Talking Balance with David Kim - Page 68

On May 02 2012 21:58 SarcasmMonster wrote:


I should have posted both. Remember to look at sample sizes before drawing conclusions.

On May 02 2012 23:21 Big J wrote:


you don't need to, that's what the standard deviation in the graphs is for.

On May 02 2012 23:28 SarcasmMonster wrote:


Standard deviation presumes a binomial distribution, and that each game is independent of each other. For a small sample size, these assumptions can be dangerous.

On May 02 2012 11:51 YosHGo wrote:


Oh yeah considering that stargate opener after FFE is pretty good way to mess with stephanos roach style so yeah i think i will just listen to MC and so should you before posting false statements

On May 02 2012 23:30 Shiori wrote:

Not to mention that in the context of a BoX, the games are not independent of each other.

On May 02 2012 23:28 SarcasmMonster wrote:


Standard deviation presumes a binomial distribution, and that each game is independent of each other. For a small sample size, these assumptions can be dangerous.

On May 02 2012 23:43 Big J wrote:


Na, I don't think this is dangerous. I mean, binominal distribution is what we have here, it's just a question of the exact probability parameter. As long as the creator of those stats has chosen something that is somewhere around 0.4-0.6, it should be pretty exact.
And games are not independend, but they will never be, you will always have certain metagame influences, certain personality influences and influences of playing a set, but a lot of those things don't change if you had bigger sample sizes of the same time periode. F.e. the overall XvY metagame is the same if you have 100 matches in a month, or if you have 1000 matches in the same month.
(if anybody would take statistics that literally, then there would not be any form of applied statistics. At some point you will always have to put up with certain assumptions, and those that are made here, are pretty common for applied statistics)

On May 02 2012 23:32 Rumpus wrote:


Lol seriously? That is your argument? An entire tech tree is validated because it can possibly fluster one play style in ZvP? It doesn't work at any other point in that match up, transitioning out is a massive risk, and it is quite easy to shut down entirely which is an immediate GG. And we didn't even get to TvP or PvP. So I don't think what I said is too far fetched or in any way a "false statement."

On May 03 2012 01:38 Shiori wrote:

We don't have large enough sample sizes (except maybe in the International winrates) to assume these things balance each other out, though...

On May 02 2012 22:18 jdsowa wrote:
Lyyna's got it right. People are stuck in a bio mindset. Bio units are throwaways, mech/air is more like playing the P deathball. You want to make expensive upgraded units, and hold onto them until you're maxed.

Early and mid-game, contain P with threat of hellion run-bys. Tech up to 3/3 and an ultimate unit composition of siege tanks, with a wall of thors in front, and banshee support. I'm not sure P can do anything against this. If you don't feel safe taking your 3rd, then maybe get a planetary or two.

On May 03 2012 07:36 TORTOISE wrote:
Give us increased Tank damage. PLEASE AND THANNK YOU , BLIZZARD ♥

On May 03 2012 07:24 zmansman17 wrote:


This post lacks so much knowledge and clearly you do not play Terran. As has been stated on here numerous times by pro terrans and rank 1 master terrans like myself, mech does not work versus Protoss. If it did, you would see Terrans using it: plain and simple.

If you haven't noticed, as Terran we do have a very flexible race, so what we lack in raw power, we make up for in versatility. This leads to a certain type of player that is attracted to Terran and a certain style that jives well with the race. We are Always trying new things and that is why Terran has been successful. Part its player base and part the race design itself. Mech does not and will not work until adjustments are made. Any statement otherwise from a Protoss player is very mis-guided and un-informed of the panopoly of posts made on mech TvP

On May 03 2012 08:02 convention wrote:


Applying your logic about why mech doesn't work in TvP would imply that mech doesn't work in TvZ either. Also applying your logic (about people not currently using mech), then the current strategy is always the best strategy, because if something else was the best everyone would already be doing it. But people find ways to make new strategies work all of the time. While I think that there might be problems with mech at the highest level, Lyyna has shown that Masters and below it is definitely viable, which I think really shows something. Everyone always says charlots + immortals + ... is why mech can't work. But if mech works at the lower levels, then it can't be that it is those units that make mech unusable (assuming masters protoss players use charlots)

On May 02 2012 10:46 Rumpus wrote:


First off look at the units and tech tree. If you think the race is okay, examine the Hydralisk, Roach, Corrupter, Ultralisk, and how many units we have compared to the other races. These units are all dull, or useless, with unimaginative design, terrible utility and extreme "situationalism." The flavor of the race, the Zerg theme of "the swarm," has been taken away entirely. Hell, not only is the Queen the source of a boring, mundane game-play chore, but is directly counter intuitive to that "swarm" style. Also, flexibility in composition and strategy is virtually nonexistent.

Warp Gates make sense for Protoss but it ruined that powerful "high cost, low count" army mystique. Oh and their tech trees are a mess too. Considering going down one of them (Stargate) is almost always a poor choice.

On May 03 2012 08:23 SupLilSon wrote:


Stop. Please. Stop. How many of these posts are there going to be? Yes, there are a few masters Terran players per server that use mech TvP. I've read the guide, I've tried mech. Yes, you can win games with it. Is it "viable"? I guess it's viable if your definition of viable is having over a 0% chance of success. The fact is, for all of bio's shortcomings, it is still the better choice. Why should we use an inferior strategy just because it has the potential to win? For the 1000th time. There is a reason mech seldom appears in professional matches. If you don't think Terran pros have tried mech and explored it as much as Lyyna, then you're hopeless. But yea, I mean, go ahead and keep questioning the collective knowledge of every professional Terran. I'm sure you have got it all figured out.

On May 02 2012 16:22 Big J wrote:
Hydralisks are weak, but this has nothing to do with any form of "incomplete design", they are simply not playable as a straight up combat unit right now.

On May 02 2012 16:22 Big J wrote:
Corruptors are boring, I agree, but they are needed.

On May 02 2012 16:22 Big J wrote:
There is absolutly nothing, that makes Zerg (or Protoss) more "incomplete" than Terran.

On May 03 2012 09:41 sunprince wrote:


Agreed. They just have awful stats; hydralisks in BW weren't anything special either, but their low mineral/gas/supply cost and speed upgrade made them an effective bread-and-butter unit.



Not necessarily. You could fill the same role with a less boring unit.



The main reason Zerg feels "incomplete" is the lack of a disproportionately cost-effective defense to slow/stop attacks and counterattacks. Terrans revolve around this with siege tanks, repaired bunkers, and MMM, while Protoss have splash damage like Colossi, Storm, and Archons, but whereas Zerg had lurkers and dark swarm in BW, they have nothing of the sort now. This lack of defensive ability drastically reduces a SC2 Zerg player's strategic options, resulting in the race feeling "incomplete" and/or "unfun".

On May 03 2012 09:45 teamhozac wrote:


Spine crawlers and spores are both great defensive units that can be uprooted and relocated wherever they are needed, not sure what exactly you want as defense but you probably have theb est defensive structure in the game