balance, and number of games played)
Medusa
1) Map size: 128 x 128 2) Number of starting positions: 3 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 4.5 5) Total number of games played: 728 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: PvP 8) Least played mirror-match-up: TvT 9) Match-ups with the losing race having a win rate of less than 40%: None 10) A good map (in terms of balance, and number of games played)
Aztec
1) Map size: 128 x 128 2) Number of starting positions: 3 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 6 5) Total number of games played: 728 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: PvP 8) Least played mirror-match-up: TvT 9) Match-ups with the losing race having a win rate of less than 40%: PvT, ZvT 10) A bad map for terran
Conclusion: The sample size is too small, but perhaps Bisu enjoyed three player maps that wasn't friendly towards the terran players. He also seems to thrive in balanced maps.
Jaedong: Highest ELO in nine maps
Circuit Breaker
1) Map size: 128 x 128 2) Number of starting positions: 4 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 6 5) Total number of games played: 489 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: TvT 8) Least played mirror-match-up: PvP 9) Match-ups with the losing race having a win rate of less than 40%: None 10) A good map (in terms of balance, and number of games played)
Match Point
1) Map size: 96 x 128 2) Number of starting positions: 2 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 4 5) Total number of games played: 1007 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: PvP 8) Least played mirror-match-up: ZvZ 9) Match-ups with the losing race having a win rate of less than 40%: None 10) A good map (in terms of balance, and number of games played)
Fighting Spirit
1) Map size: 128 x 128 2) Number of starting positions: 4 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 7 5) Total number of games played: 1048 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: TvT 8) Least played mirror-match-up: ZvZ 9) Match-ups with the losing race having a win rate of less than 40%: None 10) A good map (in terms of balance, and number of games played)
Sin Peaks of Baekdu
1) Map size: 128 x 128 2) Number of starting positions: 4 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 4 5) Total number of games played: 664 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: ZvZ 8) Least played mirror-match-up: PvP 9) Match-ups with the losing race having a win rate of less than 40%: None 10) A good map (in terms of balance, and number of games played)
Python
1) Map size: 128 x 128 2) Number of starting positions: 4 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 6 5) Total number of games played: 1013 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: TvT 8) Least played mirror-match-up: PvP 9) Match-ups with the losing race having a win rate of less than 40%: None 10) A good map (in terms of balance, and number of games played)
Arcadia 2
1) Map size: 128 x 128 2) Number of starting positions: 4 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 6 5) Total number of games played: 281 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: ZvZ 8) Least played mirror-match-up: PvP 9) Match-ups with the losing race having a win rate of less than 40%: None 10) A good map (in terms of balance)
Sin Chupung-Ryeong
1) Map size: 96 x 128 2) Number of starting positions: 2 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 4 5) Total number of games played: 284 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: TvT 8) Least played mirror-match-up: PvP 9) Match-ups with the losing race having a win rate of less than 40%: TvP, ZvP 10) A bad map for protoss
Odd Eye
1) Map size: 128 x 112 2) Number of starting positions: 2 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 4 5) Total number of games played: 138 6) Whether it was used in the ProLeague: No 7) Most played mirror-match-up: TvT 8) Least played mirror-match-up: PvP 9) Match-ups with the losing race having a win rate of less than 40%: TvP 10) A bad map for protoss
Katrina SE
1) Map size: 128 x 128 2) Number of starting positions: 4 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 6 5) Total number of games played: 189 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: ZvZ 8) Least played mirror-match-up: TvT 9) Match-ups with the losing race having a win rate of less than 40%: PvT 10) A bad map for terran
Conclusion: Favours orthodox four player map pools with good balance, and has lots of games played on it. Does not seem to favour three player maps.
Flash: Highest ELO in ten maps
Benzene
1) Map size: 128 x 128 2) Number of starting positions: 4 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 5 5) Total number of games played: 195 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: PvP 8) Least played mirror-match-up: TvT 9) Match-ups with the losing race having a win rate of less than 40%: TvZ 10) GoRush said this map was good for him because it was strategical
Eye of the Storm
1) Map size: 128 x 128 2) Number of starting positions: 4 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 6 5) Total number of games played: 314 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: TvT 8) Least played mirror-match-up: PvP 9) Match-ups with the losing race having a win rate of less than 40%: None 10) A good map (in terms of balance)
Polaris Rhapsody
1) Map size: 96 x 128 2) Number of starting positions: 2 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 4 5) Total number of games played: 206 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: PvP 8) Least played mirror-match-up: ZvZ 9) Match-ups with the losing race having a win rate of less than 40%: None 10) A good map (in terms of balance)
Katrina
1) Map size: 128 x 128 2) Number of starting positions: 4 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 6 5) Total number of games played: 507 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: PvP 8) Least played mirror-match-up: TvT 9) Match-ups with the losing race having a win rate of less than 40%: PvZ 10) A good map for protoss
Grand Line SE
1) Map size: 128 x 128 2) Number of starting positions: 4 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 6 5) Total number of games played: 194 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: TvT 8) Least played mirror-match-up: PvP 9) Match-ups with the losing race having a win rate of less than 40%: ZvP, TvZ 10) A bad map for protoss
Othello
1) Map size: 128 x 128 2) Number of starting positions: 4 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 6 5) Total number of games played: 257 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: TvT 8) Least played mirror-match-up: PvP, ZvZ 9) Match-ups with the losing race having a win rate of less than 40%: TvZ 10) A good map for terran
Fortress SE
1) Map size: 128 x 128 2) Number of starting positions: 4 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 8 5) Total number of games played: 103 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: TvT 8) Least played mirror-match-up: ZvZ 9) Match-ups with the losing race having a win rate of less than 40%: TvZ 10) A good map for terran
Montecristo
1) Map size: 96 x 128 2) Number of starting positions: 2 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 6 5) Total number of games played: 112 6) Whether it was used in the ProLeague: No 7) Most played mirror-match-up: PvP, ZvZ 8) Least played mirror-match-up: TvT 9) Match-ups with the losing race having a win rate of less than 40%: TvZ 10) A good map for protoss
Hitchhiker
1) Map size: 128 x 128 2) Number of starting positions: 2 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 5 5) Total number of games played: 325 6) Whether it was used in the ProLeague: No 7) Most played mirror-match-up: TvT 8) Least played mirror-match-up: PvP 9) Match-ups with the losing race having a win rate of less than 40%: None 10) A good map (in terms of balance)
Roadrunner
1) Map size: 128 x 128 2) Number of starting positions: 4 3) Mineral and gas count at main base: 9M 1G 4) Gas count per player if divided evenly: 6 5) Total number of games played: 189 6) Whether it was used in the ProLeague: Yes 7) Most played mirror-match-up: TvT 8) Least played mirror-match-up: PvP 9) Match-ups with the losing race having a win rate of less than 40%: TvP 10) A bad map for protoss
Conclusion: Tends to do better on maps with high frequencies of the terran-versus-terran match-up. Does not seem to favour three player maps. Does not seem to favour maps that have a lot of games played on it (Katrina is the only map to have more than 500 games played on it).
In closing, Bisu seems to favour orthodox three player macro-management maps that punishes terran players more than the other maps. Jaedong seems to favour orthodox four player macro-management maps that has lots of games played on it. Flash seems to have a preference slightly more unorthodox maps, but does not seem to like three player maps.
I'll try extrapolating some of my ideas in more detail, at the cost of getting things massively wrong.
This is the landscape of modern-day professional Brood War. The start of the 2007 season with the offline qualifiers for the next individual leagues just after the last ever finals for the 2006 season (GomTV MSL Season 1) ended.
1) Flash makes his debut as a professional gamer. 2) ProLeague schedule is extended to five days a week, roughly doubling in schedule. 3) MSL raises their overall prize pool to roughly match the prize pools given for OGN StarLeague.
So this is the data from 15th March 2007 ~ 4th August 2012, from YGOSU.
Total number of games played: 22,968 matches Total number of games played on two player maps: 7,285 (31.72%) Total number of games played on three player maps: 3,818 (16.62%) Total number of games played on four player maps: 11,217 (48.84%) Total number of games played on four player maps (sized 128 x 128) with 9 mineral fields in the main base: 10,239 (44.58%)
The most orthodox map in the modern era would have the following characteristics.
1) Size: 128 x 128 2) Starting positions: 4 3) Number of mineral fields in the main base: 9
The most familiar examples of these kind of maps would be Python, Fighting Spirit, and Circuit Breaker. The balance on these kind of maps are as follows:
55:45 in favour of terran players in TvZ 51:49 in favour of protoss players in PvT 55:45 in favour of zerg players in ZvP
Three player maps with 9 mineral fields (such as Medusa, Outsider, and Aztec) in the starting base would shift this commonly accepted form of balance (where terran shits on zerg, zerg shits on protoss, and terran players struggle slightly versus protoss) to the following:
51:49 in favour of terran players in TvZ 52:48 in favour of protoss players in PvT 56:44 in favour of zerg players in ZvP
On standard three player maps, the zerg race replaces the terran race as the most successful race. With three player maps being used only sparingly in the most recent professional scene, the large success zerg players tended to enjoy on it was somewhat restricted.
Two player maps with 9 mineral fields (such as Destination, Heartbreak Ridge, and Match Point) in the starting base would give the following results in terms of balance:
53:47 in favour of terran players in TvZ 55:45 in favour of protoss players in PvT 56:44 in favour of zerg players in ZvP
Orthodox two player maps historically had the most success in finding parity between the three races (all three races were given a tough, and an easy match-up), and were used quite often. The reason why two player maps were not used as frequently was, in my opinion, probably due to commercial reasons, with build order disadvantages being harder to overcome on these kind of maps, and the back-and-forth nature of four player maps not being seen as often. Players with a build disadvantage would often die outright, or die a slow, horrible death with no other starting locations to fight over, or having opportunies showcase unreal multi-tasking.
With this being my opinion on the these kind of maps, here are the most notable two player, three player, and four player maps.
Orthodox two player maps with over 700 games first used after 2007
Bisu is weird as fuck, because he sucks balls compared to other top players in generic two player maps, but did fantastically when he played on Destination, which is probably due to the fact that this was one of the most used maps when Bisu was killing it in both the individual leagues and the ProLeague, whereas for most of his career he always sucked in one or the other. Considering his relative mediocre performances on Sin Peaks of Baekdu, one of the few two player maps in history to have over 500 matches played on it, I think it is fair to say Bisu did not like playing on two player maps where his multi-tasking abilities were somewhat limited, and his fantastic performance on Destination is probably him performing well despite his natural aversion to two player maps.
Orthodox three player maps with over 300 games first used after 2007
Although Bisu's ever fluctuating form makes things difficult to read, I think it's correct in saying that he enjoyed more success in orthodox three player maps than orthodox two player maps.
I also think I was correct in saying that Jaedong did poorly in three player maps, despite zerg players in general performing better on them on average. The only reason for this I can think of is the difference in the total number of games played (three player maps rarely had over 500 matches played on it). Three player maps in general were not used as much, so Jaedong's tendencies to do better in maps with more games played on it, overcame the natural inclination these maps had towards the zerg race, which must mean that Jaedong was more influenced by his familiarity with the maps, than the relative small variations these maps had in terms of balance.
FanTaSy is weird because he does not seem to hate three player maps as much Flash, who seems to do worse in orthodox three player maps in general. FanTaSy also enjoyed great success in Pathfinder, the three player map used for two of the three most recent OGN StarLeagues, whereas Flash once again did not take such a liking towards it.
Orthodox four player maps with over 500 games first used after 2007
Jaedong has a great track record in almost all popular four player maps, except for La Mancha, which might be better explained by his drop in form towards the latter stages of his professional career, rather than a dislike of four player maps. His performances on Circuit Breaker was stellar as well, which contrasts to his extremely mediocre performances on La Mancha.
Stork's ELO points in general seems to drop-off as more starting positions are added, which might mean his relatively low eAPM, micro-management focused playstyle might have not been suited for four player maps.
Bisu seems to have the opposite tendencies to Stork, doing generally better in maps with more starting positions to showcase his multi-tasking abilities.
Let me see if my theories are correct. I'll be checking some statistics on YGOSU.
Stork's career win rate on:
2 player maps: 64.7% 3 player maps: 61.3% 4 player maps: 61.8%
Stork's win rate after Flash's debut on:
2 player maps: 66.7% 3 player maps: 64.4% 4 player maps: 62.0%
Verdict: My statements are true, especially if you read into his results after Flash made his debut in 2007.
Bisu's career win rate on:
2 player maps: 66.4% 3 player maps: 68.4% 4 player maps: 65.8%
Bisu's win rate after Flash's debut on:
2 player maps: 66.0% 3 player maps: 70.6% 4 player maps: 66.3%
Verdict: Yes Bisu does like three player maps, but perhaps my assessment of Bisu on two player maps may have been off.
Jaedong's career win rate on:
2 player maps: 67.0% 3 player maps: 65.2% 4 player maps: 71.1%
Jaedong's win rate after Flash's debut on:
2 player maps: 67.6% 3 player maps: 67.1% 4 player maps: 71.2%
Jaedong's career win rate on:
Maps with over 1,000 matches played overall: 74.51% (record of 76-26) Maps with over 500 matches, but less than 1,000 matches played overall: 67.48% (record of 110-53) Maps with less than 500 matches played overall: 66.54%
In this graph, you can see that Jaedong, followed by Flash are the best performers on maps with over 1,000 matches played on it. Bisu, as per usual, is hard to point down due to his erratic nature.
Verdict: Yes Jaedong sucks on three player maps, and that may be because there are less games played on it overall, and Jaedong had a tendency to do well on maps that were played more, rather than being the best at adapting to new maps with smaller game pool overall.
Flash's career win rate on:
2 player maps: 66.8% 3 player maps: 74.1% 4 player maps: 73.6%
Flash's career win rate on:
Maps with over 1,000 matches played overall: 67.9% (record of 72-34) Maps with over 500 matches, but less than 1,000 matches played overall: 72.8% (record of 99-37) Maps with less than 500 matches played overall: 71.2%
Verdict: I was totally off about Flash being mediocre on three player maps, and wildly over-estimated his abilities on two player maps. It seems Flash was the least comfortable on two player maps. Also, while it is true that Flash is good at maps with less games played on it, it is mainly because Flash is just good in general, rather than being more adaptable to new maps, it seems. I was wrong that Flash prefered maps with smaller game sample sizes, just because Jaedong sucks on unfamilar maps doesn't mean Flash prefers them to familiar ones, although Flash does falter slightly on maps with huge sample sizes, but that may becaause those maps are rather few in number.
The most played maps after 2007, going from the 20th most played map, Sin Chupung Ryeong, to the most played map, Destination. Finding any pattern is difficult, and while Jaedong doing better as maps have more matches played on it is the most apparent out of the three, it is not as clear cut as I first thought it was. Although it may be a coincidence, both Jaedong and Flash stay above an ELO of 2100 once they play on maps with over 700 matches played on it.
FanTaSy's career win rate on:
2 player maps: 58.2% 3 player maps: 63.6% 4 player maps: 63.5%
Verdict: It seems two player maps does not seem to be a friendly battle-ground for these two terran players.
One random thing I thouht of is that FlaSh has a dropped quite a few important games on 2p maps like vs Bisu on Chain Reaction, vs Sea Taekbaek Mountains ASL2 Final Match Group Stage and vs Sea again next season on Match Point, ASL3 Finals.
On August 05 2017 22:20 c3rberUs wrote: One random thing I thouht of is that FlaSh has a dropped quite a few important games on 2p maps like vs Bisu on Chain Reaction, vs Sea Taekbaek Mountains ASL2 Final Match Group Stage and vs Sea again next season on Match Point, ASL3 Finals.
I think it is due to the linear nature of the maps. On two player maps, I personally believe that it is easier to hold on to, or making a killing blow out of an advantageous build order, due to the more restricted number of options available.
That means people who are really good at having a pre-planned build (easier to plan for someone if there are fixed starting positions) and tend to do poorly as time passes by on more complex or demanding maps, are at their most comfortable on two player maps.
People (probably more to be found) with unnaturally high win rate on two player maps compared to the other maps (it usually needs them to be shit on non two player maps):
sKyHigh on:
Two player maps: 67.0% Non two player maps: 54.8%
Horang2 on:
Two player maps: 63.2% Non two player maps: 51.4%
BackHo on:
Two player maps: 55.7% Non two player maps: 45.5%
Really on:
Two player maps: 57.8% win rate Non two player maps: 48.7% win rate
Calm on:
Two player maps: 61.4% Non two player maps: 54.8%
TheROCK[3.33] on:
Two player maps: 52.9% Non two player maps: 46.3%
Iris on:
Two player maps: 60.1% Non two player maps: 53.8%
Sea.GuemChi on:
Two player maps: 54.9% Non two player maps: 47.6%
UpMagic on:
Two player maps: 54.8% Non two player maps: 49.2%
Hwasin on:
Two player maps: 60% Non two player maps: 54.8%
How many of these players are known for their over-reliance on certain traits, rather than being a solid rock of a player who thrives in orthodox games with absolutely zero curve balls? This is like the who's who list of flawed players who still somehow managed to find success.
So for someone like Flash, perhaps it is only natural that he prefers four player maps, due to how pre-planned build order designed to absolutely crush someone, has less potency. Someone like Jaedong or Stork, who both had killer micro-management throughout their entire careers, would be able to secure early game leads, or get themselves out of an impossible situation with a miraculous micro-management abilities. Someone who methodically broke down their opponents with a series of good decisions, and thrived when there were more bases to manage, would be more prone to failure on two player maps in my opinion.
I just want to say that all your blogs are great and I feel like your conclusions here and in the SC:R beta thread are especially interesting as they explore/confirm the amount of variance in skill-sets among top players. The fact that TBLS all had different ways of approaching the game lends credence to the game's vast complexity. It's fantastic how much depth Brood War has to offer and awe inspiring to see great players express themselves through the game with their perfectly honed skills.
I did have one question though concerning your post here though. According to your working definition, strategy was "how good players are coming up with cheese". Now while I don't disagree that creating unorthodox builds is definitely a part of strategy, I definitely consider it to be just as much if not more about the ability to set accurate goals and long-term planning.
So my question is, which players emphasized strategy (as I have defined it) the most and of which of TBLS has the most strategic playstyle? I feel that since decision making, macro-management, and large scale army control is such a integral part of positional play, it would only be natural to call Flash a very strategic player, which is why he excels at four player maps. I would really like to hear your thoughts and research concerning this.
On August 06 2017 09:43 chaosTheory_14cc wrote: I just want to say that all your blogs are great and I feel like your conclusions here and in the SC:R beta thread are especially interesting as they explore/confirm the amount of variance in skill-sets among top players. The fact that TBLS all had different ways of approaching the game lends credence to the game's vast complexity. It's fantastic how much depth Brood War has to offer and awe inspiring to see great players express themselves through the game with their perfectly honed skills.
I did have one question though concerning your post here though. According to your working definition, strategy was "how good players are coming up with cheese". Now while I don't disagree that creating unorthodox builds is definitely a part of strategy, I definitely consider it to be just as much if not more about the ability to set accurate goals and long-term planning.
So my question is, which players emphasized strategy (as I have defined it) the most and of which of TBLS has the most strategic playstyle? I feel that since decision making, macro-management, and large scale army control is such a integral part of positional play, it would only be natural to call Flash a very strategic player, which is why he excels at four player maps. I would really like to hear your thoughts and research concerning this.
As I have understood it, the designer of MyStarcraft, the simulation game from which took the metrics from, defined strategy as more of a player's ability and tendencies to come up with potent pre-planned build orders (basically cheese, whether it is economical in nature, or more focused on finishing the enemy at a certain timing), rather than good strategical mind for adaptation during the game. Which why a player like Calm, who was fantastic at coming up with pre-planned build orders, but wasn't exactly a pretty sight to see if the direction of the game didn't go in the way he expected, was given such a high rating for strategy in that game.
I believe the ability you are defining would be most closely matched by the statistic that is labelled as decision making within that game. Adaptation within the game, and making the correct series of decisions, is something that is rewarded more heavily in four player maps, where the power of pre-planned strategy is not as potent, and the more diverse set of options available in terms of decision making allows you to come back from a build order deficit with great decision making more easily.
I think it was Light that once said what differentiated Flash from the others was his decision making. He said that if there are ten decisions to make in a row, most gamers, even the great ones, struggle to make five good decisions within that time frame. Flash, on the hand, would regularly make eight, or nine, great decisions in a row.
I believe that Flash has not only the best decision making out of Taek-Beng-LeeSsang, but out of any player in history. Others may have had manuals for success that ws years ahead of their time such as illoveoov, or sAviOr, while others may have had superior abilities in coming up with creative solutions to cover up their mechanical deficiencies such as Silver, or UpMagic, while others may have had a slightly better sense to pull off high-risk, high-reward decisions on the fly such as EffOrt, but no gamer in history had the ability to make such consistent, reliable, and correct decisions throughout the entire game like Flash, nobody even comes close in my opinion.
I don't think Flash was the most creative genius, even amongst his peers, but I do believe that Flash succeeded in creating the most efficient algorithm in his mind for approaching the game overall. Combined with top notch fundamentals, Flash during his absolute prime is the player with the least number of flaws in history.
I was always a fan of more visceral players who were more suited for highlight reels, players who would often come up with unthinkable outplays through untouchable mechanical prowess, or a moment of sudden inspiration. Players who were so ahead of their time in terms of mechanical prowess that it looked like an adult running circles around little kids. Players like NaDa, July, and Jaedong, were the kind of players worshipped. Flash never was that kind of player in my opinion, and just played the game in a clean, efficient manner without making much mistakes.
Bisu is special because he is such an enigma. Against the zerg race he is up there with the greatest of all time, both in terms of mechanical ability and decision making. He did one superlative play after another, and on top of thought ten steps in advance, making tiny adjustments depending on how well his play turned out beforehand. Against the terran race, I don't think he ever mastered the match-up. Although his other skill-sets are outstanding enough for him to smash mediocre players, his sloppy play compared to top tier play-making protoss players such as JangBi in his prime made me wonder if he had multiple personas within him.
Stork, in my eyes, had the potential to make it big as a player with his outstanding micro-management abilities, when he became the youngest protoss player in history to make it past the offline qualifiers at the age of 16 in 2005, but he wasted his potential to become great by playing too much World of Warcraft, and although he had great results in 2007, there was a limit to how great he could become due to his lower eAPM compared freshly crowned kings of the each the races, Flash, Jaedong, and Bisu. He remained relevant as a top player, but he was a relic of the past with limited potential due to his relatively poor performances in orthodox four player maps compared to other elite tier players.
That's what I think, my views on Jaedong never changed much throughout the years, so I won't bother repeating the same stuff.
This interview has Grubby explaining how Moon became scarier if he was given time to prepare, while he may have been the better of the two in situations where that wasn't possible, such as competing in World Cyber Games.
I personally believe preparation was the most extensive in best-of-five series in tournaments that had its schedule spaced out over several months (OGN StarLeague, MSL, and GomTV Classic in the more recent years). Meaning by the time best-of-five series were played, players had plenty of time to specifically prepare for a single opponent.
Best-of-fives played in tournaments that were rushed in terms of schedule like GomTV Invitational, showmatches hosted by GomTV or SuperFights, and IEF 2011 cannot be seen to have a similar level of preparation for this reason.
On the opposite side of the spectrum, Winners League could be said to have the most hectic scheduling, and arguably has one of the least amount of time to individually prepare for each and every opponent before the games begin. I will try to contrast and compare the players who shined the brightest, in these two opposite sides of the spectrum.
For judging Winners League performance, I will add the ELO points from the three Winners Leagues, to make sure players who padded their statistics from all-killing shit teams like Air Force ACE, or Estro multiple times weren't rewarded too heavily.
Top ten players with the highest combined ELO points from all three Winners Leagues (2009, 2010, and 2011)
1. versus Calm: 3-1 2. versus BeSt: 3-0 3. versus Kwanro: 3-1 4. versus Movie: 3-1 5. versus Jaedong: 1-3 6. versus MVP: 3-1 7. versus Pure: 3-0 8. versus free: 3-0 9. versus EffOrt: 2-3 10. versus Jaedong: 3-0 11. versus ForGG: 3-0 12. versus FanTaSy: 3-2 13. versus Jaedong: 3-2 14. versus free: 3-1 15. versus Jaedong: 3-1
Jaedong's record in a properly scheduled best-of-five setting within the same time frame (17/01/2009 ~ 09/04/2009)
1. versus herO: 3-0 2. versus FanTaSy: 3-2 3. versus Canata: 3-2 4. versus Calm: 1-3 5. versus FanTaSy: 3-1 6. versus YellOw[Name]: 3-0 7. versus Stats: 3-0 8. versus Kal: 3-0 9. versus Flash: 3-1 10. vesus Midas: 3-0 11. versus Calm: 3-1 12. versus Flash: 0-3 12. versus Sea: 3-1 13. versus Light: 3-2 14. versus Flash: 2-3 15. versus Stork: 3-2 16. versus Flash: 1-3 17. versus SnOw: 3-2 18. versus Hydra: 2-3
Light's record in a properly scheduled best-of-five setting within the same time frame (17/01/2009 ~ 09/04/2009)
So in between early 2009 to early 2011, both Flash and Jaedong performed well across the board. However, probably the third best performer in a best-of-five setting in this time frame, FanTaSy, did not distinguish himself from the rest as he did with more preparation in the individual leagues, where he was thwarted only by Flash, and Jaedong.
On the other side of the spectrum is Bisu, who used to be quite the beast in a best-of-five setting, was one of the best Winners League player of all time, even if you take his massive slump in 2010 into account, but was embarassingly bad in a best-of-five setting for a player of his calibre. Bisu was one of the better players against the terran race if given minimal time to be prepared against, but once players were given more time to study his weaknesses, he couldn't repeat his fantastic performances in the Winners League into a best-of-five setting, even though we are examining the same exact time frame, only considering best-of-fives played between early 2009 and early 2011.
The list of top performers in the Winners League was dominated by terran players, which seems to suggest that games played without much preparation seems to slightly favour terran players, which may be why terran players such as Light, HiyA, and Leta struggled to showcase their talents in the individual leagues, where players were given more time to prepare for a specific opponent. Flash was just a monster who did well no matter what the format. FanTaSy seems to be the odd-ball, one of the rare terran players who shined brighter given extra time to prepare.
I wish there were more tournaments to compare and contrast, but I am not fully aware of the details of how much time players had to prepare for a lot of the tournaments. For example, the ProLeague seasons sometimes had the entries for the upcoming games released, but after the match-fixing scandal in 2010, they stopped doing that to stop players from fixing matches. Any players who faced a sudden decline, or improvement in results from Shinhan 2009/2010 ProLeague, and Shinhan 2010/2011 ProLeague may have been influenced by that change in the circumstances. The lack of exposed entries meant that all the teams had to now guess which players they had to prepare for, which meant that hardcore specific preparation for a certain player was no longer possible.
Some notable beneficiaries of the lack of exposed entries (less chance of being prepared for hardcore) in the ProLeague:
Bisu: ELO raised from 2056 (27th) to 2271 (1st) Light: ELO raised from 2022 (42nd) to 2161 (2nd)
Some notable casualities of the lack of exposed entries (less chance of being prepared for hardcore) in the ProLeague:
FanTaSy: ELO dropped from 2136 (3rd) to 2002 (49th)
Of course, this doesn't explain the change in form a player faces over the course of a year, and there may be exceptions. However, Light was killing it in the 2010 Winners League and progressed further than he ever did in the individual leagues in 2010, and actually did worse in the individual leagues and the 2011 Winners League. Yet somehow, Light did better in the regular formats of the ProLeague by a stupendous amount despite doing worse in almost every other platform, and in my opinion, the biggest reason for that is the fact that the entries being exposed in advance was no longer a thing.
So if we have a spectrum of players, starting from those who do worse with more preparation, and those who do the best with extra time for preparation. It would look something like this:
Light - Sea - FanTaSy
These three terran players were reasonably stable in form, so it was easier for me to judge their results under different settings.
The forms for players who competed in Shinhan 2009/2010 ProLeague, and Shinan Winners League 2010 should be somewhat similar since the map pools overlap, and the time frame overlaps somewhat.
The forms for players who competed in Shinhan 2010/2011 ProLeague, and Shinhan Winners League 2011 should be somewhat similar since the map pools overlap, and the time drame overlaps somewhat.
Season 2009/2010:
Top players in terms of ELO points who do better in Winners League (less time to prepare for opponents):
Flash (1st, compared to 2nd in the ProLeague) Light (2nd, compared to 42nd in the ProLeague) Leta (3rd, compared to 15th in the ProLeague) Sea (5th, compared to 18th in the ProLeague)
Top players in terms of ELO points who do better in the standard ProLeague format (more time to prepare for opponents):
Jaedong (1st, compared to 4th in the Winners League) FanTaSy (3rd, compared to 7th in the Winners League) KwanrO (4th, compared to 19th in the Winners League) free (5th, compared to 12th in the Winners League)
So, if my theory that Winners League rewards players who do better with less specific preparation against a certain player, they should have done better in the 2010/2011 Shinhan ProLeague, assuming that their forms remained the same. Having no entries being exposed in advance should help them.
Flash: No (still 2nd) Light: Yes (moved up to 6th) Leta: Yes (moved up to 13th) Sea: Yes (moved up to 12th)
In the same vein, people who were doing better in the regular format of the ProLeague compared to the Winners League in 2009/2010 should have done relatively worse in 2010/2011, assuming that their form did not change drastically over the course of a year.
Jaedong: Yes (Moved down to 3rd) FanTaSy Yes (Moved down to 49th) KwanrO Yes (Moved down to 67th) free: Yes (Moved down to 107th)
In conclusion, here are the top players of 2009 to 2011, being lined up in a spectrum from those who rely a lot on preparation for an opponent before the actual gameplay begins, to those who suffer from having somebody studying and preparing for their playstyle.
FanTaSy - Jaedong - Flash - Bisu - Light
Stork lies somewhere in the middle, but his form in general between 2009 and 2011 was not that stellar, so it is harder to say exactly where he belongs.
Remember, this spectrum only accurate when taken by the context of the time frame I mentioned above. It is difficult to say for sure how these players have changed throughout the years when there are a severe shortage of tournaments with lots of sample size, and significantly contrasting styles of format, or circumstances so that I can gauge exactly by how much these players were influenced by preparation time.
Because I have followed the career of Jaedong the most closely, I feel as if he has two eras of peak performance as a professional player. One era that is marked by his insane mechanical prowess (circa 2007 ~ 2008), and slightly questionable preparation skills, and the era that brought him arguably greater success in terms of results, when Jaedong had deteriorated as a mechanical monster relative to others, but had become an overall smarter player, and a more complete player overall, even if his in-game skills were not as eye-popping as before (circa 2009 ~ 2010).
I kind of wish there were Winners League style tournaments with hundreds of games available for analysis, but sadly that is not the case. The closest tournament with a significant enough sample size of player participation and overall game count, that took place over relatively short span of time with little time for preparation is Seoul 2007 e-Sports Festival, a tournament that was open for all professional gamers, and was completed within a span of three days, after 469 games had been played. The fact players who would go on to display good performances in the Winners League in the years to come such as Jaedong (finished 1st), and HiyA (finished in the top four) would place highly in this tournament despite being finished around one and half years before the first games for the 2009 Winners League even began, might imply that players like Jaedong and HiyA were already top players if given almost zero time to prepare for specific opponents in mid-2007, a trait that stuck with them for years to come.
For clarity's sake, players who are fantastic at pre-game preparation do not necessarily have a high in-game IQ. You can have the smartest, most abusive build order ready to be used for just the right opponent, but it does not mean you have the ability to adapt your initial plan according to how the game plays out, although a combination of both would indeed to deadly.
Although it's not as simple as I put it (and there are always exceptions to the rule), I would say it is roughly the case that people who tend to do better in individual leagues (especially further up the bracket you go) tend to do better with more preparation, while people who do better in a ProLeague setting (Winners League especially) tend to do better with less time to be prepared for.
An older player who is a perfect (no pun intended) candidate of a player who excelled under situations with minimal preparation is XellOs, who had incredible performances in MBC Game Team League (another old school team league with an All-Kill format), as well as tournaments with shorter scheduling such as WCG Korea.
I'm trying to think of bunch of tournaments with tighter schedule, but enough overall sample size so that a couple of games worth of hot streak, or a poor game or two doesn't heavily influence the statistics. Maybe when I find the time I'll revisit this blog for further investigation.
I think for a more accurate analysis, you need to consider that not all Bo5s were created equal, so to speak. Some Bo5s had tighter schedules than others like for example, some players who played in one league only may have had more preparation time than players who participated in two or even multiple leagues. But of course, unless you're going to devote a good amount of time on doing someting like this, this is not practical at all.
On August 08 2017 23:54 c3rberUs wrote: I think for a more accurate analysis, you need to consider that not all Bo5s were created equal, so to speak. Some Bo5s had tighter schedules than others like for example, some players who played in one league only may have had more preparation time than players who participated in two or even multiple leagues. But of course, unless you're going to devote a good amount of time on doing someting like this, this is not practical at all.
Early 2009 had one of the most stacked schedules in history. No other period in history had such significant overlap between the individual leagues and the Winners League.
GomTV Classic Season 2 - Lost Saga MSL - Winners League 2009 - Batoo OGN StarLeague
These four competitions had at least two weeks of overlap, even though they ended at different points of 2009. JangBi, the player with the busiest best-of-five schedule (he competed in five) during the duration of Shinhan Winners League 2009, and he had about anything from a week to a month of downtime between his best-of-five series in terms of individual league schedule.
I think a week of preparation is a significant enough, even if it is not quite the same as having your first best-of-five in months, compared to the daunting task of preparing for two teams' worth of players in a week (even if you can narrow it down to about eight players) during the Winners League. Preparing for one player for a maximum of five games during the space of a week may be tough, but preparing for four maps, not knowing for sure which opponents you will face (and in which order), two times a week is harder in terms of specific preparation in my opinion.
No matter how tight the individual league schedule, a best-of-five series in a properly scheduled tounrament will in general, have better preparation from both sides, than the matches seen in a Winner League. Of course there are exceptions, but these are the most extreme cases of the spectrum in terms of average level of preparation put into each individual games, in my opinion.
Every player had various schedules demands (games played for Group A of Lost Saga MSL were played a month in advance of games for Group H, for example), and it was sheer luck whether your ProLeague schedules demands got in the way of your individual league practice. It was something the players had to cope with, but individual league in general gave you more time to prepare for each opponent the further you progressed as a general rule of thumb, even if every player had differing circumstances.
The only players in recent history to play multiple best-of-fives for different tournaments (that had proper scheduling over a period of several months) within the space of less than a week (normal down time if you only participate for one individual league) like sAviOr did at the peak of his powers is the following:
Stork in mid-2007 (two semi-finals within the space of two days, one of which he failed)
Jaedong in mid-2009 (two semi-fianls within the space of two days, one of which he failed)
Iris in mid-2009 (one finals and one semi-finals within the space of five days, both of which he failed)
Flash in early-2010 (one finals and one semi-finals within the space of four days, one of which he failed)
Flash in mid-2010 (one semi-finals and one quarter-finals within the space oft two days, both of which we won)
Jaedong in late-2010 (one finals and one semi-finals within the space of five days, one of which he failed)
Stork in early-2011 (one finals and one quarter-finals within the space of three days, both of which he failed)
The list becomes longer if you include small invitarional tournaments, and showmatches that used best-of-five series to settle things, but I'm talking specifically about best-of-fives that had at least a week of scheduled downtime within the tournament for proper preparation, since I am trying to compare performances in best-of-five series that had at least one week preparation for one player for a maximum of five games, and performances in the Winners League format where if you actually planned anything close to the level you did for a best-of-five series in an individual league, you probably prepared for one to three likely players for each of the eight maps you were going to play in if you wanted to All-Kill both teams.
So apart from players mentioned, nobody after sAviOr was good enough to bitch about an extremely tight schedule if you ignore the usual ProLeague demands, which something all professional gamers had to deal with. Everybody, even the busiest players, all had a week of downtime, and if they had a ProLeague schedule squeezed in between that, that was something that was possible even if he only participated in one individual league, and was good enough to go far into the bracket stages.
So considering only Stork, Jaedong, and Flash were the only players in recent years unlucky enough to play best-of-five series two days in a row, perhaps they could argue about the lack of prepation even compared to the Winners League schedule (usually two matches a week played four days apart from one another, meaning you have to prepare for eight best-of-one matches if you plan to All-Kill the opposing teams), but apart from that, there's nobody who suffered an individual league schedule that was much worse than the Winners League schedule in terms specific preparation.
I wanted to add one more thing about Bisu's tendencies in the protoss-versus-terran match-up, and how he approaches a best-of-five series in that particular match-up.
FanTaSy - Jaedong - Flash - Bisu - Light
This was my rough spectrum of top players between early 2009 and early 2011, and how Bisu was more like Light, a player who exceled when everybody had minimal time for pre-game preparations (Winners League), and suffered as more time was allowed for the preparation for a specific opponent (best-of-fives in the latter stages of major individual leagues).
A comment from Mind seems to line up quite well with this theory. Mind said that when he practiced with Bisu, he was the type to just spam games non-stop on the given map pool for the upcoming important match. This is in very stark contrast to someone like iloveoov, one of the greatest minds of the game, practiced, who focused heavily on the optimization of strategies and build orders rather than perfecting his execution through endless practice.
Mind said that Bisu never spent time discussing his choice of builds during his practice sessions, and this coincides with what Kingdom once said about the difference between BeSt and Bisu, saying that BeSt was the more cerebral of the two players, while Bisu had the superior work ethic. Mind also said that Bisu wasn't the type to pay attention to all the meticulous details to really evil (as Grubby put it, in the video I linked above) builds, but someone who relied heavily on his form, and high levels of performance to succeed.
This was what I always thought about Bisu. Bisu, especially during his SK Telecom T1 days, was someone who needed to play well in order to win. This is something that Flash also said about Mind, saying that Mind was someone who cared a lot about his level of performance, rather than the end result, saying that Mind cared more about winning in a pleasing manner than winning more, a trait that Flash himself said that he didn't share with Mind.
Although it isn't as simple as I put it, if we were to divide the list of players into those with brains over brawns, and those with brawns over brains, Bisu, during his stay at SK Telecom T1 became much more focused on his brawns (Sea said that Bisu became a much "better" player after he transferred to SK Telecom T1), which is a pity because I personally believed Bisu to have had a healthy mix of brains and brawns during the early stages of superstardom.
FanTaSy, despite coming from the same team, was a player who had more brains than brawns in my opinion. Perhaps it was the presence of mentors that shaped the players to a degree, with PuSan (one of the smartest protoss players in terms of strategy in my opinion) molding Bisu to a certain degree during his stay at MBC Game HERO, and iloveoov playing a part in FanTaSy's development as a cerebral player.
EPT
![[image loading]](http://i.imgur.com/NOnoaFu.jpg)
![[image loading]](http://i.imgur.com/0mplN5k.png)
![[image loading]](http://i.imgur.com/ODXzA7q.png)
![[image loading]](http://i.imgur.com/3Brs2Sg.png)
![[image loading]](http://i.imgur.com/gCMY13p.png)
![[image loading]](http://i.imgur.com/iAxcVfh.png)
