|
19533 Posts
On October 21 2014 16:46 Thieving Magpie wrote:Show nested quote +On October 21 2014 16:42 The_Red_Viper wrote:On October 21 2014 16:37 Thieving Magpie wrote:On October 21 2014 16:25 The_Red_Viper wrote:On October 21 2014 16:23 Thieving Magpie wrote:On October 21 2014 16:21 The_Red_Viper wrote:On October 21 2014 16:18 Thieving Magpie wrote:On October 21 2014 16:14 playa wrote:On October 21 2014 15:59 Gwavajuice wrote:On October 21 2014 05:11 DinoMight wrote:
[quote]
Comments like this should get a warning. Those are exactly the type of comments we don't need in this forum. They add absolutely nothing to the conversation at all and create a poisonous air of balance whining and "dead gaem."
Rather, let's have a real conversation.
Your comment is also 100% false. There are many ways to play PvZ and the Zerg certainly never gets a free win. 100% agree but I guess this thread has just became a battle.net sub forum now. I wish we could come back to what the OP intended it to be.... When your race is at 44% in a matchup and the win rates via game length reflect Toss winning the majority of their all-in games, well, that 44% becomes even less when talking about macro games. Fact: there are not many ways at all to consistently win macro games. The math doesn't add up. There's either not many viable ways of playing or a lot of players are throwing games. This is common sense. This isn't some kinda conspiracy. 50% winrate (+/-) 6% is damn even. No it really is not What do you consider acceptable margin of error? Hard to say, but 6% points is pretty huge imo It really shouldn't be a hard thing to know. What is comfortable or you as a margin of error. You either know what you think it should be or you don't know what you think it should be. You, for example, don't know what you think it should be. To me, 10% is too high, and 5%-9% is acceptable, 0-4 is too close to call. Unless we have definitive terms on what we want achieved we will never be able to discuss anything. We need to know what numbers is acceptable and then begin discussion with the acceptance of those numbers. Otherwise we end up stuck arguing semantics when we could be arguing specifics. 56:44 is already over 10% off. So i guess it is too high for you as well? It shouldn't be higher than 53:47 imo, and even that is already somewhat questionable. Don't invent fake math please. 56:44 is +\- 6 53:47 is +\- 3 60/40 is +\- 10 According to you, a 3% variance is too high? Because that sound stupid.
Yes +/- 6 percentage points. Which is 12% off.
|
On October 21 2014 16:44 pure.Wasted wrote:Show nested quote +On October 21 2014 16:07 Thieving Magpie wrote:On October 21 2014 16:02 pure.Wasted wrote:On October 21 2014 15:04 Thieving Magpie wrote:On October 21 2014 10:08 pure.Wasted wrote:On October 21 2014 09:50 Thieving Magpie wrote:On October 21 2014 05:15 DinoMight wrote:On October 18 2014 04:02 Thieving Magpie wrote:
How much mins does Protoss normally bank?
Like, when Zerg and toss are banking money, and Zerg tries to bank 4k/4k, can toss just bank 0/4k and spend 4k on cannons?
Then the army wipe happens, you start remixing on tech units like temps and Phoenix's while Zerg has to fight through 4k worth of cannons? Problem with that is that Zerg banks at a faster rate from having more bases throughout the game. A good Protoss never really banks minerals because they're constantly trying to destroy the Zerg econ. You can't make any of the "good" Protoss units instantly (Colossus, Carriers, Void Rays, etc.) but you CAN remax on any Zerg unit in one build cycle. So Zergs are trying to bank minerals while Protoss are trying to constantly produce and harass. Banking just doesn't really benefit Protoss. You make a good point about base numbers. A "bank equal amounts" argument requires the assumption of equal economic strength. This is not the case in any of the non-mirror matchups. Then a better discussion point should be this: What economic ratio per matchup should we make the jumping point on balance discussions. The simplest benchmark to use would number of bases. For example, in mech vs bio X:X+[1-2] is the ratio where balance discussions hinges on. Where say "assuming bio makes this much" and "mech is this efficient" we can create a baseline for discussion. "Tanks are not too strong because bio has enough expendable resources to act as fodder for the initial volley" "Marauders are not OP since bio is mineral starved due to having to waste troops on tank volleys" Etc... What economic benchmark should we set as a balanced late game scenario in PvZ? Is X:X+1 balanced? Is X:X-Y balanced? Is 4:X balanced? Etc... Until we have consensus as to what constitutes an even playing field, we will never be able to discuss individual unit stats. Meta shifts completely invalidate that approach to game balancing. The only way you can use something like "X:X+1" as a benchmark for unit balance is if there is, and will always be, only one way to play every MU. It's actually not that difficult to keep SC2 MUs in the 45/55 range, I feel. There are enough units and upgrades to allow an infinite combination of buffs and nerfs. The trouble is defining "balance" to begin with. I'll invent a new sport to demonstrate. In my made up sport, Competitor A has to juggle 10 balls while Competitor B has to juggle 1 ball. Drop the balls and you lose. In this phase of the competition, Competitor A has only a 10% winrate. If after five minutes, all 11 balls are still in the air, Competitor A gets to roll a ten sided die. If it lands on any number but "1," he wins the game. If it lands on 1, Competitor B wins instead. In this phase of the competition, Competitor B has only a 10% winrate. Both competitors have an equal chance of winning. Is the game balanced? If you mean "are the winrates balanced?" then certainly. But the sport is so poorly designed, who cares if it's 50-50? One player has to work much harder for his win in the first phase, while the other has absolutely nothing to do once the competition enters the second stage. Whether he wins or loses is, at that point, completely out of his control. All he can do is hope that the die roller "messes up." No situation in SC2 has ever been so cut-and-dry as in my invented sport, but winrate balance is such a small part of the big picture, more often than not it's just a red herring. The bench mark is arbitrary. Notice I did not talk about win rate, I talked about being able to have a context in which to discuss unit interaction and gameflow dynamics. Sure metagames can always shift (and they should) but that is not the discussion. The discussion is very simply "assuming baseline A, does unit B's interaction with unit C produce the results we want it to have in an idealized scenario--and then maps and metagame shifts move in reference to that idealized scenario." Here's an example. 50 minerals is the standard cost of a basic unit (worker/tier1) Zerg gets 2 units Terran gets 1 unit Protoss pays twice as much for its unit. All 3 units (zergling/marine/zealot) are units in reference to the basic unit (the worker) and its design is based off of that baseline. By having that baseline to frame the discussion, we get to discuss how the game is expected to play out and not simply discuss how the games conclude. We should not care who wins, we should only care what interactions happen. Economic considerations are only one of many (more complex, I think) factors that influence unit interactions. Nothing in the ratio you mention helps us to predict that Zerglings will have interesting dynamics when fighting units like Hellions, WMs, and Banelings, interactions where split-second decision making and mechanical skill are rewarded, but are reduced to "A-move" units in most other scenarios in the game. If the Zergling's role changes radically throughout the game depending on what units it's fighting, its balance becomes much harder to pin down. Marines are hardly OP without Bunkers, Stim, shields, Medivacs, Terran splash, meatshields, their ability to clump... but all of those fundamentally change how the Marine behaves in the game, to the point where if Protoss didn't have Colossi, they'd be right fucked. I don't see how an economy-centric analysis of the game can possibly tell us at what point the Marine becomes OP versus Protoss. Maybe I'm overlooking something, or trying to use your thought process to do something it wasn't meant to do. Economic analysis tells us how many units each side has. Before we can talk about how unit A interacts with unit B we should discuss how many unit A's are interacting with unit B's. Sending 1 zergling vs 200 marines tells us nothing. Sending 400 zerglings vs 1 marine tells us nothing. Economic analysis tells us how many lings to expect vs how many marines to expect. Could you make some kind of statement re: the state of the game, so I can see your philosophy in action? Right now it just sounds like you're saying "how much units cost is important," which is hard to disagree with but also... not very useful. I'm sure you're saying something more nuanced than that, and an illustration would really help!
Let's take the muta tech switch that people whined about a few pages ago.
Before we can discuss whether or not tech switching to Mutalisks is OP, we must first discuss what we consider The bench mark number of metas that is considered appropriate. Am we don't consider the tech switch OP unless the Zerg player makes that many Mutalisks. Why? Because until we have that as a bench mark, we won't actually have an idea how to fix the problem because we don't have a solution to aim for.
|
On October 21 2014 16:50 The_Red_Viper wrote:Show nested quote +On October 21 2014 16:46 Thieving Magpie wrote:On October 21 2014 16:42 The_Red_Viper wrote:On October 21 2014 16:37 Thieving Magpie wrote:On October 21 2014 16:25 The_Red_Viper wrote:On October 21 2014 16:23 Thieving Magpie wrote:On October 21 2014 16:21 The_Red_Viper wrote:On October 21 2014 16:18 Thieving Magpie wrote:On October 21 2014 16:14 playa wrote:On October 21 2014 15:59 Gwavajuice wrote:
[quote]
100% agree but I guess this thread has just became a battle.net sub forum now.
I wish we could come back to what the OP intended it to be....
When your race is at 44% in a matchup and the win rates via game length reflect Toss winning the majority of their all-in games, well, that 44% becomes even less when talking about macro games. Fact: there are not many ways at all to consistently win macro games. The math doesn't add up. There's either not many viable ways of playing or a lot of players are throwing games. This is common sense. This isn't some kinda conspiracy. 50% winrate (+/-) 6% is damn even. No it really is not What do you consider acceptable margin of error? Hard to say, but 6% points is pretty huge imo It really shouldn't be a hard thing to know. What is comfortable or you as a margin of error. You either know what you think it should be or you don't know what you think it should be. You, for example, don't know what you think it should be. To me, 10% is too high, and 5%-9% is acceptable, 0-4 is too close to call. Unless we have definitive terms on what we want achieved we will never be able to discuss anything. We need to know what numbers is acceptable and then begin discussion with the acceptance of those numbers. Otherwise we end up stuck arguing semantics when we could be arguing specifics. 56:44 is already over 10% off. So i guess it is too high for you as well? It shouldn't be higher than 53:47 imo, and even that is already somewhat questionable. Don't invent fake math please. 56:44 is +\- 6 53:47 is +\- 3 60/40 is +\- 10 According to you, a 3% variance is too high? Because that sound stupid. Yes +/- 6 percentage points. Which is 12% off.
You really have no idea how margins of error works do you?
Here's how it works in competitive anythings.
50% is ideal. +\- meta shifts, +\- skill gap, +\- correlative trends
There will always be margins of error. What is the amount you consider acceptable. Saying +/- 6 is 12% off is a meaningless phrase since it is as often +6% as it is -6% which averages out to 50% over infinite iterations. So what is the margin of error you consider fair and please stop making up Maths that don't add anything to the discussion.
|
19533 Posts
On October 21 2014 16:56 Thieving Magpie wrote:Show nested quote +On October 21 2014 16:50 The_Red_Viper wrote:On October 21 2014 16:46 Thieving Magpie wrote:On October 21 2014 16:42 The_Red_Viper wrote:On October 21 2014 16:37 Thieving Magpie wrote:On October 21 2014 16:25 The_Red_Viper wrote:On October 21 2014 16:23 Thieving Magpie wrote:On October 21 2014 16:21 The_Red_Viper wrote:On October 21 2014 16:18 Thieving Magpie wrote:On October 21 2014 16:14 playa wrote:
[quote]
When your race is at 44% in a matchup and the win rates via game length reflect Toss winning the majority of their all-in games, well, that 44% becomes even less when talking about macro games. Fact: there are not many ways at all to consistently win macro games. The math doesn't add up. There's either not many viable ways of playing or a lot of players are throwing games. This is common sense. This isn't some kinda conspiracy. 50% winrate (+/-) 6% is damn even. No it really is not What do you consider acceptable margin of error? Hard to say, but 6% points is pretty huge imo It really shouldn't be a hard thing to know. What is comfortable or you as a margin of error. You either know what you think it should be or you don't know what you think it should be. You, for example, don't know what you think it should be. To me, 10% is too high, and 5%-9% is acceptable, 0-4 is too close to call. Unless we have definitive terms on what we want achieved we will never be able to discuss anything. We need to know what numbers is acceptable and then begin discussion with the acceptance of those numbers. Otherwise we end up stuck arguing semantics when we could be arguing specifics. 56:44 is already over 10% off. So i guess it is too high for you as well? It shouldn't be higher than 53:47 imo, and even that is already somewhat questionable. Don't invent fake math please. 56:44 is +\- 6 53:47 is +\- 3 60/40 is +\- 10 According to you, a 3% variance is too high? Because that sound stupid. Yes +/- 6 percentage points. Which is 12% off. You really have no idea how margins of error works do you? Here's how it works in competitive anythings. 50% is ideal. +\- meta shifts, +\- skill gap, +\- correlative trends There will always be margins of error. What is the amount you consider acceptable. Saying +/- 6 is 12% off is a meaningless phrase since it is as often +6% as it is -6% which averages out to 50% over infinite iterations. So what is the margin of error you consider fair and please stop making up Maths that don't add anything to the discussion.
Nevermind, not enough pashun to argue with you
|
I would like to remind people that Blizzard has said they aim to step in and fix any matchup once it is at 45% or worse, over a meaningful time frame. That's when it seems clear to them that something is not only imbalanced but outside of their idea of "acceptable imbalance."
When it comes to P vs Z, I really think you have to look at how the games are being won and lost; what other matchup has such a history of all-ins? If Toss won 80% of the time they decided to all-in and 30% of the time they tried to play a macro game, yet the numbers were around 50%, overall, would this mean nothing needs to change? If this were the case, people would still be in here claiming the macro games are fine, even though clearly they aren't.
At the moment, I'm of the opinion that mutas are overpowered if you don't have a lot of stargates, already, as part of your build. That's not to say that Toss isn't, in fact, OP, if they do already have them. I'm of the opinion of if you want to raise the numbers in macro games, for Toss, then you either need a new unit or you need to buff something, like cannons.
Either that, or everyone gravitates towards playing the same way, which, while it might raise the win numbers, it would probably not be ideal for viewers.
After seeing CJ Heros interview, where he talked about how clueless and lost he is versus Zerg, odds are there aren't that many viable ways of playing.
|
|
|
On October 21 2014 16:56 Thieving Magpie wrote:Show nested quote +On October 21 2014 16:50 The_Red_Viper wrote:On October 21 2014 16:46 Thieving Magpie wrote:On October 21 2014 16:42 The_Red_Viper wrote:On October 21 2014 16:37 Thieving Magpie wrote:On October 21 2014 16:25 The_Red_Viper wrote:On October 21 2014 16:23 Thieving Magpie wrote:On October 21 2014 16:21 The_Red_Viper wrote:On October 21 2014 16:18 Thieving Magpie wrote:On October 21 2014 16:14 playa wrote:
[quote]
When your race is at 44% in a matchup and the win rates via game length reflect Toss winning the majority of their all-in games, well, that 44% becomes even less when talking about macro games. Fact: there are not many ways at all to consistently win macro games. The math doesn't add up. There's either not many viable ways of playing or a lot of players are throwing games. This is common sense. This isn't some kinda conspiracy. 50% winrate (+/-) 6% is damn even. No it really is not What do you consider acceptable margin of error? Hard to say, but 6% points is pretty huge imo It really shouldn't be a hard thing to know. What is comfortable or you as a margin of error. You either know what you think it should be or you don't know what you think it should be. You, for example, don't know what you think it should be. To me, 10% is too high, and 5%-9% is acceptable, 0-4 is too close to call. Unless we have definitive terms on what we want achieved we will never be able to discuss anything. We need to know what numbers is acceptable and then begin discussion with the acceptance of those numbers. Otherwise we end up stuck arguing semantics when we could be arguing specifics. 56:44 is already over 10% off. So i guess it is too high for you as well? It shouldn't be higher than 53:47 imo, and even that is already somewhat questionable. Don't invent fake math please. 56:44 is +\- 6 53:47 is +\- 3 60/40 is +\- 10 According to you, a 3% variance is too high? Because that sound stupid. Yes +/- 6 percentage points. Which is 12% off. You really have no idea how margins of error works do you? Here's how it works in competitive anythings. 50% is ideal. +\- meta shifts, +\- skill gap, +\- correlative trends There will always be margins of error. What is the amount you consider acceptable. Saying +/- 6 is 12% off is a meaningless phrase since it is as often +6% as it is -6% which averages out to 50% over infinite iterations. So what is the margin of error you consider fair and please stop making up Maths that don't add anything to the discussion.
vnm: don't want to prolong pointless discussions.
|
On October 21 2014 17:11 playa wrote:
I would like to remind people that Blizzard has said they aim to step in and fix any matchup once it is at 45% or worse, over a meaningful time frame. That's when it seems clear to them that something is not only imbalanced but outside of their idea of "acceptable imbalance."
When it comes to P vs Z, I really think you have to look at how the games are being won and lost; what other matchup has such a history of all-ins? If Toss won 80% of the time they decided to all-in and 30% of the time they tried to play a macro game, yet the numbers were around 50%, overall, would this mean nothing needs to change? If this were the case, people would still be in here claiming the macro games are fine, even though clearly they aren't.
At the moment, I'm of the opinion that mutas are overpowered if you don't have a lot of stargates, already, as part of your build. That's not to say that Toss isn't, in fact, OP, if they do already have them. I'm of the opinion of if you want to raise the numbers in macro games, for Toss, then you either need a new unit or you need to buff something, like cannons.
Either that, or everyone gravitates towards playing the same way, which, while it might raise the win numbers, it would probably not be ideal for viewers.
After seeing CJ Heros interview, where he talked about how clueless and lost he is versus Zerg, odds are there aren't that many viable ways of playing.
From reddit, these are recorded ladder winrates over game time.
http://www.reddit.com/r/starcraft/comments/2ggsza/ladder_winrate_by_matchup_and_game_length_from/
![[image loading]](http://i.imgur.com/FzPBpCU.png)
Let's disect the game by periodes:
P1: 0-8mins
P2: 9-13mins
P3: 14-20mins
P4: 20-28mins
P5: 29mins+
P1: relative few wins come from this periode of time. Protoss can win at this point in time with canon rushing, proxy gate or 1base/1basish cheese. Zerg can do early pools and early speedling or roach attacks.
None of those are really a problem, but zerg has a strongly increased winrate. My guess would be that a lot of canonrush wins are not recoreded here, because the game drags beyond 8mins. Also I believe that in any case the defender should usually come on top more often than the attack in this periode of time.
P2: basically all the popular and not so popular Protoss allins (of 2bases) happen now. Zergs main way to win here is by pressuring a third base or holding an allin. The Protoss winrate is quite increased. A lot of games end in this timeframe.
P3: The periode in which Zerg usually has the map. The timeframe when intial mutalisk play picks up the tempo and Zerg can do a lot of 3-4basish commited aggression or kill a Protoss, like roach/hydra type of plays. Sentries and Stalkers lose their power, but higher tech tools are expensive and still take time to be out in numbers. The winrates are in zergs favor. Probably a lot of Zerg wins here also come from retaliation to Protoss P2 allinish attacks.
P4: Protoss has now acquired a massive deathball. The techs are revealed and unless a trade occurs, zerg cannot really surprise with techswitches anymore. Protoss infrastructure and set up usually make it very hard to do damage with mutalisks or straight attacks. Zerg can still go toe-to-toe in the open but will often try to go for SHs. Heavy Warpgate and blink play become relevant again for harassment. The game looks quite balanced at this point.
P5: The endgame wrap up. The map becomes more and more mined out. Players have to stretch thin, which opens them up to harassment. In particular warp prisms become very powerful.
Zergs usually mass static defenses and try to slowly kill their opponent with Swarm Hosts or emphasize on opponents movement mistakes, but straight up engagements usually favor Protoss. The winrates are in Protoss favor. The amount of games that reach this status are comparable to P1, few in numbers.
To me this looks quite ok. Everything seems to swing back eventually. The matchup is a little extreme in who it favors at what time. However, comparing these stats to gameplay, to me it looks like it always comes down to scouting and reacting properly, which is what we should embrace. Take P2 for example: Protoss seems to win a lot more than zerg and that is probably because of Protoss allins. But that is really up to scouting, right? It's not like Protoss could just always 2base allin and thereby all later periodes don't happen and the overall winrate would go way in Protoss favor. Or the other way around, Zerg cannot "just win" in P3. They have a good winchance, but that's really still up to gameplay.
|
On October 21 2014 17:11 playa wrote:
I would like to remind people that Blizzard has said they aim to step in and fix any matchup once it is at 45% or worse, over a meaningful time frame. That's when it seems clear to them that something is not only imbalanced but outside of their idea of "acceptable imbalance."
When it comes to P vs Z, I really think you have to look at how the games are being won and lost; what other matchup has such a history of all-ins? If Toss won 80% of the time they decided to all-in and 30% of the time they tried to play a macro game, yet the numbers were around 50%, overall, would this mean nothing needs to change? If this were the case, people would still be in here claiming the macro games are fine, even though clearly they aren't.
At the moment, I'm of the opinion that mutas are overpowered if you don't have a lot of stargates, already, as part of your build. That's not to say that Toss isn't, in fact, OP, if they do already have them. I'm of the opinion of if you want to raise the numbers in macro games, for Toss, then you either need a new unit or you need to buff something, like cannons.
Either that, or everyone gravitates towards playing the same way, which, while it might raise the win numbers, it would probably not be ideal for viewers.
After seeing CJ Heros interview, where he talked about how clueless and lost he is versus Zerg, odds are there aren't that many viable ways of playing.
if you really thing zerg is so favoured in a macro game why dont you like actually back it up with statistics?
and im curious where you got the 44% number you keep qouting since i see a-47% on aligulac
|
On October 21 2014 18:24 Enigmasc wrote:Show nested quote +On October 21 2014 17:11 playa wrote:
I would like to remind people that Blizzard has said they aim to step in and fix any matchup once it is at 45% or worse, over a meaningful time frame. That's when it seems clear to them that something is not only imbalanced but outside of their idea of "acceptable imbalance."
When it comes to P vs Z, I really think you have to look at how the games are being won and lost; what other matchup has such a history of all-ins? If Toss won 80% of the time they decided to all-in and 30% of the time they tried to play a macro game, yet the numbers were around 50%, overall, would this mean nothing needs to change? If this were the case, people would still be in here claiming the macro games are fine, even though clearly they aren't.
At the moment, I'm of the opinion that mutas are overpowered if you don't have a lot of stargates, already, as part of your build. That's not to say that Toss isn't, in fact, OP, if they do already have them. I'm of the opinion of if you want to raise the numbers in macro games, for Toss, then you either need a new unit or you need to buff something, like cannons.
Either that, or everyone gravitates towards playing the same way, which, while it might raise the win numbers, it would probably not be ideal for viewers.
After seeing CJ Heros interview, where he talked about how clueless and lost he is versus Zerg, odds are there aren't that many viable ways of playing. if you really thing zerg is so favoured in a macro game why dont you like actually back it up with statistics? and im curious where you got the 44% number you keep qouting since i see a-47% on aligulac
He is quoting actual aligulac-statistics.
On aligulac, if you go to "ranking/current", on the right side there are half-monthly winrates under the tag "List XXX". The current list (which is just a week), list 122 has ~46% winrate, list 121 had 43% and list 120 had ~45% winrate.
October will probably have around 44% winrate overall.
However, this isn't that long. It's like 1.5months and the current trend is trending towards normalizing again. If you go to the lists before the cited ones, it's between 48-51% all the time.
|
On October 21 2014 18:24 Enigmasc wrote:Show nested quote +On October 21 2014 17:11 playa wrote:
I would like to remind people that Blizzard has said they aim to step in and fix any matchup once it is at 45% or worse, over a meaningful time frame. That's when it seems clear to them that something is not only imbalanced but outside of their idea of "acceptable imbalance."
When it comes to P vs Z, I really think you have to look at how the games are being won and lost; what other matchup has such a history of all-ins? If Toss won 80% of the time they decided to all-in and 30% of the time they tried to play a macro game, yet the numbers were around 50%, overall, would this mean nothing needs to change? If this were the case, people would still be in here claiming the macro games are fine, even though clearly they aren't.
At the moment, I'm of the opinion that mutas are overpowered if you don't have a lot of stargates, already, as part of your build. That's not to say that Toss isn't, in fact, OP, if they do already have them. I'm of the opinion of if you want to raise the numbers in macro games, for Toss, then you either need a new unit or you need to buff something, like cannons.
Either that, or everyone gravitates towards playing the same way, which, while it might raise the win numbers, it would probably not be ideal for viewers.
After seeing CJ Heros interview, where he talked about how clueless and lost he is versus Zerg, odds are there aren't that many viable ways of playing. if you really thing zerg is so favoured in a macro game why dont you like actually back it up with statistics? and im curious where you got the 44% number you keep qouting since i see a-47% on aligulac
The 44% number is from the last report on aligulac. Going back 2 reports, 1406 P vs Z games, the number is at 45%. I tend to believe the more recent the numbers, the more relevant. As for all-ins, I saw some data a few months ago that broke down the precise win rates for stages in the game, and Toss wasn't above 40% in macro games until very late stages in the game, where if they've made it that far then they had quite the economy and infrastructure. If Toss has static D and multiple stargates, then they have the ability to counter everything. If that's not part of your build to begin with, though, then it's not easy to get to that point.
|
On October 21 2014 16:56 Thieving Magpie wrote:Show nested quote +On October 21 2014 16:50 The_Red_Viper wrote:On October 21 2014 16:46 Thieving Magpie wrote:On October 21 2014 16:42 The_Red_Viper wrote:On October 21 2014 16:37 Thieving Magpie wrote:On October 21 2014 16:25 The_Red_Viper wrote:On October 21 2014 16:23 Thieving Magpie wrote:On October 21 2014 16:21 The_Red_Viper wrote:On October 21 2014 16:18 Thieving Magpie wrote:On October 21 2014 16:14 playa wrote:
[quote]
When your race is at 44% in a matchup and the win rates via game length reflect Toss winning the majority of their all-in games, well, that 44% becomes even less when talking about macro games. Fact: there are not many ways at all to consistently win macro games. The math doesn't add up. There's either not many viable ways of playing or a lot of players are throwing games. This is common sense. This isn't some kinda conspiracy. 50% winrate (+/-) 6% is damn even. No it really is not What do you consider acceptable margin of error? Hard to say, but 6% points is pretty huge imo It really shouldn't be a hard thing to know. What is comfortable or you as a margin of error. You either know what you think it should be or you don't know what you think it should be. You, for example, don't know what you think it should be. To me, 10% is too high, and 5%-9% is acceptable, 0-4 is too close to call. Unless we have definitive terms on what we want achieved we will never be able to discuss anything. We need to know what numbers is acceptable and then begin discussion with the acceptance of those numbers. Otherwise we end up stuck arguing semantics when we could be arguing specifics. 56:44 is already over 10% off. So i guess it is too high for you as well? It shouldn't be higher than 53:47 imo, and even that is already somewhat questionable. Don't invent fake math please. 56:44 is +\- 6 53:47 is +\- 3 60/40 is +\- 10 According to you, a 3% variance is too high? Because that sound stupid. Yes +/- 6 percentage points. Which is 12% off. You really have no idea how margins of error works do you? Here's how it works in competitive anythings. 50% is ideal. +\- meta shifts, +\- skill gap, +\- correlative trends There will always be margins of error. What is the amount you consider acceptable. Saying +/- 6 is 12% off is a meaningless phrase since it is as often +6% as it is -6% which averages out to 50% over infinite iterations. So what is the margin of error you consider fair and please stop making up Maths that don't add anything to the discussion.
The concept of 'margin of error' doesn't even apply here. We are not looking at samples, but at complete numbers. We are looking at win rates and nothing else. If one race wins 100% of its matches against another race, we don't have a huge 'margin of error', we have complete imbalance. The nonsense you write and the tone you present it in made me report a tl post for the first time ever, yours.
|
One thing I always find weird in these discussions....
Since when is it accepted law that a 50% win ratio over all ladder equals perfect balance? I don't want to sound like a smartass, it's an honest question. For me 50%/50% just means that there's an equal amount of players winning/losing on both sides, which is not the kind of balance people want.
There's so many factors going into this that I believe it is a fallacy to say something like that. Maps, skill differences, conditions of players and so on.
Just look at brood war, there was no such thing as 50%/50% yet many consider brood war to be much more balanced than SC2 ever will be.
EDIT:
Maybe to further illustrate my point:
A Protoss player and a Zerg player face off against each other. Protoss does an unbeatable strategy on map X, then Zerg does an unbetable strategy on map Y. They always iterate like that and no one ever manages to beat the opponent on his respective map. They have 50%/50% - but is that "balance"?
|
On October 21 2014 22:24 Aiobhill wrote:Show nested quote +On October 21 2014 16:56 Thieving Magpie wrote:On October 21 2014 16:50 The_Red_Viper wrote:On October 21 2014 16:46 Thieving Magpie wrote:On October 21 2014 16:42 The_Red_Viper wrote:On October 21 2014 16:37 Thieving Magpie wrote:On October 21 2014 16:25 The_Red_Viper wrote:On October 21 2014 16:23 Thieving Magpie wrote:On October 21 2014 16:21 The_Red_Viper wrote:On October 21 2014 16:18 Thieving Magpie wrote:
[quote]
50% winrate (+/-) 6% is damn even. No it really is not What do you consider acceptable margin of error? Hard to say, but 6% points is pretty huge imo It really shouldn't be a hard thing to know. What is comfortable or you as a margin of error. You either know what you think it should be or you don't know what you think it should be. You, for example, don't know what you think it should be. To me, 10% is too high, and 5%-9% is acceptable, 0-4 is too close to call. Unless we have definitive terms on what we want achieved we will never be able to discuss anything. We need to know what numbers is acceptable and then begin discussion with the acceptance of those numbers. Otherwise we end up stuck arguing semantics when we could be arguing specifics. 56:44 is already over 10% off. So i guess it is too high for you as well? It shouldn't be higher than 53:47 imo, and even that is already somewhat questionable. Don't invent fake math please. 56:44 is +\- 6 53:47 is +\- 3 60/40 is +\- 10 According to you, a 3% variance is too high? Because that sound stupid. Yes +/- 6 percentage points. Which is 12% off. You really have no idea how margins of error works do you? Here's how it works in competitive anythings. 50% is ideal. +\- meta shifts, +\- skill gap, +\- correlative trends There will always be margins of error. What is the amount you consider acceptable. Saying +/- 6 is 12% off is a meaningless phrase since it is as often +6% as it is -6% which averages out to 50% over infinite iterations. So what is the margin of error you consider fair and please stop making up Maths that don't add anything to the discussion. The concept of 'margin of error' doesn't even apply here. We are not looking at samples, but at complete numbers. We are looking at win rates and nothing else. If one race wins 100% of its matches against another race, we don't have a huge 'margin of error', we have complete imbalance. The nonsense you write and the tone you present it in made me report a tl post for the first time ever, yours.
What was being talked about was a 1 month aligulac report which was different months ago and different months before that. Zerg and Protoss have had their winrates switching back and for forever. It going to 44% for 1 month does not mean it was 44% for the totality of the matchup. That is what margin of error is about. It is knowing that over infinite series the games average out to 50% and swings one way or another over short periods of time. What is being discussed when talking about margin of error is what range is it acceptable for short term winrates to be problematic. Hence why the description is 50% +\-X% wherein the matchup is 50% over a long period of time +/-X% of a short period of time.
Saying you disallow margin of error is saying you want things patched the moment one race wins 1-2 tournaments and you now assume that the matchup is unbeatable just because of how things played out in the past 30 days instead of the past year. It's a problematic symptom of gut call reactionary balance whining.
|
On October 21 2014 22:35 KeksX wrote:
One thing I always find weird in these discussions....
Since when is it accepted law that a 50% win ratio over all ladder equals perfect balance? I don't want to sound like a smartass, it's an honest question. For me 50%/50% just means that there's an equal amount of players winning/losing on both sides, which is not the kind of balance people want.
There's so many factors going into this that I believe it is a fallacy to say something like that. Maps, skill differences, conditions of players and so on.
Just look at brood war, there was no such thing as 50%/50% yet many consider brood war to be much more balanced than SC2 ever will be.
EDIT:
Maybe to further illustrate my point:
A Protoss player and a Zerg player face off against each other. Protoss does an unbeatable strategy on map X, then Zerg does an unbetable strategy on map Y. They always iterate like that and no one ever manages to beat the opponent on his respective map. They have 50%/50% - but is that "balance"?
This is precisely why I have been talking about margin of error. People are so reactionry on TL (and I'm assuming other forums) that the moment it's more than 0%-1% off of 50% winrates they cry bloody murder and believe it's an impossible matchup that requires patching.
|
It's difficult to state what exactly winrates mean, but from experience one can tell that if win rates exceed 55:45 this correlates with match-ups feeling broken. (for me personally) If the win rates go to 60:40 it's enough to create authentic mass panic on the forums.
|
Germany2808 Posts
Someone said Broodwar was much more balanced then Hots. But the rates where never at 50/50.
Quote form Liquipedia ( "ZvP win-loss ratio of all games played until 2008-11: 53.46%, PvT ratio until the same time: 51.09%" )
So maybe we should focus more in something more... usefull then just the aligulac ratings.
|
On October 21 2014 18:21 Big J wrote:Show nested quote +On October 21 2014 17:11 playa wrote:
I would like to remind people that Blizzard has said they aim to step in and fix any matchup once it is at 45% or worse, over a meaningful time frame. That's when it seems clear to them that something is not only imbalanced but outside of their idea of "acceptable imbalance."
When it comes to P vs Z, I really think you have to look at how the games are being won and lost; what other matchup has such a history of all-ins? If Toss won 80% of the time they decided to all-in and 30% of the time they tried to play a macro game, yet the numbers were around 50%, overall, would this mean nothing needs to change? If this were the case, people would still be in here claiming the macro games are fine, even though clearly they aren't.
At the moment, I'm of the opinion that mutas are overpowered if you don't have a lot of stargates, already, as part of your build. That's not to say that Toss isn't, in fact, OP, if they do already have them. I'm of the opinion of if you want to raise the numbers in macro games, for Toss, then you either need a new unit or you need to buff something, like cannons.
Either that, or everyone gravitates towards playing the same way, which, while it might raise the win numbers, it would probably not be ideal for viewers.
After seeing CJ Heros interview, where he talked about how clueless and lost he is versus Zerg, odds are there aren't that many viable ways of playing. + Show Spoiler +From reddit, these are recorded ladder winrates over game time. http://www.reddit.com/r/starcraft/comments/2ggsza/ladder_winrate_by_matchup_and_game_length_from/Let's disect the game by periodes: P1: 0-8mins P2: 9-13mins P3: 14-20mins P4: 20-28mins P5: 29mins+ P1: relative few wins come from this periode of time. Protoss can win at this point in time with canon rushing, proxy gate or 1base/1basish cheese. Zerg can do early pools and early speedling or roach attacks. None of those are really a problem, but zerg has a strongly increased winrate. My guess would be that a lot of canonrush wins are not recoreded here, because the game drags beyond 8mins. Also I believe that in any case the defender should usually come on top more often than the attack in this periode of time. P2: basically all the popular and not so popular Protoss allins (of 2bases) happen now. Zergs main way to win here is by pressuring a third base or holding an allin. The Protoss winrate is quite increased. A lot of games end in this timeframe. P3: The periode in which Zerg usually has the map. The timeframe when intial mutalisk play picks up the tempo and Zerg can do a lot of 3-4basish commited aggression or kill a Protoss, like roach/hydra type of plays. Sentries and Stalkers lose their power, but higher tech tools are expensive and still take time to be out in numbers. The winrates are in zergs favor. Probably a lot of Zerg wins here also come from retaliation to Protoss P2 allinish attacks. P4: Protoss has now acquired a massive deathball. The techs are revealed and unless a trade occurs, zerg cannot really surprise with techswitches anymore. Protoss infrastructure and set up usually make it very hard to do damage with mutalisks or straight attacks. Zerg can still go toe-to-toe in the open but will often try to go for SHs. Heavy Warpgate and blink play become relevant again for harassment. The game looks quite balanced at this point. P5: The endgame wrap up. The map becomes more and more mined out. Players have to stretch thin, which opens them up to harassment. In particular warp prisms become very powerful. Zergs usually mass static defenses and try to slowly kill their opponent with Swarm Hosts or emphasize on opponents movement mistakes, but straight up engagements usually favor Protoss. The winrates are in Protoss favor. The amount of games that reach this status are comparable to P1, few in numbers. To me this looks quite ok. Everything seems to swing back eventually. The matchup is a little extreme in who it favors at what time. However, comparing these stats to gameplay, to me it looks like it always comes down to scouting and reacting properly, which is what we should embrace. Take P2 for example: Protoss seems to win a lot more than zerg and that is probably because of Protoss allins. But that is really up to scouting, right? It's not like Protoss could just always 2base allin and thereby all later periodes don't happen and the overall winrate would go way in Protoss favor. Or the other way around, Zerg cannot "just win" in P3. They have a good winchance, but that's really still up to gameplay.
Is that from all leagues? I remember a lot of cheesy zergs and skytoss deathballs in lower leagues, maybe its still a pattern?
|
Hey guys, what about that merged templar tech? More options and tempo for Protoss in midgame.
|
Northern Ireland25642 Posts
I've never really liked DTs as a tempo unit until the very lategame. Definitely something at least worth consideration if a LoTV world entails a good bit of changes.
It would be pretty massive a change to make at this moment in time, especially with PvT not being in a horrendous state at present.
|
|
|
|
|
|
EPT
