|
On December 01 2012 04:42 Hider wrote:
If you want some numbers.
The average terran master player is at top 4.32% of his race. For zerg that is 6.26%.
The average terran GM is at top 0.17% of his race. For zerg that is 0.31%.
So is the average terran player worse than the average zerg? Possibly. Let's just assume that there are roughly 25000 (which is close to 25% of all terran players in bronze) terran players in bronze who absolutely suck and almost never play and jsut remove them from the equation.
The results for terran is now; 4.96% vs 6.26%. and 0.2% vs 0.31%
Still clear signs of superior terran players facing inferior zerg players.
Blizzard definitely needs to use the ladder distribution and combine it with other metrics (e.g. they could also use metrics such as solo games played, teamgames played, as active players probably are better than non-active players to help determine the "estimated skill level").
Regardless of what they do, it is almost avoidable that terran through an adjusted balance metrics will heavily underpowered.
I'm not necessarily saying that I disagree with you, but be careful about how much faith you put in that type of a metric.
There are several factors that influence MMR distribution that aren't related to balance. You already mentioned that Terran has a dis-proportionally large number of Bronze players (most likely because of the campaign), but there are any number of outside factors that can skew the distribution one way or another.
Maybe one race is easier to learn than the other, so more players end up in higher leagues of one race than another (not takes less skill, just more straight-forward to learn). Maybe because one race is harder to learn at low levels, so more players gave up playing while in low leagues or switched to another race robbing the player pool of skilled and dedicated players.
To be clear, I'm not implying that any of this IS the case, just that it CAN be the case with this type of metric.
|
On November 30 2012 11:29 SuperYo1000 wrote:
Global statistics don’t give us the impression that Zerg players have a general advantage, especially at the highest levels of play
In what universe are they getting there info from? At first I thought I must be taking crazy pills....Turns out...blizzard is sucking these down like candy
well zerg players losing on purpose seems to have worked.
It will be funny, when balance patches with crappy changes will be done, how instantly every zerg will start to play serious again.
|
The fact that ladder is used to balance the game at all is an issue. Do you know how many bad people there are on ladder? Even in GM, there are terrible players. A majority of the players on NA don't even have proper mechanics, they just win because they play other players who have just as bad mechanics. If you take an 3 average masters level players (1 of each race) from Korea, the American server, and the EU server, you will see clear differences in play styles, builds, and overall mechanics. Using ladder to balance the game is ridiculous .
There are players that have admittedly gotten to GM doing the same cheese over and over again (i.e the guy who got to GM with 6pool only, the maphackers, etc...).
|
On December 01 2012 05:01 ShamW0W wrote:
The only statistics that matter imo would be statistics at the very top level of play. I don't care how many players of race X or race Y are in Masters or GM, that's irrelevant to how balanced the game actually is. At the very top level though you'll get rid of factors like matchmaking and severely reduce the variable level of skill between player to get a more accurate representation of how well-balanced your game is.
I have a limited view but most of the tournament results I have seen recently suggest that P and Z are outperforming T. Not quite as much as some people think but from what I've seen it's there.
Where the Psionic change was probably too large of a change I feel this current one is too small. It doesn't address the core issue that players can still mass Infestors and hold their own because it's too much of a catch all unit.
Nerf Fungal, Nerf Infested Terran, buff the Hydralisk. Players shall rejoice.
That's one way of balancing the game. But the problem is how do we look at that?
GSL code s sample size is way too volatile (sample size to low) to be meaningfull alone. GM results from the ladder oculd be used (higher sample size) and according to those numbers, terran is heavily underpowered as well (through quite a lot of seasons).
|
On December 01 2012 05:08 TrippSC2 wrote:Show nested quote +On December 01 2012 04:42 Hider wrote:
If you want some numbers.
The average terran master player is at top 4.32% of his race. For zerg that is 6.26%.
The average terran GM is at top 0.17% of his race. For zerg that is 0.31%.
So is the average terran player worse than the average zerg? Possibly. Let's just assume that there are roughly 25000 (which is close to 25% of all terran players in bronze) terran players in bronze who absolutely suck and almost never play and jsut remove them from the equation.
The results for terran is now; 4.96% vs 6.26%. and 0.2% vs 0.31%
Still clear signs of superior terran players facing inferior zerg players.
Blizzard definitely needs to use the ladder distribution and combine it with other metrics (e.g. they could also use metrics such as solo games played, teamgames played, as active players probably are better than non-active players to help determine the "estimated skill level").
Regardless of what they do, it is almost avoidable that terran through an adjusted balance metrics will heavily underpowered.
I'm not necessarily saying that I disagree with you, but be careful about how much faith you put in that type of a metric. There are several factors that influence MMR distribution that aren't related to balance. You already mentioned that Terran has a dis-proportionally large number of Bronze players (most likely because of the campaign), but there are any number of outside factors that can skew the distribution one way or another. Maybe one race is easier to learn than the other, so more players end up in higher leagues of one race than another (not takes less skill, just more straight-forward to learn). Maybe because one race is harder to learn at low levels, so more players gave up playing while in low leagues or switched to another race robbing the player pool of skilled and dedicated players. To be clear, I'm not implying that any of this IS the case, just that it CAN be the case with this type of metric.
I kind of agree. I think Blizzard should look at several "adjusted metrics".
Then Blizzard could determine how the game is balanced at the highest level, the second highest level etc..
That the game is easier to learn for race x isn't a huge problem designwise, but still an optimal balanced game would have all races equally balanced across the various skill levels.
But Blizzard them selves could determine what kind of numbers they are justified with (maybe they are fine with the game being 40-60 at gold level if it is very close to 50-50 at master/gm).
But I think Blizzard should hire a full-time mathguy to assist Dustin Browder/David Kim. From what I gathered from their comments, I don't think they really understand how to interpret statistics or they don't know how to develop usefull metrics (which isn't that easy either but a math guy could come up with usefull metrics).
|
What an useless test.
Wonder how many Terrans or Protosses will actually bother when the intended fungal nerf changed to a goddamn 20hp egg nerf.
Make the fungal a slow moving projectile and add one more range to it? Now we have something to actually test.
Would not be too hard on the bronzies, but would add a whole another depth to prolvl play.
|
On December 01 2012 05:06 Hider wrote:Show nested quote +On December 01 2012 04:58 orBitual wrote:On December 01 2012 04:49 Hider wrote:On December 01 2012 04:45 orBitual wrote:On December 01 2012 04:26 Hider wrote:On December 01 2012 04:23 orBitual wrote:On December 01 2012 04:19 Hider wrote:On December 01 2012 04:18 orBitual wrote:On December 01 2012 04:14 Hider wrote:On December 01 2012 04:08 nottapro wrote:
[quote]
Lol exactly, concluding the game is balanced by reading the statistics of a system; "specifically enigineered" to make sure that everyone always maintain a near 50% win / loss ratio through hand selecting their opponents, is a new level of stupid. Please don't tell Dustin Browder that. It will ruin his understanding of the world. On another note.Do anyone remember when back in late beta when tanks dealt like 60 damage to everything (maybe that was slightly more imbalanced than infestors). DB said back then that the matchup (zvt) was 50-50 !!! and indicated that it was probably balanced. He probably does the same thing now, and combine his balance-comments by looking at the GSL. Obviously this is a flawed phiolosophy. They should have developed an expected win rate. One could look at the race distribution and assume that win rates has a normal distribution. Then one would could conclude that the top1% terran should be winning 65% (given the game was balanced) against the top3% zergs. But since the top1% terran only wins 50% of the game against the top3% zergs, the game is imbalanced by 15 percentage point. (though the numbers are invented, I wouldn't be surprised if they are very close to the truth). did you just invent numbers to try to criticize someone else's unscientific approach to balance? lol Great you missed the entire point. Go look at Sc2ranks. Terran is the most played race, yet the least present in any league but bronze. Clearly its not fair that a superior terran player goes 50-50 vs an inferior zerg player. The exact numbers isn't relevant to my point. Ok, I went over to sc2ranks.com and I didn't see a list of superior terran players and inferior zerg players so I could compare their records. Can you give a direct link? Or... is this just another one of your unscientific things? Wait what? YOu still don't get it? I already explained that one could look at ladder distributions and make an estimated win/rate based on an assumption of normal distribution. I wrote that in my first post. This is definitely an improved way of balancing, and one could look at other metrics as well (than just ladder distribution). As I said again, you can get the numbers from sc2ranks, and you can see that the top x% terrans plays against the top (x+b) % zergs where b is a positive number. Nah, you have yet to prove that a superior terran player goes 50-50 vs an inferior zerg player, which is what your argument is based on. The only thing I can think (now that you have backed away from that argument) is that you are assuming is that ladder distributio n should follow a normal distribution in proportion to population, which is not a given. Don't think you understand statistics. It doesn't matter how it is distributed actually (for my point to be valid - just look at the above posts). But one needs to know the distribution to quantiy the estimated win rate. The important assumption is that zerg players in average aren't heavily superior to terran players. In the above post I suggested ways that Blizzard could take average skill level into account, but no matter what, it can't completely justifiy the skewed distirbutions that we see ladder season after ladder season. I think any statistics-guy by looking at these numbers would conclude that terran is undervalued with a very high probability. I believe one has to be a zerg player to argue that these numbers are justified. Don't delude yourself, you haven't used any complicated statistics. You arbitrarily pick numbers to support your conclusion, and I can do the same using your own resources. From the world playerbase, active players in the last 30 days. If you exclude bronze, 4.5% of the Terran population is in GM and 4.8% of the Zerg population is in GM. 9.3% of the Terran population is in Master's and 8.3% of the Zerg population is in Master's. Seems close to me. *Note: this isn't statistics, just arithmetic. You are a master of missing the point (intentionally or just ignorant?). Sure this is basic statistics, but I just argued that you didn't understand the premise my assumption was build on due to lack of understnad game understnading. Also removing bronze completely is absolutely retarded. This is a hopelessly biased error because what is happening is that in an imbalanced world: low gm terran players gets into high master. Low masters --> Diamond low diamond into plat.. etc.. (you really never considered that there is a lot of terrans in bronze because the race might be UP?). So in the end there are lot of terran players who deserves to be in silver but is in bronze. Removing bronze is terrible solution. So I think I end our discussiong here as you are clearly a biased zerg player + you have little understanding of statistics/can't interpret numbers objectively.
Sorry, each of your points has relied on an arbitrary'invented choice of numbers:
'the top 1% of terrans has a 50% win rate against the top 3% of zergs'
'let's remove the bottom 25,000 number of terrans'
and then you criticize blizzard for being unscientific and me for not interpreting numbers objectively.
Is this objective: 'In an imbalanced world... <insert your conclusion of imbalance>' ? lol. Sorry if you're unable to construct durable arguments.
And hopelessly biased? Look at this person you responded to:
ZERG
GM: 0.283%
PROTOSS
GM: 0.203%
TERRAN
GM: 0.175%
so only 66 % of players are playing the game. Your response to this - criticism of their mathematical objectivity? no - "Lol exactly, concluding the game is balanced by reading the statistics of a system; "specifically enigineered" to make sure that everyone always maintain a near 50% win / loss ratio through hand selecting their opponents, is a new level of stupid."
I think you have eclipsed it.
|
|
|
Not that many viewers on the streams (25k on the main) or pages in the IPL thread imo. Really hope HotS shakes things for good because I feel a lot of people are losing interest (like me).
Really need to focus to make a fun to watch and play game first, then balance it.
|
On November 30 2012 21:06 Glon wrote:Show nested quote +On November 30 2012 11:56 TheDwf wrote:On November 30 2012 11:52 Glon wrote:
I was addressing the ZvP matchup. If you want my opinions of ZvT -- Watch bogus or Ryung for inspiration. Someone doesn't watch IPL5. Realize that the GSL Ro4 players are there for the GSL. Sniper also fell out 0-2 to STC in R2 -- they prepped for GSL and will save strategies for GSL, not use them in the early rounds of IPL. People seem to be forgetting that your entire army should not be able to be fungaled. 1. You should be split 2. Collosus in front -- focussing infestor 3. SOME templar near front to storm eggs/brood lord and feedback infestor/queen 4. SPLIT ARC w/ Stalker 5. Mothership NEAR front so that the zerg cannot poke in with Brood lords w/ out fully committing with rest of army (for fear of getting double vortexed directly on the brood lords) As for attacking into zergs, protoss should not be doing that until they have their ultimate/unbeatable army of carrier/mothership/2 collosus/arcon/high templar. Rely on warp prisms to force zerg into attacking you -- then engage the zerg in the open if possible. This egg change will also make it easier for late game engagements -- Since zergs must now actually have some roaches since infested terran eggs will die so easily before even hatching. Hope ^ helped. I can write a similar how-to engage for terran if people are interested. Maybe I'll even make a full guide, we'll see.
Some positive feedback to my post, some negative (as is to be expected when I write on a public forum, and is a VERY different reaction than when I discuss with practice partners). The main concensus is that people think that I was stating that protoss could not attack into zerg until they have a top-tier late game death ball. However, you are taking my guide out of context.
What I was writing was a guide on what to do versus the zerg that already has 10+ broods and 10+ infestors. You won't have your top-tier army yet -- in fact, you need another ~6-8 minutes to get it (or more depending on how the mid game went for you). Therefore, you must use what I wrote to its fullest effect in order to survive until your top-tier late game.
IN THE MEANTIME -- the BEFORE the 10+ brood 10+ infestor, nothing I wrote should be taken into account. For one, you won't have a mothership at that point, so take vortex completely out of the equation. Use collosus to focus fire infestors and keep sentries alive to forcefield off segments of the zerg army. If the zerg is relying on mainly infestors with only 5-6 brood lords, multi-pronged attacks work very nicely, or just simply army movement to catch the zerg player out of position.
Remember: with the egg change, 1 storm = all eggs killed, and collosus fire is much more effective at clearing infested terran eggs before they even spawn. So, the zerg that relys on 15 infestors before brood lords will be in quite a bit of trouble versus the player that pressures off of 3 bases while taking a fourth (with collosus and storm).
I am not stating that the change is perfect -- again, I still advocate for fungal growth NOT rooting High templar/Ghosts as to add more micro to the zerg late game engagement (selecting only the 3 closest brood lords to target a high templar, ect). However, it is an interesting step that I believe will help the protoss in the mid/late game where they don't quite have that top-tier army.
PS -- for those still questioning -- the transition period where the zerg player DOES have a 10 + brood lord and 10+ infestor army should be FILLED with warp prism harassment and strategic placement of the protoss army. You CAN take a fight in the middle where you have a surround, or where the infestors or brood lords are out of position. Additionally, aggresive poking with your army WILL cause zerg to pull units to a certain area, opening up pathways to the zerg main. What many players thought by "don't attack into a zerg" is compeltely wrong -- if you just sit on your ass on 4 bases, you will be attacked in a favorable position for the zerg player and will likely lose. The protoss (and the zerg player) must remain VERY active on the map with drops and main army movement.
|
I love how some people are trying to justify this change just because 1 storm can kill the eggs now. Where does that leave Terran? We don't have any kind of Collosus or HT aoe spells, what are we left to do, manually target down each egg with our mmm? Zergs are already using them to soak siege tank fire when they engage, so it's not like they're using IT for its DPS, but rather the egg to soak initial damage so nothing's changed. I just don't see anything changing for Terran with this infestor 'nerf'.
|
On December 01 2012 05:20 orBitual wrote:Show nested quote +On December 01 2012 05:06 Hider wrote:On December 01 2012 04:58 orBitual wrote:On December 01 2012 04:49 Hider wrote:On December 01 2012 04:45 orBitual wrote:On December 01 2012 04:26 Hider wrote:On December 01 2012 04:23 orBitual wrote:On December 01 2012 04:19 Hider wrote:On December 01 2012 04:18 orBitual wrote:On December 01 2012 04:14 Hider wrote:
[quote]
Please don't tell Dustin Browder that. It will ruin his understanding of the world.
On another note.Do anyone remember when back in late beta when tanks dealt like 60 damage to everything (maybe that was slightly more imbalanced than infestors). DB said back then that the matchup (zvt) was 50-50 !!! and indicated that it was probably balanced.
He probably does the same thing now, and combine his balance-comments by looking at the GSL.
Obviously this is a flawed phiolosophy. They should have developed an expected win rate. One could look at the race distribution and assume that win rates has a normal distribution. Then one would could conclude that the top1% terran should be winning 65% (given the game was balanced) against the top3% zergs.
But since the top1% terran only wins 50% of the game against the top3% zergs, the game is imbalanced by 15 percentage point. (though the numbers are invented, I wouldn't be surprised if they are very close to the truth). did you just invent numbers to try to criticize someone else's unscientific approach to balance? lol Great you missed the entire point. Go look at Sc2ranks. Terran is the most played race, yet the least present in any league but bronze. Clearly its not fair that a superior terran player goes 50-50 vs an inferior zerg player. The exact numbers isn't relevant to my point. Ok, I went over to sc2ranks.com and I didn't see a list of superior terran players and inferior zerg players so I could compare their records. Can you give a direct link? Or... is this just another one of your unscientific things? Wait what? YOu still don't get it? I already explained that one could look at ladder distributions and make an estimated win/rate based on an assumption of normal distribution. I wrote that in my first post. This is definitely an improved way of balancing, and one could look at other metrics as well (than just ladder distribution). As I said again, you can get the numbers from sc2ranks, and you can see that the top x% terrans plays against the top (x+b) % zergs where b is a positive number. Nah, you have yet to prove that a superior terran player goes 50-50 vs an inferior zerg player, which is what your argument is based on. The only thing I can think (now that you have backed away from that argument) is that you are assuming is that ladder distributio n should follow a normal distribution in proportion to population, which is not a given. Don't think you understand statistics. It doesn't matter how it is distributed actually (for my point to be valid - just look at the above posts). But one needs to know the distribution to quantiy the estimated win rate. The important assumption is that zerg players in average aren't heavily superior to terran players. In the above post I suggested ways that Blizzard could take average skill level into account, but no matter what, it can't completely justifiy the skewed distirbutions that we see ladder season after ladder season. I think any statistics-guy by looking at these numbers would conclude that terran is undervalued with a very high probability. I believe one has to be a zerg player to argue that these numbers are justified. Don't delude yourself, you haven't used any complicated statistics. You arbitrarily pick numbers to support your conclusion, and I can do the same using your own resources. From the world playerbase, active players in the last 30 days. If you exclude bronze, 4.5% of the Terran population is in GM and 4.8% of the Zerg population is in GM. 9.3% of the Terran population is in Master's and 8.3% of the Zerg population is in Master's. Seems close to me. *Note: this isn't statistics, just arithmetic. You are a master of missing the point (intentionally or just ignorant?). Sure this is basic statistics, but I just argued that you didn't understand the premise my assumption was build on due to lack of understnad game understnading. Also removing bronze completely is absolutely retarded. This is a hopelessly biased error because what is happening is that in an imbalanced world: low gm terran players gets into high master. Low masters --> Diamond low diamond into plat.. etc.. (you really never considered that there is a lot of terrans in bronze because the race might be UP?). So in the end there are lot of terran players who deserves to be in silver but is in bronze. Removing bronze is terrible solution. So I think I end our discussiong here as you are clearly a biased zerg player + you have little understanding of statistics/can't interpret numbers objectively. Sorry, each of your points has relied on an arbitrary'invented choice of numbers: 'the top 1% of terrans has a 50% win rate against the top 3% of zergs' 'let's remove the bottom 25,000 number of terrans'and then you criticize blizzard for being unscientific and me for not interpreting numbers objectively. Is this objective: 'In an imbalanced world... <insert your conclusion of imbalance>' ? lol. Sorry if you're unable to construct durable arguments. And hopelessly biased? Look at this person you responded to: ZERG GM: 0.283% PROTOSS GM: 0.203% TERRAN GM: 0.175% so only 66 % of players are playing the game. Your response to this - criticism of their mathematical objectivity? no - "Lol exactly, concluding the game is balanced by reading the statistics of a system; "specifically enigineered" to make sure that everyone always maintain a near 50% win / loss ratio through hand selecting their opponents, is a new level of stupid." I think you have eclipsed it.
Ok last time I respond.
1) I just came up with possible solutions or ways the could improve. But I actually suggested they should hire something, because the exact way of develpoing adjusted metrics is more complicated. WHat I tried to do is illustrate the flaws in their balance philosophy. Exact numbers are completely irrelevant (even though you are clearly biased I still don't get how you haven't figured that out yet).
2) Removing 25,000 bronze terrans is an example of taking into account that their could be "campaign-terrans", awfull terrans or w/e. This is quite a a high number. Blizzard probably have access to mroe data and could determine a more proper number. But what I showed that even after removing this number, terran was clearly still underrepesented. But I also came up with other ways Blizzard could determine the average skill level difference between zerg and terran. Again don't focus too much on the exact numbers. Rather focus on the methodology.
|
On December 01 2012 05:21 Salteador Neo wrote:
Not that many viewers on the streams (25k on the main) or pages in the IPL thread imo. Really hope HotS shakes things for good because I feel a lot of people are losing interest (like me).
Really need to focus to make a fun to watch and play game first, then balance it.
Sure for a good 2-3 months viewers will increase by 20%-30%. Then it's goanna go downhill again.
When that is said, I don't think balance or infestors being OP has that much to do with declining viewer numbers. It's more due to natural product maturity.
|
On December 01 2012 05:11 Hider wrote:Show nested quote +On December 01 2012 05:01 ShamW0W wrote:
The only statistics that matter imo would be statistics at the very top level of play. I don't care how many players of race X or race Y are in Masters or GM, that's irrelevant to how balanced the game actually is. At the very top level though you'll get rid of factors like matchmaking and severely reduce the variable level of skill between player to get a more accurate representation of how well-balanced your game is.
I have a limited view but most of the tournament results I have seen recently suggest that P and Z are outperforming T. Not quite as much as some people think but from what I've seen it's there.
Where the Psionic change was probably too large of a change I feel this current one is too small. It doesn't address the core issue that players can still mass Infestors and hold their own because it's too much of a catch all unit.
Nerf Fungal, Nerf Infested Terran, buff the Hydralisk. Players shall rejoice.
That's one way of balancing the game. But the problem is how do we look at that? GSL code s sample size is way too volatile (sample size to low) to be meaningfull alone. GM results from the ladder oculd be used (higher sample size) and according to those numbers, terran is heavily underpowered as well (through quite a lot of seasons).
Me personally? I'd prefer the small sample size, even though it's volatile, over using the larger sample size of GM players across the ladders.
For instance, in tournaments or leagues where the prize pool is $5k+, what are the racial win rates since the last patch? How have those rates evolved over the last few months as players have adapted to the current meta game? Are we trending closer to a more balanced game as players learn the metagame or is one race becoming dominant over the others?
You can use statistics to inform balance changes but ultimately there's a level of intuition about the game that a designer must have to make the best call for the game.
edit: If there's data for this already it'd be awesome to see. 
|
On December 01 2012 05:15 Hider wrote:Show nested quote +On December 01 2012 05:08 TrippSC2 wrote:On December 01 2012 04:42 Hider wrote:
If you want some numbers.
The average terran master player is at top 4.32% of his race. For zerg that is 6.26%.
The average terran GM is at top 0.17% of his race. For zerg that is 0.31%.
So is the average terran player worse than the average zerg? Possibly. Let's just assume that there are roughly 25000 (which is close to 25% of all terran players in bronze) terran players in bronze who absolutely suck and almost never play and jsut remove them from the equation.
The results for terran is now; 4.96% vs 6.26%. and 0.2% vs 0.31%
Still clear signs of superior terran players facing inferior zerg players.
Blizzard definitely needs to use the ladder distribution and combine it with other metrics (e.g. they could also use metrics such as solo games played, teamgames played, as active players probably are better than non-active players to help determine the "estimated skill level").
Regardless of what they do, it is almost avoidable that terran through an adjusted balance metrics will heavily underpowered.
I'm not necessarily saying that I disagree with you, but be careful about how much faith you put in that type of a metric. There are several factors that influence MMR distribution that aren't related to balance. You already mentioned that Terran has a dis-proportionally large number of Bronze players (most likely because of the campaign), but there are any number of outside factors that can skew the distribution one way or another. Maybe one race is easier to learn than the other, so more players end up in higher leagues of one race than another (not takes less skill, just more straight-forward to learn). Maybe because one race is harder to learn at low levels, so more players gave up playing while in low leagues or switched to another race robbing the player pool of skilled and dedicated players. To be clear, I'm not implying that any of this IS the case, just that it CAN be the case with this type of metric. I kind of agree. I think Blizzard should look at several "adjusted metrics". Then Blizzard could determine how the game is balanced at the highest level, the second highest level etc.. That the game is easier to learn for race x isn't a huge problem designwise, but still an optimal balanced game would have all races equally balanced across the various skill levels. But Blizzard them selves could determine what kind of numbers they are justified with (maybe they are fine with the game being 40-60 at gold level if it is very close to 50-50 at master/gm). But I think Blizzard should hire a full-time mathguy to assist Dustin Browder/David Kim. From what I gathered from their comments, I don't think they really understand how to interpret statistics or they don't know how to develop usefull metrics (which isn't that easy either but a math guy could come up with usefull metrics).
I'm sure they have access to the numbers you're talking about. You can only take educated guesses about their meaning, but that's all. A mathematician can't tell you what the numbers mean, only present the numbers.
|
On December 01 2012 05:01 ShamW0W wrote:
The only statistics that matter imo would be statistics at the very top level of play. I don't care how many players of race X or race Y are in Masters or GM, that's irrelevant to how balanced the game actually is. At the very top level though you'll get rid of factors like matchmaking and severely reduce the variable level of skill between player to get a more accurate representation of how well-balanced your game is.
I have a limited view but most of the tournament results I have seen recently suggest that P and Z are outperforming T. Not quite as much as some people think but from what I've seen it's there.
Where the Psionic change was probably too large of a change I feel this current one is too small. It doesn't address the core issue that players can still mass Infestors and hold their own because it's too much of a catch all unit.
Nerf Fungal, Nerf Infested Terran, buff the Hydralisk. Players shall rejoice.
I completely agree with this. To be honest i think all races would be happy with this.
|
On December 01 2012 05:11 Hider wrote:Show nested quote +On December 01 2012 05:01 ShamW0W wrote:
The only statistics that matter imo would be statistics at the very top level of play. I don't care how many players of race X or race Y are in Masters or GM, that's irrelevant to how balanced the game actually is. At the very top level though you'll get rid of factors like matchmaking and severely reduce the variable level of skill between player to get a more accurate representation of how well-balanced your game is.
I have a limited view but most of the tournament results I have seen recently suggest that P and Z are outperforming T. Not quite as much as some people think but from what I've seen it's there.
Where the Psionic change was probably too large of a change I feel this current one is too small. It doesn't address the core issue that players can still mass Infestors and hold their own because it's too much of a catch all unit.
Nerf Fungal, Nerf Infested Terran, buff the Hydralisk. Players shall rejoice.
That's one way of balancing the game. But the problem is how do we look at that? GSL code s sample size is way too volatile (sample size to low) to be meaningfull alone. GM results from the ladder oculd be used (higher sample size) and according to those numbers, terran is heavily underpowered as well (through quite a lot of seasons).
The other problem with using GM stats is that it is only 200 accounts per region, not players. Any professional player with multiple practice accounts will totally throw off the statistics if those accounts get into GM. That and the fact that getting into GM is more based on who is playing at the time GM opens up for that season makes it unreliable at best. Maybe taking for stats for an entire year might be useful, but there to many external factors to make the numbers for a specific season useful.
But arguing about stats is not what we are here to do, since we should be cutting our teeth on that test maps and proving that the changes are worthless or not.
|
well either way, from a spectator pov
This game is getting boring to watch
Especially games involving Zerg. (PvZ/ZvP, TvZ/ZvT)
|
On December 01 2012 05:31 Hider wrote:Show nested quote +On December 01 2012 05:21 Salteador Neo wrote:
Not that many viewers on the streams (25k on the main) or pages in the IPL thread imo. Really hope HotS shakes things for good because I feel a lot of people are losing interest (like me).
Really need to focus to make a fun to watch and play game first, then balance it. Sure for a good 2-3 months viewers will increase by 20%-30%. Then it's goanna go downhill again. When that is said, I don't think balance or infestors being OP has that much to do with declining viewer numbers. It's more due to natural product maturity.
Yep that's a big factor too.
Infestors being OP is part of the game being solved so fast IMO. I mean just look at the evolution of the game in the last year, after all early to mid game strategies got nerfed to the ground. Every protoss attack is an allin, even when on 3 bases. We see valuable ghost, thor, raven or BC play in 1/10 games at most. The most innovative thing by zerg I've seen was Leenock stutter stepping hydras against Infested Terrans xD
In BW we had surprising shit and new strats even after 10+ years in it.
|
On December 01 2012 05:33 ShamW0W wrote:Show nested quote +On December 01 2012 05:11 Hider wrote:On December 01 2012 05:01 ShamW0W wrote:
The only statistics that matter imo would be statistics at the very top level of play. I don't care how many players of race X or race Y are in Masters or GM, that's irrelevant to how balanced the game actually is. At the very top level though you'll get rid of factors like matchmaking and severely reduce the variable level of skill between player to get a more accurate representation of how well-balanced your game is.
I have a limited view but most of the tournament results I have seen recently suggest that P and Z are outperforming T. Not quite as much as some people think but from what I've seen it's there.
Where the Psionic change was probably too large of a change I feel this current one is too small. It doesn't address the core issue that players can still mass Infestors and hold their own because it's too much of a catch all unit.
Nerf Fungal, Nerf Infested Terran, buff the Hydralisk. Players shall rejoice.
That's one way of balancing the game. But the problem is how do we look at that? GSL code s sample size is way too volatile (sample size to low) to be meaningfull alone. GM results from the ladder oculd be used (higher sample size) and according to those numbers, terran is heavily underpowered as well (through quite a lot of seasons). Me personally? I'd prefer the small sample size, even though it's volatile, over using the larger sample size of GM players across the ladders. For instance, in tournaments or leagues where the prize pool is $5k+, what are the racial win rates since the last patch? How have those rates evolved over the last few months as players have adapted to the current meta game? Are we trending closer to a more balanced game as players learn the metagame or is one race becoming dominant over the others? You can use statistics to inform balance changes but ultimately there's a level of intuition about the game that a designer must have to make the best call for the game. edit: If there's data for this already it'd be awesome to see.
Well I think Blizzard should look at several metrics.
Then could of course design a high-level tournament win rate metric and weight it according to the signifcance of the results and for how important they believe it is.
But logically, why would there by any signifcant difference between top-top level and GM results on the ladder? (I know there is one argument which states that people play different on the ladder than tournament but that is not really relevant here as it probably is the similar for all three races).
Sure there are also players who are bad at tournamaents/good at ladder, but again, on average we expect this to even out.
In the end, I just think GM results on the ladder is a more important metric than high level tournament play which suffers from low sample sizes and could be biased due to many korean terrans being invited and few korean toss's/korean zerg's (over a large sample size this would probably be evened out as well, but this just leads to very volatile).
So I guess we can't just look at this tournmanet metric quantiatively. We have to combine it with a qualitive analysis which Blizzard also do, but this is a very difficult task for a small team, and honestly I don't think they are doing a particular good job (though they aren't awfull either - the game is still somewhat balanced, but due to design flaws of the game it makes it difficult to balance the infestor etc.).
Also, I dont agree that it's completely irrelevant that the game is balanced at master level/diamond level etc. Sure top level play should have highest priority, but balance matters for the playing experience and if players are unsatifised they could stop watching the game --> killing esports etc.
|
|
|
|
|
|
EPT
