|
On November 04 2010 10:19 Cloak wrote:Show nested quote +On November 04 2010 04:37 EriktheGuy wrote:
Not really significant.
Given complete balance, the chance of no toss in the final 4 is about 20%. So, since they were precluded twice in a row, that's a 1/25 chance. Which is greater than the 95% confidence that statistics and the sciences use.
Good point, I wasn't taking additional GSLs into consideration. But still, you have to calculate the probability that this could happen to ANY race. This thread could exist after all, if terran or zerg are underrepresented in the GSL. This gives about a 3/25 chance (not exactly, but a good approximation) of one race being down to 1 player or less by Ro two tournaments in a row. That's about 12%, not very rare, and within the 95% confidence interval you suggested.
I'm being nit-picky here since it's my job, but statistics/science does not 'use' the 95% confidence interval. It's arbitrary what interval you use, and 95% is just a common one.
|
On November 04 2010 17:24 EriktheGuy wrote:Show nested quote +On November 04 2010 10:19 Cloak wrote:On November 04 2010 04:37 EriktheGuy wrote:
Not really significant.
Given complete balance, the chance of no toss in the final 4 is about 20%. So, since they were precluded twice in a row, that's a 1/25 chance. Which is greater than the 95% confidence that statistics and the sciences use. Good point, I wasn't taking additional GSLs into consideration. But still, you have to calculate the probability that this could happen to ANY race. This thread could exist after all, if terran or zerg are underrepresented in the GSL. This gives about a 3/25 chance (not exactly, but a good approximation) of one race being down to 1 player or less by Ro two tournaments in a row. That's about 12%, not very rare, and within the 95% confidence interval you suggested. I'm being nit-picky here since it's my job, but statistics/science does not 'use' the 95% confidence interval. It's arbitrary what interval you use, and 95% is just a common one.
If it was your job you would also know because player skill and human mistakes introduce a large amount of variance, and as such, any result from any tournament, unless you have a huge sample of tournaments to account for all that variance(which even if we took all the tourneys to date with the current patch, it would still be arguable that it is not enough), a statistical approach is a waste of time.
|
On November 04 2010 17:24 EriktheGuy wrote:Show nested quote +On November 04 2010 10:19 Cloak wrote:On November 04 2010 04:37 EriktheGuy wrote:
Not really significant.
Given complete balance, the chance of no toss in the final 4 is about 20%. So, since they were precluded twice in a row, that's a 1/25 chance. Which is greater than the 95% confidence that statistics and the sciences use. Good point, I wasn't taking additional GSLs into consideration. But still, you have to calculate the probability that this could happen to ANY race. This thread could exist after all, if terran or zerg are underrepresented in the GSL. This gives about a 3/25 chance (not exactly, but a good approximation) of one race being down to 1 player or less by Ro two tournaments in a row. That's about 12%, not very rare, and within the 95% confidence interval you suggested. I'm being nit-picky here since it's my job, but statistics/science does not 'use' the 95% confidence interval. It's arbitrary what interval you use, and 95% is just a common one.
Well, that's what the vast majority of methods use to show that something is statistically significant. The only thing off the top of my head that doesn't use 95% is beta error correction. It's arbitrary, but if it satisfies the standard by which we hold our scientific findings to, that holds a lot more clout as an argument. Right with the 12% but I got to nitpick. Any one given race with a double loss would have 3x the permutations, but we're talking about Protoss. What is the probability that Protoss lost twice in a row due to normal drift? 1/25. The difference is that Terran and Zerg aren't close to being knocked out in the Ro4 twice in a row. Especially not Terran.
|
Well, its only the second GSL. And we are most likely looking at a Terran winner. It would be my prediction that a protoss will win the 3rd GSL. And because of that tournament wins will seem pretty balanced.
|
On November 04 2010 18:14 adeezy wrote:
It would be my prediction that a protoss will win the 3rd GSL.
What are you even basing this off? Even if 3 race is completely balanced in every single aspect you can't make a prediction like that.
|
On November 04 2010 15:36 dbizzle wrote:
As for GSL, I watched all the protoss matches, and only a few games where the protoss went standard robo colossi build. Everyone tried doing DT, VR play, which I don't get why people are trying fancy builds as every solid top player are doing standard builds. PvP is a killer too, the main reason i switched from P to Z, I hate pvp with a passion and its pretty lame. Overall I think bigger maps is the key for Protoss for doing better.
Awsome, so you are saying everything that isn't robo colossus is "fancy" - THANK YOU for proving my/our point, exactly that's the problem. You can either play ridiculously predictable with gateway/robo OR you can play some sort of cheesy, all-in-ish stuff. Completely the opposite to terran/zerg who got many possible "standard" ways to play their match-ups (in PvZ muta-tech and roach/hydra/corruptor are basicly equally viable and still completely different; don't wanna get started on terran)
You also nearly got me here with bigger maps being the key for protoss success. On bigger maps protoss in its current stance would suffer even more because Terran could go crazy with their expansions and Protoss would have no way to defend all the possible drop-locations. PvT is currently the easiest on maps with a close natural and really hard to take thirds...this is why we see protoss doing so good on xel naga, terran can't abuse their army-advantage in midgame because gold ist hard to take/hold even with PF and the expos in the back are hard to defend as well.
|
Protoss in itself is fine (or close to) imo, they can still win tournament : NEXGenious in the blizzcon and HuK in the EG masters cup. And they need a little bit of time to figure out a way to play against the new roach.
You cant compare those things to GSL, they are at a different level as evidenced by your examples. Nexgenius was one of two koreans in the blizzcon tournament, of course he won, not much competition, didnt even have to play Maka. IdrA came into the last MLG unseeded and destroyed everyone including a 2-0 of HuK and he just lives in Korea, hasnt even made the top 8 of GSL. Its when these protoss reach the top 8 that these problems are cropping up.
|
If you just replace the word protoss in this thread with zerg you have all the balance threads looked a month ago. Just a bit funny.
|
On November 04 2010 19:08 nam nam wrote:
If you just replace the word protoss in this thread with zerg you have all the balance threads looked a month ago. Just a bit funny.
playxp jinjing agrees - protoss-QQ is showing a steady rise, with currently being at 73%
zerg maxed out at 80%, gogogo korean protoss-whiners, you can do it!
http://www.playxp.com/sc2/jingjing/
|
i just found the best solution to make protoss balanced
there are already 2 different models for DTs. make the one with 1 blade be a normal DT and the one with the 2 blade be a MARAUDER ASSASSIN, who deals 0 +126 to units that are both armored and biological
also, make that DT cut off the heads of the rauders and decorate your nexus with them
|
On November 04 2010 20:49 sleepingdog wrote:Show nested quote +On November 04 2010 19:08 nam nam wrote:
If you just replace the word protoss in this thread with zerg you have all the balance threads looked a month ago. Just a bit funny. playxp jinjing agrees - protoss-QQ is showing a steady rise, with currently being at 73% zerg maxed out at 80%, gogogo korean protoss-whiners, you can do it! http://www.playxp.com/sc2/jingjing/
The last time any race got over 70% it was 6 days in a row in the week before 1.1 for zerg. Sadly, that patch with possible P buffs (and some counteracting nerfs that may or may not be needed) won't be coming out until at least after MLG (and more likely not until after GSL finals if there's any non-Protoss changes). So if the % for P keeps up they'll set a couple records on that chart (they already broke the pure # of people voting for one race in a day).
|
On November 04 2010 20:49 sleepingdog wrote:Show nested quote +On November 04 2010 19:08 nam nam wrote:
If you just replace the word protoss in this thread with zerg you have all the balance threads looked a month ago. Just a bit funny. playxp jinjing agrees - protoss-QQ is showing a steady rise, with currently being at 73% zerg maxed out at 80%, gogogo korean protoss-whiners, you can do it! http://www.playxp.com/sc2/jingjing/
Could you please explain this site a bit? I cant read anything there but I am really interested
|
On November 04 2010 21:12 Sewi wrote:Show nested quote +On November 04 2010 20:49 sleepingdog wrote:On November 04 2010 19:08 nam nam wrote:
If you just replace the word protoss in this thread with zerg you have all the balance threads looked a month ago. Just a bit funny. playxp jinjing agrees - protoss-QQ is showing a steady rise, with currently being at 73% zerg maxed out at 80%, gogogo korean protoss-whiners, you can do it! http://www.playxp.com/sc2/jingjing/ Could you please explain this site a bit? I cant read anything there but I am really interested
There is (was?) a whole thread about this, try search, shouldn't be too difficult to find it
In a nutshell: every day everybody is allowed to vote on one race he likes to QQ about:
red: zerg
green: terran
blue: toss
purple: balanced
As you immediately see when you scroll down, the red bar was huge before the recent patch and lately the blue bar has grown drastically.
|
There were many pvp's this gsl, has to do with brackets.
|
On November 05 2010 00:18 ppshchik wrote:
There were many pvp's this gsl, has to do with brackets.
You mean, the only reason we got a P in the RO8 at all is because there was a mirror in the RO16? 
|
On November 05 2010 00:18 ppshchik wrote:
There were many pvp's this gsl, has to do with brackets.
If there were many PvP´s then there had to be on the other hand many TvT, ZvZ and TvZ which makes the whole equation of brackets beeing the reason for Protoss beeing eliminated early senseless. And if it is PvP u can be sure that one Protoss will advance.
|
On November 04 2010 18:22 sleepingdog wrote:Show nested quote +On November 04 2010 15:36 dbizzle wrote:
As for GSL, I watched all the protoss matches, and only a few games where the protoss went standard robo colossi build. Everyone tried doing DT, VR play, which I don't get why people are trying fancy builds as every solid top player are doing standard builds. PvP is a killer too, the main reason i switched from P to Z, I hate pvp with a passion and its pretty lame. Overall I think bigger maps is the key for Protoss for doing better. Awsome, so you are saying everything that isn't robo colossus is "fancy" - THANK YOU for proving my/our point, exactly that's the problem. You can either play ridiculously predictable with gateway/robo OR you can play some sort of cheesy, all-in-ish stuff. Completely the opposite to terran/zerg who got many possible "standard" ways to play their match-ups (in PvZ muta-tech and roach/hydra/corruptor are basicly equally viable and still completely different; don't wanna get started on terran) You also nearly got me here with bigger maps being the key for protoss success. On bigger maps protoss in its current stance would suffer even more because Terran could go crazy with their expansions and Protoss would have no way to defend all the possible drop-locations. PvT is currently the easiest on maps with a close natural and really hard to take thirds...this is why we see protoss doing so good on xel naga, terran can't abuse their army-advantage in midgame because gold ist hard to take/hold even with PF and the expos in the back are hard to defend as well.
Drops can be easier to defend with good placement observers and a few stalkers in every expansion. I feel like PvT is heavily in favor for P. If T goes mech build you are able to macro more, if T goes for a bio push, counter it with templar play (bigger maps = easier 2 base) then macro from there. And about xel naga, the map is bigger than standard maps but it still kinda small with small choke points for mech play.
|
On November 05 2010 00:57 dbizzle wrote:
Drops can be easier to defend with good placement observers and a few stalkers in every expansion. I feel like PvT is heavily in favor for P. If T goes mech build you are able to macro more, if T goes for a bio push, counter it with templar play (bigger maps = easier 2 base) then macro from there.
The thought of "just squeezing out the templar tech and taking it from there" is very sweet when basicly the whole PvT is about if I CAN get my templar-tech without being either dead or terribly behind.
Also randomly leaving some observers (100 frickin gas each) and units everywhere is so naive....I don't even....
Bigger maps means much more succesful drops because you just "can't" protect all locations at once. Logically impossible, if I spread my units out evenly 25/25/25/25 to four locations, the one location terran is going to hit will be dead, because 25% of the units aren't gonna cut it.
Also bigger maps means more necessary observers to keep track of all possible drop-paths (that's what YOU suggested should protoss do...it's not really viable due to the heavy gas-investment that observers are anyways)
|
The reason is twofold- skill and some semblance of balance. Before you flame, I do NOT mean I think its imba.
Firstly, nexgenius won blizzcon and still isn't anywhere near the level of the top gamers. SSKS on the other hand is. Look through the vP wins, most players later in the tournament tell you they train with him. When you train with the best protoss, you will beat lesser ones.
With that being said, the second problem is protoss predictability. There are not very many viable openings, as too many toss openers leave you vulnerable to glaring matchup flaws (large map T drop harass comes to mind).
One of three things will happen
1. Someone will come up with a new safe early opener for protoss and the entire problem is solved (imagine as T not automatically knowing no FE and no 4gate equals 2gaterobo?)
2. This will continue for a few more weeks of protoss imba rage, and blizzard will make an unnecessary patch
3. sSKS comes back to qualify for GSL3 and finally puts this senseless argument to rest.
For the record, I main protoss. Its not imbalanced, we just still suck.
|
One base straight high templars tech: the T will over-expand you.
Two base straight high templars tech: high probability of dying to 3 rax or 2 starports openings before the templar tech kicks in.
|
|
|
|
|
|
EPT
