EPTThe Deepest Team in Proleague - Page 2
| Forum Index > BW General |
|
TheNessman
United States4158 Posts
| ||
|
Caihead
Canada8550 Posts
CJ (and to a lesser extent, SKT and Team 8) puts too much expectations on its players being able to play vs any match up on any map, they are sending players dependent on the score / momentum of the match rather than preparing them for specific map snipes. Nearly all of their players have played on every map / match up. When you consider this, CJ / SKT still has the "deeper" line up considering that they are still able to pull out wins vs map / race snipers; but Khan definitely has better coaching and decision making which is much more important in a pro-league (non winners league) format. | ||
|
Bagration
United States18282 Posts
| ||
|
Count9
China10928 Posts
| ||
|
Louuster
Canada2869 Posts
| ||
|
Xiphos
Canada7507 Posts
KHAN is the nimble owl, carefully stalking their prey while camouflaging with the darkness in making up for their lack of muscle. The bird takes extra time to analyze the playing field to determine the best angle to strike at the opposition. This risky mindset can also lead to fatal results of getting humiliated by solid teams. They just have to make sure that they don't get caught with their pants down. Looks like we got some ying & yang thing going on here. | ||
|
Hesmyrr
Canada5776 Posts
+ Show Spoiler + ![[image loading]](http://i39.tinypic.com/33w1lon.png) deeper and deeper... T_T | ||
|
TheNessman
United States4158 Posts
On February 23 2012 10:45 Bagration wrote: Cool analysis there. For some reason I expected a poll, not sure why. thats interesting hahah maybe it would've been a poll long , but then everyone had to explain there answers, but i don't think everyone would have written as in depth answers as this guy  | ||
|
VGhost
United States3616 Posts
On February 23 2012 10:33 TheNessman wrote: i don't understand that graph >< The short (and not precisely accurate) version: The numbers I ran give a value for each player's record which can be thought of as approximating an "equivalent" undefeated record. Flash is the only one I included here whose 13 wins translate directly to a value of 13, because he hasn't lost. Contrast BeSt, whose 9 wins are offset by 7 losses... but those 9 wins really do matter, so he rates out around 5. The orange "players" graphs represent the sum of all players' values on the given team. But the biggest problem is that my number just generates "0", no matter how many games you lose, until you win a game, and then no matter how bad you are you get a rating that's at least a little positive. In reality, the record of players like Grape (0-4) or HoeJJa's (2-5) have what's really a negative effect on the overall team record. So the team bar (blue) is shorter; how much shorter relatively indicates I'm not really sure what, but I offered some guesses. My current hypothesis is that the difference reflects how well teams manage whatever players they do have. If you're still curious, PM me and I can give you a fuller rundown of what I'm calculating, why I think it's reasonably accurate, and what I think it "means". | ||
|
pappelee
China34 Posts
| ||
|
ChriS-X
Malaysia1374 Posts
| ||
|
shaftofpleasure
Korea (North)1375 Posts
+ Show Spoiler + in the rankings -_- | ||
|
Heimatloser
Germany1494 Posts
stupid and biased post. edit: second paragraph just reflects proleague results, too. non more and non less third paragraph makes me angry, since i have no clue wtf you talking about. so you are comparing "the number"? which number? 42? also the, graph, what does it mean? you could at least give us some hints so that non psychic ppl also can follow your thought, like naming the god-damn axes... | ||
|
Bill Murray
United States9292 Posts
| ||
flamewheel
FREEAGLELAND26781 Posts
| ||
|
intotheheart
Canada33091 Posts
| ||
|
JohannesH
Finland1364 Posts
Take Khan for example, they're not deep because Stork, Turn, Jangbi and Brave perform so well. They're deep because they've still got Reality, Shine, Roro (+others still) to back them up. | ||
|
Soft`Soap
Canada865 Posts
I'm inclined to think that it's not a coincidence the top four teams right now are the four teams with the four best players in the game Very nice analysis but I just gotta pick on this it's not a coincidence, there is a direct correlation between the top four teams and the four players with the best record in pro league. This is because the wins from a player help a team win. | ||
|
VGhost
United States3616 Posts
On February 26 2012 20:17 Heimatloser wrote: If you would have done the analysis when kt was on top, they would be the deepest team. Possibly. They came in #2 anyway. On the other hand, Stars (#7 here) were solidly in 4th when this was written and have to be at least in 5th still despite some epic failures. On February 26 2012 20:17 Heimatloser wrote: Second paragraph just reflects proleague results, too. non more and non less. Yes, I mentioned that so that people could factor that into how seriously they were willing to take the results. Though what else are we supposed to go on right now? On February 26 2012 20:17 Heimatloser wrote: ...I have no clue [what] you [are] talking about. So you are comparing "the number"? Which number? 42? Also the graph, what does it mean? You could at least give us some hints so that non psychic ppl also can follow your thought, like naming the [] axes... The number in the initial 18 players I listed is a number generated by a calculation based on wins and games and is right there next to each player's name, with their stats. For the graph, I initially said, "...By comparing the number for the whole team to the numbers put up by each player..." so that the numbers on the graph are on that same scale of calculation. The orange bar is the sum of player values for each team, and the blue is the overall team value, as I explained here when someone else asked the same question. On February 26 2012 20:32 Bill Murray wrote: I would imagine you should analyze number of players fielded in relation to the % shared regarding playtime, while also balancing it out with overall efficiency I'm sure this would improve accuracy in calculating how well a team is managed, but I haven't the vaguest idea how to weight these things and am not sure I want to spend the time working it out. Most teams follow the same pattern anyway: 2-4 players who play almost every match, and then a bunch who rotate through based on maps, recent play, etc. There are sort of two extremes: ACE, and to a lesser extent KHAN and CJ, play almost everybody about equally; on the other end T8 (also Stars and SKT) play the same players pretty much all the time. STX and KT (except Flash) are sort of in the middle. But ACE still plays Kal almost every time and T8 still rotates Killer and Tyson: I called them "extremes" but I'm not sure that term is actually justified. On February 27 2012 01:05 JohannesH wrote: I don't know if that really is a good measure of depth - having many players place in top 18 is not really a measure of that since having plenty of players with a decent ~50% winrate are what make a deep team. This just tells which has the best performing top players. Take Khan for example, they're not deep because Stork, Turn, Jangbi and Brave perform so well. They're deep because they've still got Reality, Shine, Roro (+others still) to back them up. Yes and no. A team - especially in a short match format like this season's Bo5, though not as much as in a strong man format like Winners League - can't really get by with a bunch of mediocre players. This year's example is, so far, STX. I take depth to mean, "How far down do you have good players?" and we can talk about comparative depth as answering a question like, "If we play my #4 against your #4, who wins?" If KHAN can send Brave up against Canata, for instance, it's bad for the team fielding Canata. Not that upsets don't happen, but in general. And of course the manager of a team is trying to manufacture matchups where Leta plays against Calm and not Shuttle, while Hydra plays Shuttle and not Bogus, and so forth, so we rarely test the "#4 vs #4" proposition - but it's still a big difference in the long run that CJ's number four is Movie and STX's is hyvaa. On the other hand, a deep team is going to have more good players. When you have more good players not only do those players do well, they do better than they would do against "equal" competition because it's easier for a competent coach to set up more games where Stork's not playing ZerO and winning a tight game but playing Flying and winning without breaking a sweat - padding stats, as it were. The result is that going just off records (which I did here) the deeper team will put more players into the "top results" category. | ||
|
NB
Netherlands12045 Posts
The scale looks like this: Ace>Decent>sniper. This mean that when they goes up vs KT, they will do their best to avoid jangbi or stork vs flash and use these to snipe decent players like stats instead. Mean while they keep sending out rookie to snipe other team ACE to : 1/promote rookie and give them stage exp. 2/Avoid spending good players on unavoidable loses or uncertain win. This explain why their players are getting such good w/l record: they were drafted to win, not to 'claim' the win. | ||
| ||
