Prayanavita Benchmarks the Beta
This article gives us a first impression of Starcraft II: Beta system performance by benchmarking the game on several computers. The results of these benchmarks can be used in conjuction with the [Guideline] SC2 Computers thread which contains information about user experiences running Starcraft 2. This article is focused on putting a number on that performance and to quantify differences in performance at the various resolutions and levels of detail.

Performance tests performed in this article can be used as an indication of expected performance on comparable systems. I used the systems most readily available to me and as evidenced by the specifications these are by far not top-of-the-line systems. In my opinion using top-of-the-line systems is not in the interest of most future Starcraft II players. Also it will be interesting to see if Blizzard made another game that is able to run well on older systems. These are my reasons for using this selection of sub-$500 systems.


Methodology

To test the performance in Starcraft II, I decided to judge performance by the frame rate. Measuring the following values:

• Minimum Frames per Second
• Average Frames per Second
• Maximum Frames per Second

Measuring the frame rate was done by running Fraps for 5 minutes and exporting the values to the graphs showed in this article. Because there is no built-in benchmarking tool, I tried to emulate situations consistent with the general experience of playing a game of Starcraft II. I did this by watching the same replay starting at the same point in time mark and keeping the screen centered on an area with the most activity. I benchmarked for 3 minutes watching a troop of Zerg units fighting midgame battles against a Terran army. During the replay approximately 60-70 units were seen battling on screen. A bigger performance drop can be expected when transitioning to late game play with 200/200 armies, but this was left out of scope for these benchmarks
The computers have not had any special preparations prior to installing Starcraft II, all systems have had windows installed for several months. Any additional programs not required for testing had been shut down prior to benchmarking. Because I do not have any 16:10/16 : 9 monitors available all tests have been done in 4:3/5:4 ratios.


System Specifications

For the test I used three of my computers, all with different specifications and form factors. I will shortly describe the relevant specifications of all systems in this paragraph. The names mentioned here will be used to refer back to the system specifications in the graphs of the article.
Both computers mentioned are desktop machines. Latest updates and drivers are installed. The first computer, the Q6600, is a PC I bought 2 years ago. My expectation is that this computer will not have any problems running Starcraft II. The other system, which I named A3000 for the purpose of these benchmarks, is a five year old computer that has had a upgrade in the form of a budget graphics card. Before starting the tests I am most interested in seeing my old system run Starcraft II; I wonder if it is even possible at all to run Starcraft II on high settings.

Q6600
Intel Core 2 Quad Q6600 (4x2.4GHz)
MSI P6N SLI Platinum, nForce-650i SLI
8GB OCZ Platinum DDR2 PC6400
Nvidia Asus EN8800GTS 512MB
Windows XP SP3

A3000
AMD Athlon 64 3000+ (1x1.8Ghz)
MSI K8N Platinum
2GB DDR2-400
Ati Sapphire Radeon HD 4670 512MB
Windows XP SP3

Small oops during installation on the older system:
+ Show Spoiler +

Resolutions and level of detail

I used three different resolutions for my benchmarks: 1024x768, 1280x1024 and 1600x1200. For every resolution I benchmarked two settings, low detail and high detail. The settings used can be seen here accompanied by a screenshot to give an idea what the difference between low and high detail is. When changing resolutions all other graphic settings remained the same.

Low Detail
+ Show Spoiler [Show Low Detail Screenshot] +
At low detail there is clearly a lack of detail on the units and the ground textures are a lot more plain. Compared to high detail this level of detail is a lot friendlier on the eyes, albeit not very pretty. Screenshots do not show the complete picture but what I have seen so far is that units seem to disappear a lot quicker when playing on low detail. It would not surprise me if most competitive players would select this level of detail.

High Detail
+ Show Spoiler [Show High Detail Screenshot] +
High detail definitely looks a lot more crisp and alive; the level of detail on the units is enhanced quite a bit--also there are a lot more doodads placed on the ground. Explosions tend to last longer and debris can be seen flying through the air whenever a Terran vehicle blows up.


Test Results

Here the test results will be displayed. The bar graphs are using a fixed width to provide visual clarity when comparing the results. On the horizontal axis are the frames per second measured and on the vertical axis are both systems. Each system has three bars displaying the minimum fps measured, maximum fps and the average fps for the complete duration of the benchmark.

+ Show Spoiler [Show 1024 Resolution Test Results] +
It is shown here that there is the expected difference between high and low detail. Framerates are roughly halved by selecting the higher setting. That’s something to keep in mind when configuring Starcraft for optimal gameplay.

+ Show Spoiler [Show 1280 Resolution Test Results] +
It is interesting to see that there is no reduced performance when switching towards a higher resolution. The most significant drop is seen in the max frame rate on the Q6600.

+ Show Spoiler [Show 1600 Resolution Test Results] +
Even at the highest resolution numbering almost double the amount of pixels it does not reflect back on the frame rate of Starcraft 2. I have yet to explain this behavior, but obviously this is interesting to say the least. There is a slight drop, but that could also be explained by the nature of the tests.


Conclusion

Based on the test results it can be noted that Blizzard put in a lot of effort to ensure everyone can enjoy Starcraft II. Although the results might not completely represent the situation after launch, but it can be used as an indication for someone considering an upgrade to their current system. I was pleasantly surprised by the performance of my older system, the results by far exceeded my expectations.

Requests for additional test or specifications can be placed in the topic below this article and I will see if I can squeeze it into my rigorous schedule of actually playing the beta. Sheets with raw data gathered during the benchmarking are available on request. I welcome any suggestions and feedback, so discuss!

Hot_Bid: Chrono Boosts and Underdogs
I played a lot of 2v2

Insane and I did quite well on the 2v2 Ladder. We were #1 for a day or two, and after a week we're the #2 team in Platinum Division 1. We play an economy-oriented style, and our goal is to react to the opponent's first rush. The main basic "rushes" are Zealot/Marine, Marine/Roach, and Zealot/Roach. We rarely face teching builds, or teams that play passively like us (except on Twilight Fortress, but that's just because of the shared mains). However, you can be very successful simply massing and rushing with basic units if you have streamlined builds. I would say you can reach Top 10 in any Platinum 2v2 division if you can just transition smoothly and micro your first attack well. Against the better pairs, a pure rush build probably won't work, but against the vast majority of teams, you'll be at the advantage if you attack first.


Notice that Z/X dominates the 2v2 Ladder, just like in BW.

2v2: General Tips

• Use the chrono boost on your ally's stuff. The building that by far receives the most benefit is the hatchery, as you get your lair and extra queen larvae faster. Insane often uses the boost on my upgrades like ling or baneling speed or even something like +1 air attack. [Editor's note: Blizzard has since patched this, you can't boost your ally's stuff anymore.]
  
  
• If you're P, always build a pylon in your ally's base so you can warp in units in case he gets attacked. A sentry is especially awesome as it can block the ramp by itself.
  
  
• Pay attention to Overlord placement! This seems obvious, but so much of 2v2 is positioning and army movement.
  
  
• Share resources. It's not uncommon for us to ask each other for resources. The obvious case is when you only have a worker and not enough to build a Nexus, but don't hesitate to send minerals if your ally's econ was hurt. A few hundred minerals is far more valuable to a player hurting for workers.
  
  
• Give control to each other when needed! Many players don't know this option exists, but you can allow your ally to control your units from the menu in the top right.

Race Specific Strategy

Against a PP or TT or ZZ, just build your respective counters. We mass zealots and roaches against PP. I go speedling/baneling while my ally techs to colossus against TT. Playing against ZZ can be dicey if they both roach rush, so you have the option of attempting to hit one super early, or holding out for robo tech.

Against Z/P, the Zerg almost always does some sort of roach build. We have two options: 1) build more zealots and roaches and outmicro, or 2) tech to immortals, the hard counter to roaches. The second option is far more difficult to pull off, especially on smaller distance maps or ones without ramps, but when it works, it can instantly win you the game. You are constantly trying to get into the "right" fights, as in your roaches vs enemy melee or your immortals vs enemy roaches and stalkers. Your P ally should have a nice army mix -- make sure to include sentries for the shield bubble ability.

Don't neglect your econ and tech transitions. It's easy at first to just rush rush rush off cookie cutter builds, but this sort of all-or-nothing play will only get you so far. You'll need a solid mid and late game plan to beat good teams.


One of our losses. Chrono boosted muta rush fails, and they outmacro us.

T/P is the most straightforward. We try to bust the P with roaches+ally or bust the Terran with speedling/baneling+ally. Try not to let their armies meet, and be careful about reapers. T/Z is the most difficult -- we're not truly sure what to do against this combo yet. We always can just 10pool + proxy gate rush, or bust the T's wall in. Just don't let that marine/roach army meet, because that unit combo in large quantities is so difficult to deal with. The roach is the perfect compliment to the marine, as it tanks all the damage and kills everything the marines are weak against. The rush also comes at a timing that is faster than the tech needed to deal with them.

Four Fun 2v2 Strategies for Twilight Fortress

Twilight Fortress is the map where you share mains with your ally. There are four total bases in your shared area, accessed by a single choke. It's a very fun map and allows for some truly great strategy diversification. Some of these aren't viable, but they are definitely fun.

• Proxy Gate + Sunken Rush
  Given the long distances on this map, most Zergs feel very safe doing some sort of 13 drone opening, so proxy your gates with a 10 pool. Send two drones ahead of your lings to sunken rush too. Make sure to wait for three zealots and go with the lings+drones so your opponent isn't tipped off and spends all his larvae on drones.
  
  
• Boosted Proxies
  The screenshot says it all. Another less fun one involves proxying two stargates for three void rays instead, but your ally will need to send you gas. [Editor's note: Blizzard has since patched this. You can't boost your ally's stuff anymore.]

Yes, that's a double chrono boost for reapers+upgrade at the 3:30 mark.

• Boosted Lair
  
  Go Z/P. Let your opponent's scout see no gas started in your main. As soon as you kill off the scout, make two gas and get lair. Your P ally chrono boosts the lair, and you can either go spire or nydus and completely surprise your opponents with the timing. [Editor's note: Blizzard has since patched this. You can't boost your ally's stuff anymore.]
  
  
• Macro Micro UMS
  
  Ally 1 (preferably Protoss) goes pylon, nexus, nexus, taking three of the four bases. Ally 2 builds a lot of production facilities. At the 5-minute mark, when resource sharing is allowed, Ally 1 constantly sends all resources to Ally 2, and also uses his nexuses (which all have saved energy) to chrono boost all the production buildings. If you want to be especially nasty, hidden expo to your opponent's gold mineral expo.

1v1 Adventures

After a good hundred games of 2v2, I did manage to get in some 1v1s. I had not touched the 1v1 ladder since the first day (when I went 9-1 to get into Platinum Div1), so my rating was only 1000. I don't know if the AMM was on the fritz, but I got matched up with Tasteless (1300+ rating) for my first game.


Roaches can't hit air, but in ZvZ, they don't have to.

I lost to Tasteless in a ZvZ where both of us admittedly didn't know what to do other than build mass roaches. I won the first roach fight but he won the second. As inReach said in his thread after facing Artosis: "I may never make another muta in ZvZ." They simply cannot kill roaches fast enough and eat up your gas.

While I consider myself far better than my 1000 rating, I was a relative noob at 1v1. It was weird playing someone who was #2 in his division right off the bat. So imagine my shock when I hit the "Find Game" button only to draw 1650+ rated #1 Artosis three times in a row. How does this even happen? Does it know I'm not really 1000 from my 1600+ 2v2 rating? I can't imagine it's beneficial to either party to have players facing each other with a 600 point rating difference.


Artosis jinxes himself.

I get owned very badly by Artosis in two games (ZvZ and ZvP), but in game three I successfully hid my lair from his scout and killed him with mutas. I think I set him back fifteen wins. I'm sorry Artosis, I never wanted to play you.

See you guys next week!

Insane's Thoughts on Starcraft 2.1

Amidst cries of dismay as the Battle.net server gave a mere five minutes warning before going offline, players realized that this may herald the arrival of the first ever Starcraft 2 patch--and what do you know, the fans were not disappointed. Or rather, quite a few were disappointed, but that was because of what the patch actually contained!

First though, there's a definitely cool feature I want to call out which isn't mentioned in the patch notes!



What do the arrows mean? Well, we don't know over what time period it's referring to going up and down as right now, but it's definitely a step in the right direction!

Let's take a more detailed look at what the patch actually did and how this is going to affect the game.

Protoss
The Protoss race has been by far the most popular thus far in the Beta. Whether it's actually the strongest is still under debate by many, especially those complaining about the strength of Mutalisks and other such units.

Observer
Gravitic Boosters: The cost of this research has decreased from 150 minerals and 150 vespene gas to 100 minerals and 100 vespene gas.
Maybe this will encourage people to get Observer speed more... I feel this change isn't very significant overall. There is simply such a huge lack of cloaked units in Starcraft 2 that Observers don't need to keep up with your army so much. No longer do we see Spider Mines all over popping up and ruining careless Protosses. No longer are Zergs containing with Lurkers and sniping Observers.
Until cloaked units and sniping Observers plays a larger part in the game, things like this are simply unimportant and will continue to be skipped in favor of more valuable upgrades / units.

High Templar
Phase Shift replaced by Feedback



Looks like the Dark Archon is worming his way back into SC2! As weak overall as the Dark Archon was, Feedback was a pretty good spell in a few limited situations. I have yet to use this spell in SC2, and I do wonder if it'll end up being useful here. For not much more energy, you could simply cast a Psi Storm. If you look at the units that currently define the matchups, most of them do not use energy. It looks to me like that ability that is "kind of" useful, but will be overshadowed by the fact that Psi Storm is the other High Templar spell.
Still, it's better than Phase Shift, so it's at least a step in the right direction...


Colossus
The building pathing radius for this unit has decreased from .75 to .5625.
I don't have precise information for exactly what this will mean from a functional perspective, but I can say the change is a nice one. The way the Colossus is, he should be pretty good at walking around stuff. I've had several games where the Colossus got stuck because he's fatter than an SC1 Ultralisk, so I'm glad to see this change.


Mothership
Vortex: Energy cost increased from 75 to 100, radius decreased from 3.0 to 2.5
Temporal Rift: Removed
Wormhole Transit: Removed
Mass Recall: New ability, teleports all of the player's units in the target area to the Mothership.
The biggest change for Protoss, without a doubt. This change basically changes the Mothership into a big slow Arbiter and removes its fantastic escape mechanism. I do wonder if, since they removed two abilities and only added one: will they add another ability in a later patch?

Regardless of having another ability added, I feel that these changes affect the Arbiter Mothership negatively. The Mothership lacks the speed of the SC1 Arbiter and there's only one Mothership, so Recall is not such a surprise weapon. It's really too early to say whether Mass Recall will be too strong, not strong enough, or just right. What I do know is that it dilutes the flavor of the Mothership and reduces its harass capability--before you could be aggressive and harass with it, secure in your ability to Wormhole Transit away.

Nexus
Chrono Boost can no longer target allies.
Was this considered a bug before? I don't know, but Hot_Bid and I had some cool strategies we worked out involving Chrono Boosting an ally's Lair upgrade, as well as Zergling Speed and so on. One can argue that the other macro mechanics were less "adaptable" for allies, but it was a part of the Protoss experience in 2v2 and added a lot of additional strategies and possibilities that simply were not there otherwise.

It probably won't affect 2v2 hugely, but it does detract from some cool strategic novelties and builds, and for that I'm sorry to see it go.

Gateway
The build time of this building has increased from 50 to 65.
This is quite a major change to me. Previously a Zerg going 13 Hatchery at his Natural was extremely vulnerable to the Protoss risk-free killing him--a Pylon/Gateway in the main base (in case the Zerg opened with a Pool instead) could quite safely be followed up with two proxy Gateways outside the Zerg base, which combined with Chrono Boost was extremely strong due to the fact that Roaches come out so late with a Hatchery first opening.

This change is going to make Hatchery first a much easier defensive possiblity-perhaps even feasible. The current playstyle has mostly been gravitating towards 1-base Protoss play with strong timing attacks. With the Gateway proxies not being so easy a transition against Hatchery first expansion play, this may end up pushing PvZ into a more macro-oriented style.
As a Protoss player I am obviously sad to see this change, but I think it's probably a good change, as Chrono Boosted proxy Gateways really were pretty ridiculously strong.

Zerg

Infestor
Fungal Growth: Damage decreased from 48 to 36, affected units can't burrow
Neural Parasite: Now channeling spell, lasts up to 10 seconds; all units now vulnerable.
Terran Infestation: Energy cost halved from 50 to 25 and Infested Terrans now spawn 1 at a time.
I have to confess--I've never played against anyone who used an Infestor, and I have very limited experience with them. The changes outlined above are mostly positive--I imagine the damage decrease is partially to prevent them from one-hitting (un-HP-upgraded) Marines.
Infested Terrans are frankly quite weak from what I've seen. They're not the Infested Terran of SC1--they now seem to be pretty weak. The unit still feels a bit gimmicky to me (as Infested Terrans always are....), and I can't see Neural Parasite being very useful unless it has an ungodly cast range.
As soon as you take over a Thor, the Terran will see your Infestor channeling away, and will instantly kill it--bam, useless. OK, its mana cost may not be as prohibitive as Mind Control, but in every single other aspect, it's a much worse spell.

I appreciate that they are trying to make the Infestor a useful unit, but I don't think these changes are it.

Terran

Viking
The armor value for this unit in Fighter Mode has decreased from 1 to 0.
The damage done by this unit in Fighter Mode has changed from 6 (+8 armored) to 10 (+4 armored).

The main change here is that Vikings now suck much less against Mutalisks. Vikings in ground mode are already quite effective units, but have struggled a lot against Zerg, since once the Zerg gets Mutalisks they start overturning the Vikings with ease. I like the change--I think it'll help make Starport play a lot more viable against Zergs. Banshees are already quite good; this will help them not become useless beyond early game unless accompanied by ground forces.

Miscellaneous
Addressed an issue causing people to receive an “Internal Battle.net Error” message with their game client.
Muting your microphone will no longer cause your microphone to be turned off in the operating system once the program has exited.
Please note that the voice chat functionality has only partially been implemented. There are many known issues with voice chat that we expect to address in a future patch.
The “Medium” graphical settings were reconfigured to work better on appropriate machines. The video settings “auto-detect” functionality has been reset as a result.
Added a frame rate limiter to the game menus to prevent some graphics cards from running at higher frame rates than necessary
We are still missing a lot of basic command functionality in Battle.net (/w ), and the biggest thing that bugs me right now is that the latency is, frankly, quite bad.
It's not like playing on Extra-High Lat, but it's still not like playing on #LL either. If 3rd party groups can achieve such latency settings for an ancient game like SC1 without even having the source code, then Blizzard should definitely do so on SC2, as well.

What Was Missing?
Well, there are a lot of things that were not in this patch that many people were hoping or expecting. On the whole, the patch changed very little apart from weakening Protoss early-game options. The things that many people have been calling for were a nerf to Roaches and/or a nerf to Mutalisks.

I'm glad Blizzard did not cave to the cries of players calling for this. People right now have a tendency to lose to something many times in a row trying to play like this is SC1, or simply not understanding how to beat it. Even good players, especially those who have had a history of success in other RTSes, are guilty of this--they get far because of mechanical skill and figuring out some good strategies, but can grow quick to call imbalance rather than simply asking others and experimenting with various unit combos until working out a solution.

My personal opinion is that the patch didn't miss much. Some changes were pretty insignificant, as were some of the spell changes. These are minor complaints, however, and I feel that they did not significantly affect the (quite good) balance. I'm especially glad they did not nerf Zerg into oblivion with the wailing about Roaches and Mutalisks.

I give Blizzard a grade of A- for their first SC2 patch.