[GSL] Code S Ro8: Day 1 - Page 94

On January 18 2011 19:38 imbs wrote:
did u guys not see how tester totally outplayed him in game 2, his blue flame hellion drop/push did nothing. doesnt matter for terran tho. if a protoss dt rush fails they lose the game. if a terran blueflame/banshee does nothing it doesn't actually matter nearly enough as it should.

ofc when a good terran plays hes gonna win everything. just look at how well marineking did - making 1 type of unit with good micro was enough.

On January 18 2011 20:06 Ysellian wrote:


Tester didn't outplay MVP. He simply did a good job blocking blueflame harassment using forcefields.

On January 18 2011 20:21 imbs wrote:

sry but when u invest in harass and it costs ur opponent nothing to crush it thats the very definition of outplayed

On January 18 2011 20:29 Ysellian wrote:


Watch the vod again please. He loses 2 probe + 1 Sentry against 2 hellions and "invest" is a funny word to use when he clearly uses hellions in his army later. I might also want to add that DT rushes fails, but DT harassments do an amazing job at times. Tester used it quite well against Boxer and part of the Tech to get DT are needed for HT anyway.

On January 18 2011 21:06 Victim wrote:
The hellions did exactly what they were supposed to: force the Toss to make Stalkers, making the marauder+PDD attack much better.

On January 18 2011 20:50 imbs wrote:

2 probe + 1 sentry vs a hellion drop? u kidding me that should him way ahead. and the fact that hellions fit into his army so well because of how they bbq zealots further emphasises my main points; terran harrass doesnt cost nearly enough techwise or investmentwise, and that their rushes are way too strong

On January 18 2011 21:35 netherDrake wrote:


2 hellions for 2 probes and a sentry seems rather fair to me.

On January 18 2011 09:13 redemption wrote:
This strategy works very well against bad and mediocre players who will make bad decisions against you with worse hands. However, to a good player, this makes you very predictable. And if you become predictable, it's not very difficult to figure out a counter-strategy to defeat you. So what many good players do is play unpredictably. They not only raise with their good hands, but they sometimes raise with their bad hands too. This is called polarizing your range.

On January 18 2011 19:38 imbs wrote:
did u guys not see how tester totally outplayed him in game 2, his blue flame hellion drop/push did nothing. doesnt matter for terran tho. if a protoss dt rush fails they lose the game. if a terran blueflame/banshee does nothing it doesn't actually matter nearly enough as it should.

ofc when a good terran plays hes gonna win everything. just look at how well marineking did - making 1 type of unit with good micro was enough.

On January 19 2011 01:21 bbkpoker wrote:


Hi, I'm a top 100 rated online tournament poker player, longtime starcraft fan and lurker, and just needed to chime in to correct this because you have the concepts correct but the terminology wrong.

polarization
- 7 dictionary results
po·lar·i·za·tion
   /ˌpoʊlərəˈzeɪʃən/ Show Spelled[poh-ler-uh-zey-shuhn] Show IPA
–noun
1.
a sharp division, as of a population or group, into opposing factions.

They are actually balancing their range, not polarizing it. Having a polarized range means that you take such an unusual line its very unlikely that you have any semi-strong hands such as one pair with a decent kicker, because the line you took makes no sense in terms of getting value from slightly worse hands in a traditional sense, and as a result your range is very unbalanced toward being the extremes of bluffs or super strong hands.

I guess the closest Starcraft analogy would be players who near universally play one style with little variation even when circumstances call for adjustment.

Idra is absurdly polarized to a macro style for example, and bitbybit is very polarized to all ins. A balanced player will play primarily one style (generally macro) but be capable of the occasional marine rush or 4-gate or whatever to keep their opponents on their toes re: the possibility that they might just get rolled if they don't prepare which is strong because it forces them to build some amount of preparation for it at the sacrifice of some economy. This is precisely what makes IMMVP so good. If you take for granted that he is a sick macro player (which he is) and as a result that he is unlikely to attack early and go for the super greedy FE that he Tester went for, he is balanced and egoless enough to NOT think "zomg, hope people on Team Liquid don't think I'm a noob" as he builds the perfect one base all in to crush you. He may prefer to play macro since that is an arena where few can keep up with him but if you play in an exploitable style, he has no qualms about exploiting you.

To go back to poker for a second, this need for some amount of balance to stop your opponent from being able to play game theory optimal vs a narrow range of hands is exactly the same concept of why the best players will raise or 3-bet primarily with stronger hands toward the top of their range, but occasionally mix in speculative hands or complete air from time to time as well. If a player only opened a specific range of hands (for example, 99+/AQ+) then their range of hands would be extremely predictable and it would be very easy to either just flat their raises a bunch speculatively until you find a flop you've cracked premium hands on, or to bluff them off their hands frequently on coordinated boards that are fairly bad for premium hands.

All in style builds intuitively seem to be "countered" more than macroing builds since macroing is considered the "standard" for Starcraft play, and people often misinterpret just building 4 gates and killing your opponent to be a specifically "aggressive" play, in reality though you can just as easily look at a one base Terran all in or a 4 gate rush vs a fast expand to be defensive, as they are essentially defending themselves from a mid to late game against an opponent with superior tech and economy against whom they will almost certainly lose. Super greedy macro builds such as Idra's beta build where he'd go for 14 hatch 100% of matches even on Steps of War, or, in this case the Tester/IMMVP match where Tester went for a FE with no Core are countered by amassing a large army and exploiting the timing window of vulnerability where the FE'r has just gotten their economy going but doesn't have the appropriate units to defend.

You could also talk about polarization in re: to build orders. For example, if you run into a terran base and see him taking his natural, two rax and a factory, you could assume he has a pretty balanced build that is capable of going a number of tech directions. If on the other hand you scout and they have a bunch of marines on the ramp, no gas, etc, their build order range is extremely polarized to being either a mineral only fast expand, or some sort of proxy barracks marine and/or bunker rush. Because their range of builds is so polarized to extremely specialized builds favoring either economy or aggression rather than a balance, additional scouting becomes exponentially more important since either of those specialized builds will likely roll a balanced build. If they are essentially all in on aggro then you need to build bunkers, get sentries, build a spine crawler or whatever else or you will lose to early pressure, and if they are super FEing you need to essentially either immediately re-arrange your build order around taking your own expo (or two if you are IMMVP) to keep up/get ahead, or you need to get up production facilities fast and get prepared to all in your opponent. Those are the two ways to respond. Simply continuing building your rote-memorized build will lose to EITHER way your opponent has chosen to specialize, which is the strength of building in a polarized way. The weakness of course is that if your opponent does manage to scout what you are doing enough and adjusts quickly, pretty much you're going to lose. If they specialize toward aggression and their all in build is defended they have almost no chance of winning due to being significantly behind in economy/tech and if they specialize in economy and you all in them, then they just die then and there.

Sorry for the TLDR, etc.

On January 18 2011 21:35 netherDrake wrote:


2 hellions for 2 probes and a sentry seems rather fair to me.

On January 19 2011 05:07 HinagikUx wrote:
honestly i agree with imbs. The idea of going hellions is to crush ur opponents economy. for 200 mins, mvp coulda done way more dmage, but the FF's were brilliant and tester stopped it. For 200 mins you can kill 15+ probes easily. If your drop fails, you can transition to marauder because you know ur opponent is going stalker heavy. You already have the blue flame tech, so that will crush any zealot base army. And since you already had hellions in the protoss base, you know what tech hes going and if there are any potential collosus or anything like that besides gateway units...its pretty much another fullproof scout on top of scans. There is just too little risk in going hellions. Marginal benefit = very high, marginal cost = low. No other race can do that. Theres a reason theres no more protoss left in both CodeA AND CodeS...