To the Bitter End

Four remain.

Beginning the week was the titanic clash of the two top teams in the tournament so far, which was a shocking rout. (More on that match in next week's post.) Then came the carnage, Gone in one night was the team led by the best zerg in the world, and the next night claimed the reigning GSTL, KSL, and IPL TAC champions. Surviving are the plucky underdogs of Team Liquid, led by TaeJa and HerO, and supported by an irrepressible foreign line-up. Startale has also made it a round further, with their line-up deep with GSL experience. Tomorrow, these two excellent teams will go head to head. One will keep their dreams of traveling to San Fransisco alive, the other will have to content themselves with selecting a player to go to IPL5, and thinking of what might have been.

In this IPL TAC newspost, Ver of (TAC qualifier) x6 joins us to recap and analysis of the Liquid-MVP match, tree.hugger writes his on-the-fly notes from watching the Startale-Prime broadcast, and NrGmonk takes a look at what epic things await us in tomorrow's match.

Losers Bracket Semi-Finals Review

By xSixVer and tree.hugger

      Game One: Liquid`TaeJa (7) < Antiga Shipyard > MVP.TAiLS (1)
This was one of these games where due to some small risks and bad luck nothing went right for one player. Tails didn't scout until his first zealot, aiming for a 2 gas blink or dt build. Ordinarily this would not be a problem. However, Taeja had decided to employ the rarely used proxy marauder build. Thus Tails sent his first zealot to Taeja's main to meet a supply wall, while Taeja walked up the Protoss ramp with two marauders. Instead of meeting a stalker and zealot, the marauders sniped the astonished stalker for free, and then kill the isolated zealot later after it slowly lumbered home. Probes were no match for the three marauder force with Taeja's clever positioning using the natural gases as a makeshift choke. Tail's tech allin gave him few units in the initial few minutes, and with the few that he had mercilessly killed piecemeal, his hopes vanished.


      Game Two: Liquid`TaeJa (1) < Daybreak > MVP.Vampire (7)
Taeja opened reaper expand, a generally weak opening that is favorable against the popular nexus first on Daybreak without committing too much. However, Vampire chose a standard stalker expand, so Taeja went into the midgame at a slight but enduring disadvantage.

Exploiting the extreme ease of defense on Daybreak, Vampire took a slight risk in making a forge after nexus instead of the standard two gateways, giving him a very fast attack upgrade. Like many high level TvP's on Daybreak between risk-averse players, absolutely nothing happened for the entire game. Taeja aimed for a 2 base allin, taking his 3rd command center after six barracks, but in response Vampire took a late third as well and left no holes. Vampire tried to trick Taeja into overproducing vikings by making a single colossus without range and transitioning into templar on two bases with double forge. However, Taeja's acute sense and constant threats let him feel Vampire out and judge the latter's intentions.

Without any holes to take advantage of and neither player in the mood to try to harass and risk losing the 1a battle, both players went up to four bases without any action. After both players maxed and started banking resources, Taeja sniped an observer and then marched around Vampire's cannons in the middle to threaten his third base from the top. Vampire hurriedly reacted to the threat to his third base, but in his haste his army became disorganized and fell easily to Taeja's prepared position. Vampire lacked the infrastructure to recover from such a crushing defeat so close to home, and Taeja a-moved to victory a minute later.

      Analysis: Neither player took any risks the entire game. Vampire possessed a very strong position, but made a couple of fatal errors that threw away an advantage. First, he lacked the scouting necessary to see Taeja's army movement into the exopsed area above his fourth. Second, his main army only contained one observer, which was not well hidden and easily killed early on. Taeja's cloaked ghosts were able to have free reign with no observers and no storms. Third, he did not stagger his templar or approach from multiple angles like Parting has demonstrated, and thus was not able to deliver a single storm. Finally, Vampire lacked the gateway count - despite a sizeable bank - to perform the dreaded mass zealot midbattle reinforcement.

In the end, Taeja simply made fewer mistakes than Vampire and took a solid and simple win. If both players choose to play such a rigid game, then the outcome is determined mostly by who can get the slightly better position for the big fight. He correctly sensed and scouted Vampire's build and made the proper numbers of vikings and ghosts at the right time. Perhaps most importantly, he consistently sniped observers, keeping Vampire in the dark and off balance when he made the decisive attack.


      Game Three: Liquid`TaeJa (1) < Cloud Kingdom > MVP.Sniper (7)
Sniper took a slightly late third and made fewer queens than normal. On the other side, Taeja rolled the dice with his favorite TvZ opening: triple cc into hellion/banshee with a heavy marine/late tank midgame. While very vulnerable to early allins, if the Zerg chooses to go for the standard three base saturation it gives the strongest possible economy while still retaining some form of early threat and scouting. However, Taeja only landed his third cc at the 11 minute mark, a cautious and somewhat illogical decision given his scouting information (3 base saturation) and map dominance with hellion/banshee, so he did not actually keep an economic lead.

In the end, none of the above mattered. Sniper curiously went for muta/ling/bane on a map most unfavorable to that style. After some light harass, Taeja postured in the middle while building up tank count. However, highly uncharacteristically, Taeja did not prespread his marines nor send a scout ahead. His army sat below a ramp in the middle unsieged while he was busy macroing, and was promptly encircled killed by Sniper's marauding ling/bane force. The defeat was so complete that Sniper easily converted the advantage into victory with several attacks into Taeja's third.

      Analysis: Mistakes like that as Terran are simply not allowed and cannot be recovered from. Against dedicated muta/ling/bane Terran wants to stay at home and wait for critical tank count mass before engaging. A critical part of Taeja's TvZ success is his abiltiy to successfully feign aggression, delay creep, and keep the Zerg off balance with his large marine/medivac force until he has the tank numbers ready to make a serious push. However, this type of play is much riskier than simply turtling at home and waiting passively. Taeja's build demands this kind of proactive play, as he sacrifices tank count for more marines and medivacs early on. Here, we saw what happens from a moment of inattention with this highly risky strategy.


      Game Four: Liquid`Zenio (11) < Tal'Darim Altar > MVP.Sniper (1)
Zenio went for 15 hatch, while Sniper chose 10 pool speed, a hefty build order win for the 10 pool player. Instead of hitting the natural hatchery with his faster zerglings Sniper instead elected to attack the drone line. With excellent micro Zenio managed to repel the zerglings with no losses until his own spawned; unfortunately, a great deal of mining time was nonetheless lost.

Back at home Sniper had pulled drones off gas and was quietly massing zerglings in his main, going for a no bane speed allin. Once speed finished he rushed into Zenio's natural. Zenio started a spine when his overlord saw the zerglings leave the natural, but that was far too late. Zenio's natural was blown over by a torrent of zerglings before the spine could complete.

      Analysis: Sniper won the opening coinflip and was lucky his allin was not scouted or reacted to properly. Zenio defended well against the initial zerglings but made a critical mistake in not fighting for the watchtower and/or sending a scouting zergling into Sniper's base. With his overlord spotting no second hatchery from Sniper, he should have put down at least one spine asap. Taldarim is a notoriously difficult map to stop allins in ZvZ and Zenio simply did not prepare quite enough. There are dozens of ways to die instantly in SC2 if a single thing goes slightly wrong: this was simply one of many examples.


      Game Five: Liquid`Ret (5) < Ohana > MVP.Sniper (11)
Sniper chose a slightly different allin this game, a 10 pool no gas spine rush. Unfortunately for the MVP upstart, Ret chose to go for 15 pool 15 hatch, a much safer opening against 10 pool. As soon as Ret's overlord saw the zergling/drone combo leave Sniper's natural, he immediately canceled his hatchery and threw down two spines in the mineral line. When Sniper tried to plant his own spines, Ret correctly pulled his army of drones to fight while repositioning the spines. As his spines were much faster, he easily held.

      Analysis:: Unlike Zenio, Ret got a better build order scenario and reacted properly. This time, Sniper's coinflip landed on heads instead of tails.


      Game Six: Liquid`Ret (11) < Entombed Valley > MVP.Keen (7)
Ret went with a normal hatch first followed by a later pool and gas immediately after, not afraid of a bunker rush. Keen became nervous after seeing Ret's reasonably early gas, and sensed an allin. Instead of the usual rax cc double gas reactor hellion build, he instead made a little known deviation, getting an ultra quick reactor on his barracks, very fast third command center, and went for banshees without hellions. Instead of the typical hellions, however, Ret's very early roach/ling allin met a well placed bunker and a hoard of marines. With only the roaches able to attack the packed bunker and too many marines on the high ground, Ret could not punch through.

      Analysis: Ret wanted heads, Keen chose tails. Both players took risks but this game Keen won the blind build order coinflip. Ret could also have simply gotten speed and taken a third base, getting a solid position against Keen's late tech and nonexistent hellions. However, the roach/ling allin floundered horribly against Keen's opening, which was specifically designed to beat the allin. Keen's build would have automatically lost to a slayers style 1 or 0 queen 6 min~ baneling bust and would have severe issues against a delayed ling/bane allin, but was safe against the very early roach/ling.

Ret could have still tried to salvage the game and fought back from a disadvantage had he selected to drone instead of making additional zerglings as soon as he saw Keen's marine count. However, like his loss against MVP on Antiga Shipyard from the Winter Arena, he chose to press the allin instead of droning up behind it and couldn't do the required damage.

His early resignment mystified some viewers, but with only zerglings in production, Ret was lost. He had no economy and could not fend off the incoming banshees without immense sacrifice of what little he had. Meanwhile Keen's third base was already done and his initiative would let him float the CC to his third.


      Game Seven: Liquid`HerO (1) < Atlantis Spaceship > MVP.Keen (7)
On such a massive map, players are generally inclined to go for very greedy openings or openings that punish such greed. Hero went for Nexus first while Keen chose triple cc in response. However, he made a grave mistake in preventing Hero's rascally probe from scouting the third command center. Thus Hero felt completely safe taking a third nexus off of 1 gateway, giving him the initial advantage. Like last game, Keen put a twist on a standard scenario, going 4 rax no gas after the triple cc instead of the usual double gas. He then pressured early with his first ten marines, forcing Hero back to his natural. If Hero had been too greedy and neglected gateway production in favor of additional structures and probes, his marines could have done damage as they were so unexpected. Keen's 4 rax build gave him late tech and an army deficit but did let him feel out Hero's build without wasting money on scans that might reveal nothing.

Unfortunately, Keen misjudged Hero's build and most likely assumed he was going for greedy tech, instead of mass gateways. Quite possibly he expected Hero to have more units if he was going for the Parting 8 gate. Instead, Hero went for a greedy 7 gate without probe cutting. Keen moved out to take his third, a necessary step if Hero would simply be teching and obtaining upgrades. Instead Hero attacked with his delayed aggression and Keen rolled over and died.

      Analysis: Both players rolled the dice and Hero came out on top. This appears to be the theme of the night, with five matches being decided by blind coinflips. Keen let Hero see the triple cc and then guessed Hero's build incorrectly from his initial pressure and paid dearly.

However, one cannot fault Keen too much. Had Keen kept his third cc in his main and hugged the natural until medivacs and Hero simply went double forge, chronoed probes, and teched on 3 base - which is what Keen believed Hero was doing - Keen would have been so far behind it would scarcely have been a contest. This game reflected the main TvP problem on large maps: there is simply no safe way to open that won't get you massively behind versus one build or another. Both in this game and last, Keen and Hero showed the value of slightly altering builds against experienced players. Both players chose inferior strategies vs common builds, but they were very effective at beating the one plan their opponent blindly ended up doing.


      Game Eight: Liquid`HerO (12) < Metropolis > MVP.DongRaeGu (8)
Following the standard triple hatch and Nexus forge openings, Hero elected to take a quick third backed with immortals, while DRG in response grabbed a fourth and went roach/ling/infestor. Little happened for the first thirteen minutes, as DRG thwarted Hero's warp prism while the latter planned to turtle until he could assemble a mighty host.

As if the 14th minute marked a magic number, DRG rushed his entire Zergling force into Hero's third, trying to snipe the nexus. However, he sent far too many zerglings for the task and quickly faced a conundrum: he could conceivably snipe the nexus but he would lose far too many units in the process. Thus he chose to run his entire army out when the Nexus lost its shields, but was caught by forcefields and suffered severe casualties. Then he tried to dash into Hero's natural, but the Protoss simply warped in more units and pulled his army back.

A more ideal scenario for DRG would to have split his forces into two groups from the start, sending perhaps 40 zerglings into the third to snipe the nexus, while hoarding the rest in the fog to wait for Hero's army to move to the third, then running into the natural or main. Instead, DRG's simplistic army movement let Hero repel the attacks with no losses and many dead zerglings.

With DRG having sacrificed far too many units for no gain, Hero was able to attack move through DRG's depleted forces, attacking from only one angle in a somewhat narrow choke. DRG's spines were too few and too late, and a chain of fungals was not enough to hold Hero's vastly superior army.

Analysis: A highly untypical performance from the MVP ace. DRG made lazy attacks, lacked spines to cover the broodlord transition, and did nothing to complicate the game once he was behind. On the contrary, the solid Hero calmly handled DRG's misplaced aggression and killed him with a simple counterattack before the Zerg could rebuild.






      Game One: ST_Hack (7) < Antiga Shipyard > MaruPrime (1)
Both players went for cloaked banshees on Antiga, but while Maru ran early hellion into Hack's base and was prepared, at least psychologically, for Hack's strategy, Hack completely fell apart against Maru's banshee. The Startale terran got a raven out first, only to have lost all his marines. After he finally squeezed a viking out, the banshee had already racked up 20+ kills, and Maru was running away with it. To Hack's advantage was a much earlier natural expansion, but Maru kept the pressure on all game, and despite a series of engagements that left something to be desired, he made his cheaper and more mobile bio force work wonders against Hack's thinly stretched mech army. As efficient as mech is, in small numbers it's fairly useless, and Maru' bludgeoned Hack to death through sheer weight of numbers.

- One sided
- Watch if you like banshees (and only if)


      Game Two: ST_Life (7) < Cloud Kingdom > MaruPrime (1)
Life's debut on Startale did not go as planned, as Maru executed an 11/11 rax with pulled scvs, and Life simply was unprepared, tried to fight with drones, and just died.

- Not sure what Life could've done to win, but giving up the hatchery would've been a great start.


      Game Three: ST_Curious (11) < Ohana > MaruPrime (5)
Curious pulled Startale back into contention with a very well played game on Ohana. Both players went for macro-oriented styles, with Curious's fast third hatch being the greediest of the two. After building strength and fending off mutas, Maru went for a big marine-tank push, and Curious swept around from behind with a baneling carpet and muta flock, erasing everything and taking a solid lead. Ultralisks followed, and Maru actually did a very good job of holding off the zerg forces flooding in, but eventually Curious was able to get a few broodlords up as well, kill Maru's critical fourth base, and take the win.

- Watch this game if you ever doubt the importance of hitting your opponent from multiple angles, and angles they don't expect. Excellent management of ultralisks as well.


      Game Four: ST_Curious (7)< Entombed Valley > MarineKingPrime (1)
MarineKing was supposed to cut Curious's momentum down before it could begin, but instead, he built it up even higher. On Entombed Valley, he went for a marine-marauder-hellion attack to take advantage of the rocks splitting off the zerg third. Curious had excellent creep coverage, however, and caught MKP's small force with a handful of banelings, against which the king of splitting split horribly poorly. Curious then went for a huge baneling bust that MKP was woefully unprepared for which led to most of his scvs getting destroyed. He eventually cleaned up the first wave, only to properly die to the second.

- Totally brain dead play from MKP. He looks like he did when he was slumping through last year. No confidence in his play.


      Game Five: ST_Curious (7) < Sandshorn Mists > BBoongPrime (1)
Curious was on a roll, and felt confident enough to 10 pool against BBoongBBoong, which killed 10 drones, canceled B4's early hatch, and just about ended the game in every respect. B4 would stick it out, attempting to win the lottery in zergling-baneling wars, but Curious's control and attention to his units was superb, and he was able to hold on to his hefty lead, eventually overrunning B4 with a wonderful roach-baneling attack and excellent micro to once again, and for the last time in the game, preserve his lings from B4's banelings. Total domination, and a huge wasted game from Prime to lose to a build order.

- You knew Curious was on a roll, because of how confidently he played his out. He just knew exactly what his opponent would do, and picked the perfect build and execution to go with it.


      Game Six: ST_Curious (1) < Daybreak > ByunPrime (7)
An excellent back and forth game on Daybreak took away Startale's momentum for a moment. Byun played superbly, but the midgame belonged to Curious, who surprised Byun with muta tech, killing a ton of workers and a few addons and turrets. The primary damage was done with Curious's first few mutas however, and even though he made quite a bit more, they were significantly less effective as Byun successfully defended on three bases with marines, turrets, and a very well placed sensor tower. Byun eventually moved out, forcing Curious to give up his fifth, but the zerg has broodlords coming, and Byun lost his whole army in the field attempting to kill Curious's fourth. Curious looked well ahead, but pushed a little too aggressively to deny Byun's fourth, and the terran suddenly was able to eliminate the broodlord cloud. After that stunning reversal, it was all Byun.Curious attempted to switch to ultras, but Byun crushed those with superior numbers and excellent control. Even though Curious fended off Byun's attacks for a while, the terran had reached the stage in which it is both cost-effective and able to produce huge numbers of units. Curious was on the back foot the whole game after losing his broodlords, never able to counter attack well, or kill off enough of Byun's army. Eventually Curious tried to get back to broodlords, and Byun walked in and steamrolled him.

- Best game of the series by a long shot. Really good patient play from Curious to kill off Byun's initial army, and really good patience from Byun to bring it back.
- Felt cheap that Curious basically dominated the whole game, then lost it all in one moment, but even though Byun was getting his fourth up, Curious could have stood to wait an extra minute or two to be sure. Needed more infestors especially, and he should've been prepared for the vikings.


      Game Seven: ST_Bomber (1) < Atlantis Spaceship > ByunPrime (7)
Byun went for an early proxy marauder rush, but Bomber's early scout, despite missing the barracks, was early enough to spot the absence of one in the main, and he was well prepared. Byun, evidently expecting a FE from Bomber, simply pulled back when he didn't scout one, and proceeded rushing up to banshees. But Bomber moved out with a handful of marines and simply walked right into Byun's base, where he was able to snipe cloak before it finished and a host of SCVs. The game went downhill from there for Byun, who had to contend with being behind in absolutely every facet of the game. Eventually, after weathering several hail mary attempts and barely failing to kill Byun on several occasions, Bomber moved out with his full army and (with manner mule, of course) simply crushed Byun's vastly inferior forces.

- What on earth was Byun doing when he left his front door open to Bomber's marines. Just an inexcusable mistake, he was being just a bit too greedy, and completely lost the game because of it.


      Game Eight: ST_Bomber (11) < Tal'Darim Altar > CreatorPrime (1)
With their back against the wall, Prime sent out the first protoss of the evening; their PvT ace Creator. The protoss started off with his very typical robo and double forge, then transitioned to his very predictable storm and chargelot play. Bomber was fairly obviously working to counter Creator's style from the beginning though, getting out a massive amount of units and worrying about tech later. Creator was hit right as storm finished, but he had been way too greedy, and had few templar or units in general to deal with Bomber's marine-heavy chargelot shredding machine. Perhaps great micro could've saved him, but Creator fell way too easily for an obvious bait, chasing half of Bomber's army while the other half waltzed in and killed his third with no resistance. Picking up the army and dropping it into the main, Bomber killed both forges and saved his drop as Creator's multi-task completely and utterly fell apart. From then on, there was little that Creator could do, and in typical Bomber fashion, he macro'd never ending wave of bio to win the game and take the series for Startale.

- Creator really had no multi-task whatsoever this game, it was dispiriting to see. He had less stuff and it was in the wrong places always. Bomber tore him apart, and Creator's vaunted double forge build looked Jurassic compared to the way Bomber played him. Second best game of the series.



It was a great performance from Startale, and a show of strength as well to not win with any of their vaunted protoss players. Curious's play was the turning point, as taking out MKP with no damage done eliminated Startale's #1 boogeyman. Of course, defeating Maru and BBoong was pretty important too. With those crucial wins, it was a great, low pressure situation to send Bomber into, and he delivered, taking out Prime's second and third scariest players to finish the set. With this huge confidence boosting win, Startale moves into the first losers finals match, where they'll be heavy favorites against Liquid.

First Losers Bracket Finals Preview

By NrGmonk

It has come to this. Liquid miraculously beat MVP last week and is now fighting in the Losers Bracket Finals for a chance to face up against IM, for a chance to make it to the LAN finals, for a chance to win it all! This marks Liquid's second win over a GSTL championship team, proof that Liquid has what it takes to beat any team on the planet. However, their next challenge will be arguably their biggest yet, facing Startale, a team that even toppled the Primes. Against Startale, there will be unique challenges that have not faced versus any other team they've played, not versus MVP, not versus Prime, not even versus SlayerS.

First of all, Squirtle and Parting will be especially daunting obstacles for Liquid; in contrast to Prime, whose strength is their Terrans, and MVP, whose strength is their Zergs, Startale's strength is in its Protoss, in particular two of the absolute best Protoss in the world. Against Taeja, both Squirtle's solid Colossi-based play and Parting's Storm-based play rank among the top five PvT in the world . And Hero dreads the PvP match-up almost as much as Oz these days, being by far his weakest match-up and the most common reason for his exits out of tournaments. In contrast, Startale, being a team heavily dominated by Protoss including Naniwa and Sase for pratice, contains the absolute best PvP players in the world. Squirtle and Parting hold 73% and 58% win rates in this match-up respectively, in stark contrast with Hero's 38%. Even without going into specifics, numbers alone favor Startale in this match.

And despite how good Liquid members are, Startale players just seem to have their number in tournaments. In last season's GSL, TaeJa and HerO were eliminated 0-3 by Squirtle. At the Red Bull Battlegrounds LAN, Ret was beaten by both Squirtle and Parting. And at the same LAN, TaeJa suffered defeats to both Bomber and SaSe, a player who may not be eligible to play today, but still resides in the Startale house and is a good indication of StartaleQ Protoss strength. In terms of team leagues, ZenexStartaleLife all-killed Liquid in this season of GSTL. And even Curious has had his shots, knocking Hero out of Code S this season of GSL. In recent memory, Liquid players have not managed to topple any of Startale's players, a fact I'm certain they will have in the back of their heads when going into this match. (Ok, Taeja went 3-2 vs Ace in GSL. Big deal.) With this mental handicap, Liquid has a disadvantage even before the first games are played.

But even with all these other hurdles, I strongly feel that Liquid's biggest difficulty to overcome is that they seem more and more like a two man team these days. Hero and TaeJa are obviously among the absolute best players in the world, but beyond that Liquid leaves a lot to be desired against top Korean teams. Ret, Zenio, and even Sheth can beat one or two top Koreans in team leagues, but you never feel that they can go just a bit further than that. Both the rest of the burden, which is the vast majority of the burden, rest on just the shoulders of two players. The lack of strong supporting players can be seen most evidently in Liquid's run of this season of GSTL. When both Hero and TaeJa failed to clean up their respective matches, Liquid's remaining players provided very little backup, with only Zenio being the only other Liquid member scoring a win over two matches. We've seen team leagues being dominated by just one player (DRG), but more consistent team league performances can be seen from well balanced teams full of strong players; this is your Primes, your SlayerS, your Startales.

Team Liquid vs Korean teams are always hard to predict. So far in IPL TAC, I've gone 0-3 in predicting this match-up. But more than ever, unfortunately, I am sure of Liquid's defeat. Startale just seems to have the perfect combination of players to counter Liquid. But miracles have happened. Who's to say the magic will run out just now?

Prediction: Startale 5:3 Liquid