[NvPinder]

Originally Posted by Tom

I don't know whether others have thoughts on this, but it seemed to me Protoss performed very well in Dreamhack.

[DREAMHACK SPOILERS]
Where HuK got into trouble is where he did expansion/heavy macro builds on maps like Shakuras. Macro Protoss just doesn't seem terribly viable at the moment, but perhaps this is more than compensated for by strong early game timing attacks.

I'd disagree with that analysis of shakurus, with better positioning at the start of the fight and better FF huk definitely could've held with minimal sentry loss and been on 6 gas with 2 robo, 5 gates, blink and still a fantastic sentry count against an even worker even base count zerg on lair tech. He lost that game via an army miss-step, not via his build being not superior. Also the fact that moon instantly saw thru his build and saw the early third as opposed to (what huk was wanting) moon assuming a 5-7 gate pressure was coming.

In terms of the build itself, more or less everything that could go wrong went wrong for him, and yet it was still a salvageable game for huk.

[divinesage]

So what happens now? It seems protoss is lacking new options to squeeze out advantages in this new metagame.

[Dispatcher]

I feel that the High cost of Units vs the Difficulty of control of protoss is what is holding a lot of players back in higher league play. I may be way off when I say this but I have recently became very interested in a lot of the GSL play, where Terran seem to be dominating. It is important to know that I am a sub-average starcraft player and strictly play protoss. I am not complaining about anything that I post and I apologize if I repeat anything someone has said, as I did not read the entire forum (but most of it). I will break down the thoughts that lead me to the above conclusion from the events of protoss games I have watched in the past 2 months of GSL (code A and code S) and MLG tournaments.

I have noticed how immobile protoss units are vs zerg and terran. I may be completly wrong here but let me explain this in detail by pointing out the strong points of terren and zerg towards mid to late game, non heavy cheese matches (as these are the types of matches that "most" high end games tend to be played).
First off Terran. Almost every Terran win I have seen has come from strong map control and multipronged attacks. Drops and Banshee harass in multiple locations make it very difficult for protoss and even zerg at times to defend not only the drop but also entries to their natural. One big point here is to note that it is not difficult to scout a drop ship and react at this level of play but the problem comes from the frequency of drops and the ease of evacuating if the drop was dealt with properly. If protoss does not send enough units to the drop they tend to make uneven trades, and if they do send the proper amount of re-enforcements terran can easily escape rarely unscathed. Keep in mind that even scouting out a drop ship does not guarantee that you will select the right amount of units to send to the drop as you may not always know the exact # of drops or even the complete composition of them. Drops "usually" put players in a defensive mode where they usually lose map control and or opportunities to intercept expansions.
Now to Zerg. As zerg are not very mobile early game they develop a huge advantage to mobility in the mid game when battles are fought in their territory i.e. creep. A good zerg player will have creep extend past the halfway mark of the map towards the end of the game, time permitting obviously. Also speed upgrades and good muta micro influence this mobility in similar ways to terran drops as previously stated. This usually results in protoss trying to contain zerg to 2 bases which becomes increasingly difficult towards late game play.
I have seen 1 game in the last 2 months where a protoss player has shown great mobility and this occurred just recently in the GSL code A tournament between oGsHero and Noblesse (set 2) where oGsHero showed off his ever so impressive control of the phoenix. I do know that it takes a lot of very precise phoenix play and what appeared to be some extreme micro of units, which I believe Hero is the only person at this level with these types of units. Also forward pylons increase mobility in a way but this is only in the sense of army replenishment not actual unit movement. Also anther form of protoss mobility is not mobility at all but more immobilizing other units with force-fields, which tends to disappear late games as gas becomes more important to spend on higher tech and most observers have been diluted through small battles through mid game.

In conclusion it appears that the ease of micro and the utility of the zerg and terran units increase their mobility against the slower protoss units. Also noted that it should be very understood that protoss should be a slower unit as 1 protoss unit is generally more powerful than any unit of the other races. I guess the point here is more on the ease of harassment for the other races.

I am breaking my other observations into multiple posts as this seems to be very long.

[Dispatcher]

I may be extremely biased on this one as I only play protoss and well compared to the level of play that is being discussed I play like a 2 year was slapping the keyboard because it was fun. I will make this post brief and up for the utmost criticism as I may be way off in my judgments on this one.

Terran. As most terran victories against protoss com towards mid game heavy pressure and counters I do not see to much micro from winning battles. To me it seems to be stim, stay in range of seige if you fallowed that tech pattern, and single target down huge threats aka colossi, sentries and occasionally immortals. Yes kiting is involved in some scenarios but this is just second nature at this level. For the most part terran micro what would appear to require a mid to low level amount of APM (strictly talking army vs army, game breaker scenarios). I also think the most difficult macro scenarios from terran come from harassment i.e. banshee and helion.

Zerg. I feel that zerg micro may be slightly more difficult than terran's as they do have to manage several groups of units in game breaking battles. But all and all the A.I. design for zerg seems to do a lot of the work for surrounds and con-caves. Might be a little more art to a zergling surround than what A.I. can provide. They also have movements similar to what high templar's and ghosts have with their infestors.

Protoss. One thing I notice with any large battle with protoss is as the ball moves around the units become jumbled together and where most protoss players get in trouble is their slower meat shield units, zealots, tend to get stuck behind their concave. To actually separate these units properly while moving around the map seems to be difficult for players when they approach the 200 supply limit. Without the zealots in front it becomes very easy for colossi to be picked off by terran and for surrounding units of zerg to encapsulate the protoss ball. Not to mention the extreme micro that it would take to keep up with any blink stalkers with low shields that is almost required to win battles in mid to late game.

This is where most of the cost vs control portion of protoss comes in. Higher tech units become very hard to replenish late game similar to terran tech. I do feel that the top of the tech tree colossus are easily killed by the the ranged and air units of the other races if not macro'ed correctly and by single targeting these units down the protoss can fall behind very quickly in army cost trades. I guess what I am trying to stress when I discuss these issues is that generally the high cost of even tier 1.5 protoss units become very precious in the game and if there is the slightest mistake in any micro of protoss it becomes very difficult to recover from. This has a lot to do with the mechanics of the more powerful but lower number of units that match to the same supply #'s of terran and zerg. All in all micro becomes more important to protoss when trying to keep units alive as long as possible, due to the fact that their cost and replenishment are not as cheap and accessible as the others.

I know a lot of this info my be false so feel free to hammer me as much as you want on this one.

[Dispatcher]

The bread and butter units for protoss are its sentries. Unfortunately they become less of a priority over higher tier units towards late game play. This is due to the gas that it would take away from the production of these units. Keeping early game sentries alive is almost the key to any later game protoss victory. The bad part is it's inevitable that there will be several casualties to these units in early game play that is required at higher levels. I think this is why most protoss players are pushed to cheese or even commit themselves to early game all ins. By the time they reach the late game level usually around the 2 or 3 expo they have fallen victim to higher tech and the issues discussed in the other threads. Higher tech meaning that they no longer produce sentries in order to produce higher tier units to keep up with higher tier units of other races.

Quick Comments

SnoWPanda:

that is one long post. thanks for contributing your thoughts

[Dispatcher]

Combining the last 3 post I will now add my thoughts to the state of protoss and possible but very subtle solutions the protoss balance.

I have noticed that most protoss victories come off of 3 types of play. One being the ever hated cheese. Two being the more popular early game all in. Three being the clever mid game 2 base well played timing attacks. Protoss late game suffers due to the lack of sentries, low mobility, costly (meaning resources) mistakes of imprecise micro and lack of scouting abilities.

Of these issues I feel that scouting any sort of air movement on the outer edges of the maps punishes protoss mineral lines and expansions. I also feel that the difficult micro of protoss punishes them late game.

Overall I feel that the game is very balanced but I think some slight changes towards protoss in the late game would balance things out for them. Some ideas that may possible rectify this would are listed below.

Observers are great scouting tools but they also take away from production of immortals and colossi. If they were included in a build option for gateways, with a change in cost and a requirement of a robotics facility I think this could help with evening the scouting playing field. Lets face it the scans and versatility of zerg and terran units and structure are far superior to that of protoss. Protoss just doesn't have a way of monitoring air movement like the others can.

Another idea I have had is to somehow allow movement through sentry forces fields for protoss units. The only fair solution I can think of is to have some sort of cancel all current force field option to sentries.

I also think a slight change to zealot A.I. would ease the micro required by protoss. I might be way out in left field on my micro topic and not saying that they need to totally change how protoss works but zealots constantly getting stuck in the back of the protoss ball just doesn't cut it some times.

My final thought and the worst thought of them all would be to have some sort of tech option perhaps in the robo bay to possibly exchange gas for mineral cost or even reduce the gas cost of sentries.

With that said feel free to correct anything that have improperly stated.

[Monk]

Woah shit. You Protoss out there might need to take a page out of NsHoSeo.Sage's book. This random guy

+ [GSTL Spoilers] +

All-killed fOu's line up of a Zerg, a Terran, a Terran and a Zerg

. It was very impressive play even though is army control wasn't the best.. close to the worst too.

[YJY]

you guys shuold also consider the fact that because protoss is immensely easy to play compared to the other races, the higher level skilled players dont actually have real skill compared to terran and zerg (as it requires a higher level to play these races). I know as a fact that a lot of terran / zerg players are much better than protoss players though the results dont say it, because they require much more macro/micro.

i think this is the most plausible reason why protoss are shit at high levels.

[Tom]

Originally Posted by YJY

you guys shuold also consider the fact that because protoss is immensely easy to play compared to the other races, the higher level skilled players dont actually have real skill compared to terran and zerg (as it requires a higher level to play these races). I know as a fact that a lot of terran / zerg players are much better than protoss players though the results dont say it, because they require much more macro/micro.

i think this is the most plausible reason why protoss are shit at high levels.

Maybe this is trolling.

Lets test the thesis

Lets test this thesis. Assume arguendo:

1. protoss is the easiest race to play.

2. there is an equal distribution of "skill" accross players of all races.

If both premises 1 and 2 are correct, you would expect that protoss would be substantially out-performing all other races across the board. That is, if two players of equivalent skill play each other, but one is playing an "easier" race, you would expect the player using the "easy" race to win in the majority of cases.

On the other hand, if two players of equal skill play each other using different races, and have a win/loss ratio of 50/50 (over a significant number of games), in what sense can one race be seen to be "easier" than the other?


Are our premises correct?

The data I have seen suggests protoss are certainly not overperforming in high level ladder and tournament play.

Now, this tells us (at least) that premise 1 or premise 2 are incorrect. Which should we reject?

Premise 1

If premise 2 is correct, it follows that we should reject premise 1. This is because, were both 1 and 2 correct, we would expect to see significant outperformance by protoss players in high level play. This is not the case. Lets look at premise 2 then.

Premise 2

But is premise 2 correct? There is no data on the relative "skill" of the players who pick the different races, and I suggest that in the absence of data, the most reasonable assumption to make is that the skill distribution across all races is generally equivalent.

The absence of data on the "skill" of players picking the various races is of course why it is necessary to look at results to gauge the performance of a race. By balancing the game in this way, Blizzard effectively make the "most reasonable assumption in the absence of data" that I have set out above, and they are in my view right in doing so.

It seems to me highly improbablethat all the "bad" players decide to play Protoss, whereas the "good" players decide to pick say Terran or Zerg. The other problem with this is, of course, that even if the masses of Protoss players are generally terrible, it only takes a small proportion of "super gosus" to pick Protoss and you would expect to see equivalent results in top level play.

Summing up (or TL;DR!)

There is what I would call a "logical disconnect" between the premise "Protoss is the easiest to play" and the conclusion "therefore, high level Protoss players have no real skill".

I would have thought that if Protoss was truly the easiest race, professional players (who play the game with a view to winning big prize money) would all be switching to Protoss. This is because in a contest between a super-skilled player using a "hard" race, and a super-skilled player using an "easy" race, the super-skilled player using the easy race will generally triumph. And we would therefore be seeing the results of this in high level play, with Protoss taking the majority of big-money tournaments.

We are not seeing this, and this suggests to me that YJY's thesis is quite wrong. Just to stir things up a little bit, I think it could more plausibly be suggested that terran is the easiest race to play, given the saturation of terran players in high level, big money Korean tournaments (where players wish to obtain every (legal) advantage, no matter how small, in order to win).

Quick Comments

	psygoh:	Good systematic way of analysing
	TidaL:

[YJY]

you guys shuold also consider the fact that because protoss is immensely easy to play compared to the other races, the higher level skilled players dont actually have real skill compared to terran and zerg (as it requires a higher level to play these races). I know as a fact that a lot of terran / zerg players are much better than protoss players though the results dont say it, because they require much more macro/micro.

i think this is the most plausible reason why protoss are shit at high levels.

[Digby]

Originally Posted by YJY

you guys shuold also consider the fact that because protoss is immensely easy to play compared to the other races, the higher level skilled players dont actually have real skill compared to terran and zerg (as it requires a higher level to play these races). I know as a fact that a lot of terran / zerg players are much better than protoss players though the results dont say it, because they require much more macro/micro.

i think this is the most plausible reason why protoss are shit at high levels.

So then, you're saying if a Protoss player and a Terran player for example, switched race, neither having previous 1v1 experience with either, the Terran player playing as Protoss would win?

I beg to differ my misguided friend, (this is just something that I did, whether or not it provides any conclusive evidence is up for scrutiny) I played a 1v1 with my friend who is a very capable masters Terran player after a lovely balance dicussion, myself being a diamond Protoss player challened him to play Protoss and I would play Terran. The outcome, it was a long macro game on Xel Naga Caverns and in the end, I ended up winning with MMM Vikings and Ghosts. Going up against his "Unstoppable" deathball which i may add, included Carriers and a Mothership along with Colossi.

Vikings > Colossi AND Carriers
Ghosts and EMP > Protoss

[xceL]

I have to agree with the first page about Protoss forcefields needing a slight nerf, nothing drastic.
Perhaps something like 75 energy to cast FF so sentries have a max of 2.5 forcefields on full mana instead of 4 or something like that.

I feel that that at levels below GM that the 200/200 deathball of protoss is near unstoppable. While the top tier players can deal with it, I think the simplicity of executing an A-Move with storms or forcefields is what makes it so challenging for zergs and terrans to deal with it.

I would have agreed that protoss was underpowered early game before the time when they figured out 3 gate pressure with expand was really solid, but now I think it's even across the races early on.

Quick Comments

deL:

Gotta be GM level before you can comment on GM balance.

[Tom]

http://us.battle.net/sc2/en/forum/topic/2878771457:

Overall balance has also proven to be an issue on Metalopolis -- even factoring in close position spawn issues. It’s among the least balanced maps currently in the ladder pool, and along with Scrap Station (also being removed) and Tal’darim Altar, has a heavy (60%+) bias toward zerg at the highest levels of play.

I find this post very interesting for a number of reasons.

First, I didn't realise Metalopolis was so biased towards zerg. I've personally found it to be a pretty good map. Obviously, as a Protoss player I have vetoed Scrap Station since early on in season 1, but its surprising Blizzard took this long to do something about it. I'm also quite surprised that Tal'Darim Altar is also a Zerg-favoured map at "the highest levels of play". I wonder why Blizzard didn't remove it as well?

Secondly, it suggests Blizzard are focused on trying to balance the game through map choice rather than making further changes to the races. This was the "third variable" identified in my post, which I said I didn't have any statistics on (which I thought was a significant limit to the conclusions). I suppose we need to wait and see what happens.

Finally, and most interestingly for me, this post confirms Blizzard are looking at data for different "levels" of play in balancing the game.

[Next_rim]

Why would you go DT's against terran?

Makes it hard for terran to push, buys time to catch up in macro, gives map control, opens HT tech path, drains OC energy. Overall, very cost efficient, if used correctly.

I barely see hallucination ever used in any kind of high level play

It's used a lot for scouting and proxy warp, when you put a pylon below cliff, and use hallucination for vision. And in battle, you need every precious bit of sentry energy for emergency forcefields, which obviously have much more impact on the battlefield than hallucinations. Besides, i play desert strike a lot, and a lot of people like doing sentries with illusions to tank damage. In short, it doesn't work that well.

Ghosts and EMP > Protoss

Agreed, I lost countless number of games to money EMPs. I think SC2 is macro-oriented, and should not be so unforgiving for mismicro in battle.

Brotoss needs better micro in addition no 100% usage of game mechanics. That is mommaship, carriers and warp prisms.

[Tom]

^ Hmm, yeah I never realised I need to be DEFENSIVE early game. You are right, I definately need to stop crying UNBALANCED and change my stratagys to include a few sentries! How to build?

In the future I will take some mroe time to delve into facts a little more before I post. I am grateful for your well-researched post, which I will use as a guide for my future posts.

(Apologies to others for lowering the tone of the debate, but this one was just too hilarious to resist.)

[TAScarecrow]

This needs to be closed, it was a great thread to begin with but the last few pages have gone to shit with trolls and clueless Protoss whiners.

[Tom]

^ agreed - please close!

[nGenLight]

http://www.teamliquid.net/forum/view...opic_id=252090

New data. According to the data - The illusion that Protoss was surging back up against Zerg that a few of us were so hopeful of seems to have turned the other direction. & Terran players in general has been surging lately (Past month).

Discuss.

[Tom]

Updated data

Many thanks to Ray for drawing my attention to the latest data.

Here are the direct links to the charts

- International data: http://i.imgur.com/CPVub.png
- Korean data: http://i.imgur.com/bdP2e.png

International tournament data as at July

A few months back when I first noticed the distinct trend, people were quick to point out that one data set, International tournament results, appeared to show the races were generally balanced. This is no longer case.

The International data shows:

- Protoss have consistently performed at less than 50% in International tournaments since March 2011
- Protoss is currently the worst performing races across all International tournaments (as at July)
- PvT slightly favours Terran in International tournaments
- PvZ heavily favours Zerg in International tournaments

Korean tournament data

The trend that emerged in April 2011 has contined, indeed what we now see is a volatile (but consistent) downward trend in Protoss results in Korean since December 2010.

Obviously, as the author of these data sets point out, Korea is a smaller data set and should always be treated with caution. Nonetheless, it is not that small, and we can be more confident in our conclusions if we start to see the same thing over and over for several months.

The Korean tournament data shows:

- Protoss is by far the worst performer in Korean tournaments, batting around 41%, compared to Zerg at approximately 52% and Terran at approximately 56%.
- PvT heavily favours Terran (indeed, in July, Protoss won only one third of PvTs)
- PvZ heavily favours Zerg (Protoss won approximately 43% of PvZs)

[GSL August Spoilers follow]

The above data goes to July. The trend has, if anything, accelerated in August.

Nearly every Protoss player was annihilated in the opening round of the GSL in both Code S and Code A (like the games weren't even close, it just wasn't funny). I think one P player in code A got through (Smart) in a mirror match up, and one player got through in a non-mirror (Tassadar).

When we look at the race breakdown a little more closely (courtesy of TLPD), we see:
- Code A Protoss had a win rate of about 22% in PvZ (admittedly a very small sampe size)
- Code A Protoss had a win rate of about 38% in PvT (again, a small sample)
- Code S Protoss, round of 16 is not even worth mentioning as there were only two Protoss to start with and both failed to win a game.

Blizzard ladder data and recent map balances

Blizzard have been very hedgy since the start of this year about actually releasing any data. However, I have heard them say (eg the infamous David Kim interview) that the match ups are "generally balanced" but (and I paraphrase), there are a couple of match ups they are keeping an eye on.

However, in introducing Season 3, Blizzard noted they were removing Scrap Station and Metalopolis on the basis that these were heavily Zerg favoured maps (60%+). They also stated that Tal'Darim Altar is a heavily favoured Zerg map (60%+), but did not remove the maps.

What I want to know is, if the match ups are "generally balanced", how is it that we have such a significant number of maps with a very heavy Zerg bias. One explanation is that all the other maps favour Protoss and Terran, although we can infer the bias must not be very high, as Blizzard did not cite balance as the reason for removing those other maps.

What do the Koreans think?

Lets have another look at that Korean QQ chart: http://www.teamliquid.net/forum/view...opic_id=144269

Obviously, this chart measures what people "think" about balance (sentiment) rather than balance, but its an interesting measure of the "spirit of the times". As at today, it shows around 70% think Protoss is underpowered, 8% think Terran is underpowered, and 20% think Zerg is underpowered.

Summing up

More than two months on, the evidence that Protoss is underpowered in high level play has become compelling. There were two reasonable arguments against this being the case, the first being a "shifting metagame" argument (ie we just need to allow Protoss time to adapt, and results will improve) and the International tournament data (which appeared at an earlier time to be around 50%). In my view, neither of these arguments is persuasive any longer. Five months have gone by, and there is no evidence of any shift in the meta game back towards Protoss. Most recent GSL data shows another season of terrible Protoss performance. International data has gone the way of the Korean data, with Protoss now significantly underperforming across all International tournaments.

Blizzard appear not to recognise there is a problem, apparently viewing the match ups as generally balanced. However, Blizzard have not released the data to support this claim, and it appears to be at odds with what any person looking at the Season 2 GM League standings can observe - a clustering of Zerg and Terran at the top of the ladders on each server. (In relation to the current season, I would be careful not to read anything into the standings at this point as we are only a couple of days in.) At the same time as making this claim, Blizzard openly stated three maps were heavily Zerg favoured (no such claim is made in relation to any maps favouring the other races), which at seem at odds with any claim that the match ups are "generally balanced".

I for one would like to see Blizzard's current ladder data, broken down not just by race but also by cohort (eg MMR bands, from High GM, low GM, High Master etc all the way down). Quite frankly, I do not find Blizzard's claims that the match ups are "generally balanced" to be credible. To describe something as "generally balanced" also hides a multitude of sins. To me, it begs the questions - are the match ups are generally balanced because Protoss is overperforming at low level play but underperforming at high level play? Are the match ups generally balanced in the sense that the match ups other than PvZ and/or PvT are balanced? I don't know, and I want to see the data!

Quick Comments

nGenLight:

Really amazing how quickly you write these near-flawless reports.

[SnoWPanda]

I just want to watch more GSL protoss matches honestly. I felt so cheated when I bought the last GSL ticket and ended up not having anything to watch.