|
On November 09 2010 16:55 Metalwing wrote:Show nested quote +On November 09 2010 10:17 Impervious wrote:On November 09 2010 09:53 CanucksJC wrote:
Pretty entertaining to read... pretty pointless otherwise I'd argue against it being pointless..... I was doing this type of stuff in the CSL, and I had ~80% accuracy on it. Imagine you have 4 players. The best one is 1, then 2, then 3, and the worst is 4. Your opponent has the same. Your best vs their best is an even match, it could go either way. Your 2nd vs their 2nd is also an even match, etcetera. However, if your best is facing a worse player from their team, you're more likely to win it. Same if your best is facing their third or fourth, etcetera. Therefore, the ideal way to set your lineup is to do it where you have 1v2, 2v3, 3v4, and sac the 4v1 to likely win the series in a 3-1 victory. Worst case scenario, you end up losing one of the 3 games you should win, and you get a 2-2 with a 50% shot at winning anyways. This can mean you win 3-0 as well, since it is a race to 3 wins and you don't have to play games 4 or 5. It's more complicated than that, when you have people with different strengths, different weaknesses, different map preferences, etc..... But that's the idea. I took a misfit group to the playoffs last year. And I used this type of strategy for it. It's also why I liked ICCUP as the server for the CSL, because the stats of players could be tracked and whatnot. I had some information to go by. This time around, I'm flying blind..... And I'm failing miserably (granted, it doesn't help when the average skill level of our opponents has been higher than our own in each match so far)..... But this stuff isn't pointless. Wow, a mastermind here. First, thanks for your feedback. About CSL, did you think of opening a LR thread for your games? Because I'd LOVE to follow that scene as well.
www.cstarleague.com - it's SC2 now. Last time was BW.
As for a LR thread - there aren't enough followers to really justify it. But if you search CSL on TL, you'll find a lot of info anyways.
|
Truth be told, once you're really good at anything, practicing 10 hours a day is a complete waste of time unless the environment changes so fast you're always 'new' to it.
|
Good read hojam, how are you doing?  Where in Turkey are you from? Spent the summer in buyukgada and around :D
|
On November 10 2010 01:44 Impervious wrote:Show nested quote +On November 09 2010 16:55 Metalwing wrote:On November 09 2010 10:17 Impervious wrote:On November 09 2010 09:53 CanucksJC wrote:
Pretty entertaining to read... pretty pointless otherwise I'd argue against it being pointless..... I was doing this type of stuff in the CSL, and I had ~80% accuracy on it. Imagine you have 4 players. The best one is 1, then 2, then 3, and the worst is 4. Your opponent has the same. Your best vs their best is an even match, it could go either way. Your 2nd vs their 2nd is also an even match, etcetera. However, if your best is facing a worse player from their team, you're more likely to win it. Same if your best is facing their third or fourth, etcetera. Therefore, the ideal way to set your lineup is to do it where you have 1v2, 2v3, 3v4, and sac the 4v1 to likely win the series in a 3-1 victory. Worst case scenario, you end up losing one of the 3 games you should win, and you get a 2-2 with a 50% shot at winning anyways. This can mean you win 3-0 as well, since it is a race to 3 wins and you don't have to play games 4 or 5. It's more complicated than that, when you have people with different strengths, different weaknesses, different map preferences, etc..... But that's the idea. I took a misfit group to the playoffs last year. And I used this type of strategy for it. It's also why I liked ICCUP as the server for the CSL, because the stats of players could be tracked and whatnot. I had some information to go by. This time around, I'm flying blind..... And I'm failing miserably (granted, it doesn't help when the average skill level of our opponents has been higher than our own in each match so far)..... But this stuff isn't pointless. Wow, a mastermind here. First, thanks for your feedback. About CSL, did you think of opening a LR thread for your games? Because I'd LOVE to follow that scene as well. www.cstarleague.com - it's SC2 now. Last time was BW. As for a LR thread - there aren't enough followers to really justify it. But if you search CSL on TL, you'll find a lot of info anyways.
Among my Firefox bookmarks now. Call it BW, call it SC2. I like it. I'll definitely check this.
On November 10 2010 02:33 Slardarxt wrote:Good read hojam, how are you doing? Where in Turkey are you from? Spent the summer in buyukgada and around :D
Fine hocam I'm from Ankara, the capital city, at METU (Middle East Technical University). I heard Büyükada is a nice place, but had no chance to see. Nice to see people like Jibba around. Shoutout to Jibba btw
|
Interesting read... kind of confusing at some parts.. but i think that is more or less how the post is structured. Well done!
|
United States12607 Posts
On November 10 2010 02:53 n0xi3 wrote:
Interesting read... kind of confusing at some parts.. but i think that is more or less how the post is structured. Well done!
I agree. These are really fun to read but I'm left confused as to whether the player you would have sent actually did play, or you are forecasting match results. Or are you looking at matches that you know the results of already, and just analyzing them ex post?
|
On November 10 2010 02:23 Roaming wrote:
Truth be told, once you're really good at anything, practicing 10 hours a day is a complete waste of time unless the environment changes so fast you're always 'new' to it.
Beginning a falsehood with 'truth be told' is kinda misleading. 10 hours practice time is probably not the best way to improve but it's far from a complete waste.
|
Interesting and educating read man, where do you chat during games? I would be interested in seeing more of your opinions.
Also, i dream of the day where perfectman and killer have > 50% winning percentages, that is the day that oz will rule (proleague 011-012.)
Edit: Man oz sucked before so i didn't care, but now that they show some promise I am rooting for them in all of my posts.... Lol sorry.
|
On November 10 2010 03:05 JWD wrote:Show nested quote +On November 10 2010 02:53 n0xi3 wrote:
Interesting read... kind of confusing at some parts.. but i think that is more or less how the post is structured. Well done! I agree. These are really fun to read but I'm left confused as to whether the player you would have sent actually did play, or you are forecasting match results. Or are you looking at matches that you know the results of already, and just analyzing them ex post?
My sample matches:
MBC vs Hentus: Played yesterday, results were forecast except 1 (that 1 match was played by exact same players at the exact same map). My selection of players played against Hentus players who really showed up at that match at exactly that sets, results are "logical forecast". That means I just did the assumption of "If this protoss player beats a pvp sniper, he can also beat another protoss who sucks at pvp and did the exact same strategy in his irl pvp game as the pvp sniper" at some points. Or "If this player wins a ZvZ against a stronger ZvZer, he could easily pwn the weaker one".
Look at spoilers for more detailed "logical forecast":
+ Show Spoiler +The statement of: "If Jaehoon can beat Horang2 with a strategy that Movie has just lost to in a PvP against Tyson, he would lose to Jaehoon as well" seems logical to me, what do you think?
Or, "if HyuN beats Hydra in a ZvZ, he could own HoGiL so badly that HoGiL wouldn't be able to sit for 3 days" assumption doesn't seem too insane either, agree?
Others were just particular examples about fictionary games.
+ Show Spoiler +In a world that Jaehoon beats Horang2 in a televised PvP, by.hero can beat Killer or Calm can beat Jaedong or HiyA can beat Kal. All these 4 cases carry exact same oddness imo. Or "if a rookie snow can beat Flash in an ace game, a well trained soulkey can do it as well" statement also seems logical.
On November 10 2010 11:59 TheNessman wrote:
Interesting and educating read man, where do you chat during games? I would be interested in seeing more of your opinions.
Also, i dream of the day where perfectman and killer have > 50% winning percentages, that is the day that oz will rule (proleague 011-012.)
Edit: Man oz sucked before so i didn't care, but now that they show some promise I am rooting for them in all of my posts.... Lol sorry.
I'm doing ACE vs SKT1 right now, I don't hang around on stream chats or tl irc whatsoever. But I'll be talking about game here and I'll be also here in game too.
|
So, let's come to today's match, where I'm coaching ACE against SKT1. Maps:
Set 1: Central Plains
Set 2: Aztec
Set 3: Fortress
Set 4: Icarus
Set 5: Benzene
Set 6: Grand Line SE
Set 7: Empire of the Sun
So, Central Plains is a map that  BeSt showed up often on. But, SKT's grand total of 0 protoss starters make me doubt if BeSt will really show up or not. This map wasn't played by any Terran players whatsoever, so i think there are 2 possible cases for this: A zerg player will play or BeSt will play. And, this makes me re-think the dodging strategy, and either an SKT zerg or BeSt doesn't matter. I'm going for the win.  YellOw, you're out on first set.
Note: I don't know shit about lineups for now.
EDIT: I opened the stream just about now, and:
By.Sun < Central Plains > YellOw
My view on this set: Well, I did the exact same thing as the ACE coach, but this set makes me feel awkward. Yes, I may score the win, but even if I do, SKT still has better players and I just used my best Zerg. My goals for other sets:
1- Create a M18M vs  Fantasy set.
2- Create a M18M vs Bisu set. Dodge Bisu.
3- Create a BeSt vs GoRush set.
4- Create a FBH vs SKTZERG set. Nvm, I just used him to dodge Bisu. I wish I didn't.
If I can do either 3 of them, and my players win, I win the game as well.
+ Show Spoiler +Yellow indeed lost to By.Sun. There is nothing I can do from here.
2nd set, I assume that BeSt will stay in his place, so will Bisu. So, I send FBH on this one.
EDIT for 2nd set:
Bisu appeared on 2nd set and I'm sure he would wtfpwn FBH. On the actual game, Rock is playing.
+ Show Spoiler +ROFL I'M CHANGING THIS TO FBH WTFPWNZ BISU AND GETS OUT OF THE BOOTH LIKE A BAWS.
Reference: FBH vs BeSt on Fortress
3rd set:
Looks like after Sun's win, SKT is going for 4-0, and they got reasons for it. I should have predicted that one and put Ruby on game 2. In this set, I think Fantasy will play. So, my choice is M18M
EDIT for 3rd set:
I'm making blunders over and over and over. I didn't pick any correct players after Yellow and I'm gonna get 4-0'ed. At least this is the look right now.
+ Show Spoiler +I have no friggin idea about how this would result. But DJW is not living his best day ever. He lost a PvT to FBH. M18M actually has a chance in this.
4th set:
Zerg? Fantasy? Map? Icarus.. Well, I don't take risks. Go Much
EDIT for 4th set:
This ends up being Much vs Soo_SKT. I can't forecast its ending, but Much has a losing PvZ streak and he's below %50. Still, I think he'd do better than ggaemo.
5th set:
This one may be a fakeout as SKT coach may try to overpredict me, the ACE coach and send another zerg. Anyway. GoRush, it's go time.
EDIT for 5th set:
This ended up being Fantasy and I feel like I'm thinking too much or Coach Park thinking too little. Because if Park thought the same as I wanted him to think, he would catch PvT sniper M18M, who is the actual choice of ACE coach, with Kwanro like a sitting duck.
|
On November 09 2010 10:10 Scarecrow wrote:Show nested quote +On November 09 2010 09:16 Sentient66 wrote:
Wow, this is extremely well-written and insightful. Thanks for putting the effort into doing this! I'm also fairly sure proleague lineups are submitted in advance (at least they used to be) so the analysis of set by set decisionmaking is pretty misguided.
Players are picked on a match by match basis ever since the match-fixing business.
|
you need to take into consideration that some maps are just imbalanced and some players may have better skills in some matchups than their statistics suggest. look past the stats and at the games if you really want an accurate assessment.
|
On November 10 2010 16:43 seRapH wrote:
you need to take into consideration that some maps are just imbalanced and some players may have better skills in some matchups than their statistics suggest. look past the stats and at the games if you really want an accurate assessment.
Yes, that needs consideration.
|
awful read.
first thing that cought my eyes was this line "...IChooseYouMagikarp, nor AssHo are vZ snipers, PerfectSlump..."
try to learn some respect dude.
|
Yeah, that Time has come. You saw that SKT1 has a good taste when it comes to  Victory, but me as the STX coach will make them taste different things. Like defeat.
Set 1: < Central Plains >
Set 2: < Aztec >
Set 3: < Fortress >
Set 4: < Icarus >
Set 5: < Benzene >
Set 6: < Grand Line SE >
Set 7: < Empire of the Sun >
Map 1: First map is the map that coach Park only sends Protoss. The last time, it was By.Sun vs YellOw and By.Sun won by outrageously high number of cannons. Actually, coach Park is not the only one. Coaches send hardly any zergs or terrans on this map. This map has only 3 games that both players aren't protosses. Even Stork won against Jaedong on this map. Coach Park would guess that Shuttle will be the player to be sent by STX. But, I don't think he will risk either Bisu or BeSt here and just dodge this map with By.Sun. But, when I look at the other maps, I don't see anything that BeSt may play on. So, this map will probably be a BeSt vs Shuttle and there is a possibility that these 2 players will make more units than rest of the players combined. Still, go Shuttle.
Note: On this map, 19 PvPs are played and 6 of them belongs to Shuttle.
Map 2: If coach park plays a terran on this map, I'll eat my hat. Or first, buy a hat and then eat it. Terran winrate on this map is below 37% for both non-mirror matchups. ZvP is above %50 as well, but not against Bisu. He won his both PvZs on this map and PvT here is just ezpz. Even BeSt won against a ZvP sniper, namely HoeJJa on this map. But I don't think that I'll see BeSt here, because coach Park will be smart enough to think that Calm would kick the shit out of BeSt. But, there is another thing: Kal has won against Bisu on this map in this season. The player actually depends on the outcome of the first game, but still we shouldn't ignore the probability of an SKT zerg popping out. So, against an SKT zerg or BeSt, Calm would be a good idea. But against Bisu, Calm would be a bad idea. But, Bisu played more on this map than all SKT zergs combined. So, this should be the battle of aces. Kal should be sent on.
Map 3: This is the map for Fantasy. I doubt that he can make himself a better map. Terran has more than 50% winrate in both non-mirror matchups and TvZ is 15-6. I don't think that coach Park will change his attitude or overcommit himself to thinking and he'll just send Fantasy on. Bogus is actually worth trying here.
Map 4: This map is actually a Zerg map, just like Aztec. And Terran would be a stupid choice, just like on Aztec. I just have no idea about what to do right now. Just nothing. But, SKT tends to send Zergs on this map. So, Modesty or Calm actually have good chances. I send Modesty here.
Map 5: Even though this map looks like a Terran map, people send zergs here. But, Fantasy played 2 TvPs recently on this map, but still SKT coach would assume that there will be no TvPs here. And Fantasy has a high probability to already have played up to this point. If he really has played, SKT will send either another zerg or Canata. So, I'll send either Classic or Calm. But I'm pretty much undecided.
Map 6: This map looks like a Terran map as well. And, if canata hasn't played on Map 5, I'm sure he'll be here. Classic-Calm again.
Map 7: Kal. Definitely.
|
i don't really disagree with your player choices for stx, but for skt i definitely do.
central plains is going to be either best or by.sun. you claim that best has no other possible map, but i believe park will be heavily considering sending by.sun here and best on fortress.
aztec is very likely to be bisu. i'd give it 60% chance to be bisu. the remaining 40% i'll say will be s2, soo, or kwanro
fortress could be either best or fantasy. if park plays it the way i'd play it, this would be best
icarus is a tricky map, but i like the idea of sending kwanro here, or soo. i would not trust canata here. bisu too if he wasn't on aztec
benzene i still believe is completely terran favored. i'd save fantasy for this set. but i wouldnt be surprised to see a zerg or canata, or even ssak
i hate the idea of sending protoss on grand line. zerg would be my choice.
By.Sun <Central Plains>
Bisu <Aztec>
BeSt <Fortress>
Kwanro <Icarus>
Fantasy <Benzene>
s2 <Grand Line SE>
Bisu <Empire of the Sun>
so if i were skt and you were stx we'd have:
By.Sun <Central Plains> Shuttle
Bisu <Aztec> Kal
BeSt <Fortress> Bogus
Kwanro <Icarus> Modesty
Fantasy <Benzene> Classic/Calm
s2 <Grand Line SE> Classic/Calm
Bisu <Empire of the Sun> Kal
i can't call ZvZ or PvP, but this is the likely situation. Bisu/Kal in the MSL wildcard were the results of not taking it very seriously. and no, my lineups are not an attempt to snipe every one of your players
|
Wow, this turned into something that I'd really want it to turn into. Actually even if you don't include the wildcard results, Kal is 3-2 against Bisu and 3-0 in last 3 games. About sending BeSt on Fortress, yeah. This map is the one that he owned fOrGG on last year's PL finals. But, he also lost to FBH on this map in SKT's last PL match, and STX Terran line up is doing good right now and they definitely have better TvPers than FBH, but they are not even close to fOrGG's TvP.
But, about thinking all of these, you got a point. Throwing BeSt against Shuttle is less reliable than throwing him against an STX Terran. But again, Park may go for all-in and send BeSt on Shuttle and Canata on Fortress, or Fantasy on Fortress, Canata on Benzene. Still, you got better choices than Park.
|
On November 09 2010 10:17 Impervious wrote:Show nested quote +On November 09 2010 09:53 CanucksJC wrote:
Pretty entertaining to read... pretty pointless otherwise Therefore, the ideal way to set your lineup is to do it where you have 1v2, 2v3, 3v4, and sac the 4v1 to likely win the series in a 3-1 victory. Worst case scenario, you end up losing one of the 3 games you should win, and you get a 2-2 with a 50% shot at winning anyways.
lol
On November 09 2010 10:32 CanucksJC wrote:
What I meant is, stuff like this is pretty cool, but I could say anything about a given team and players and manipulate the results so what I say makes sense.
ex) STX is probably going to send out Kal on this map but lolol I has Jaedong, he's going to sweep ya'll, Kal, Calm, and whoever that follows in that order.
=/
and that would make sense how
|
On November 10 2010 12:59 Metalwing wrote:Look at spoilers for more detailed "logical forecast": + Show Spoiler +The statement of: "If Jaehoon can beat Horang2 with a strategy that Movie has just lost to in a PvP against Tyson, he would lose to Jaehoon as well" seems logical to me, what do you think?
Or, "if HyuN beats Hydra in a ZvZ, he could own HoGiL so badly that HoGiL wouldn't be able to sit for 3 days" assumption doesn't seem too insane either, agree? Others were just particular examples about fictionary games. + Show Spoiler +In a world that Jaehoon beats Horang2 in a televised PvP, by.hero can beat Killer or Calm can beat Jaedong or HiyA can beat Kal. All these 4 cases carry exact same oddness imo. Or "if a rookie snow can beat Flash in an ace game, a well trained soulkey can do it as well" statement also seems logical.
what the hell is all this. you are so stupid. you just spent that much text saying that flash can beat mind because mind got beaten by lomo and flash beat lomo.
one hell of an assumption. and turn does not have the capabillity of beating (insert any player)?
User was temp banned for this post.
|
|
|
|
|
|
EPT
