[WL] In Search of an All Kill

At long last, folks, Winner's League is upon us. We've waited a good long time for this, and after making it through two rounds of normal proleague, we've finally arrived at the league that everyone has been eagerly anticipating and looking forward to. In case you're wondering what all the hullabaloo is about, we've taken the effort to summarise the rules of Winner's League for you below.

Winner's League employs a BO7 format. The first team to get 4 wins, wins the set.
Before each match, only the map order, and the players of the first game are known.
For each game, the winner stays, and the loser is eliminated. If player A wins the first game, he stays for the second game, and keeps playing until he loses.
The next player to be sent out is decided, on the spot, by coaches.
There is no restriction on the race of the players sent out. In normal proleague, each team must field at least 1 of each race.
Winner's League has its own playoffs and associated prize money.

Special Features

Game of The Week

Statisfaction

There is a reason for the hype behind Winner's League. It just works in a completely different way from normal proleague. What also adds to its success is the fact that a majority of what is going to happen is left unknown until the last minute, and the suspense and tension, both immensely contribute to the overall atmosphere - you never know what strategies you're going to face, what maps you'll be playing on, or even who and what race you're going to be playing against.

Do you see, now, the potential of such a format? Players are forced to practice not just on 1 map, and for 1 matchup, but on every single map, and against all 3 races. This brings a whole new element into the game; gone are the times of 1 map or 1 matchup wonders. Strategies employed by coaches have to reflect this, and indeed, coach decisions on who to play next are not simply going to be based on map statistics anymore. They have to look at other factors such as the opponent's win ratio against a certain race (think Casy vs P), the maps coming up, and the potential players who might be fielded.

A summary of the coach's thoughts

The "mystery" aspect aside, what truly makes Winner's League stand out from normal proleague is the possibility of an all kill scenario happening. The fact that one player can possibly, singlehandedly take down an entire team, changes the entire dynamic of the game. Teams like Hwaseung OZ that usually rely on one player to bring in the bread and butter, suddenly become powerhouses in this format. For the players, nothing can feel better than ruining an entire team and coming back from 0-3 down to win the set. For the same reason, this format is also endearing to the fans - to us, watching a player all kill a lineup like Flash, Luxury, Violet, ForGG, is almost as good as watching him win an individual league.

However, all killing is not an easy task. Winning, and winning multiple games at that, is a true test of skill, endurance and fortitude. Heyoka crunches some numbers from last season's Winner's League and shows us how fatigue might play a role in player performance.

Statisfaction
By Heyoka

Winner's League offers a unique opportunity to look at new facets of competitive Brood War. Players having multiple games, coaches making a series of on-the-spot decisions, and unlikely stars rising to fame, all make for an experience very different from the normal proleague.

It raises all kinds of questions. The biggest feature of Winner's League, the defining feature, is the ability to play as many games in a row as it takes for you to lose (or your team to win). Originally, I wanted to look at how fatigue played a role in a player's game, but I ended up discovering something quite different.

To begin, I decided to look at the list of games played during last year's Winner's League section. In total there were 415 games played during the season (including the playoffs at the end). They are broken down by which game the 'leading' player was on - essentially how many games he had won. Since you're out as soon as you lose one, one player will always be playing his first game of the set and the other (with the exception of game 1) will be, at least, on his second.

There were 69 instances in which both players were on their first game (because the only time this can happpen is the first game of a match - there were 66 regular matches plus 3 in the playoffs).

There were 210 games where one player was playing his second (note this doesn't mean he won two, only that he had two games in a row). There were 97 in which a player was playing his third game, and 39 instances where a player found himself playing for an allkill (going for the 4th win). Here is that information in a chart.

As you can see, as the number of games increases, the number of players who have achieved this number goes down. Intuitively, that makes sense. It's tougher to win multiple consecutive games (plus the strongest player of a team typically comes out later).

As I said before, this had been done to see if fatigue played into performance. The plan was to see if players performed worse on their second, third, or fourth game. To do so, I used the same categories and got the winrates for each of these situations, ignoring the individual players completely. This is what they look like.

All are played against someone who is on their first game

When both players are on their first game, half of them win and half of them lose (mind blowing). This is looking at the population as a whole, which will (theoretically) balance out the difference seen in skill between individuals so the numbers for the other categories should also be near that.

From there, the results differ, somewhat, from the expected. Players on their second game, as a whole, actually win more than 50% of the time. The effect is amplified when a player is on their third game, jumping up to nearly 55%. Yet, when looking at potential-all kill scenarios the winrate is far, far lower. Throughout the season, there were 39 times a player had a chance at capturing the elusive all-kill and being a hero for the day. Out of all those times, only 17 times was that player able to manage it - a rate of 43.5% that is a far cry from any other category.

Why are players so much worse in their fourth game than second or third? Is it fatigue? Likely a relevant factor, however, the data seems to be hinting at something else: the coach strategy.

Watching the games, I always get the feeling that coaches save their best for last. It could be for morale, it could be for the express purpose of sniping a player after he has played a few games. I can't say for sure, but just this year, even Snow has seen play as CJ's third player. This data certainly seems to suggest that this phenomenon is indeed happening - weaker players are going out for games 2 and 3 and when all is on the line the team finally steps up with their A-game.

Assuming it is actually happening, how valid of a strategy is it? On one hand, it seems incredibly risky to put that much faith in a single player. On the other hand, the winrate in the 4th straight game is considerably lower than the average of 50% so there is a possibility that it works. Exactly how optimal it is in terms of strategy, though, will have to wait for another day.

So, after that insightful look into the statistics of the past Winner's League, we move on to some of the goings on of the week.

One thing we noticed was that every week you see names like Jaedong, Flash, Sea, Effort, brought up, again and again. These players either crush their opponents mercilessly, and deservedly, are mentioned in these updates, or they lose spectacularly, and are crucified in the newspost. Amidst these superstars, other names have been forgotten, lost in the limelight of teammates, or opponents, buried under the deluge of "omg Bisu slumping" or "Flash BONJWA!"

You might remember our feature, in R2, of the teams that occupied the middle of the table. Well, today, we talk about players that have been similarly neglected. In particular, we want to mention a player that achieved the first all kill of the season: Guemchi. A long time ago, Guemchi, along with Tazza and Horang2, was thought to be the next big Protoss. Unfortunately, he seldom saw play, often benched because of the superior skills of his teammate, Free.

How many games did you win tonight?

This week though, he was sent out in Woongjin's second game against CJ, after the defeat of Soulkey, at the hands of Iris. He went on the wow everyone by ploughing through CJ's entire lineup, which included Effort and Movie. In fact, his game vs Movie on Fighting Spirit was a pretty exciting PvP that is definitely worth a watch if you enjoy the match up. What was an even larger surprise was him successfully overcoming CJ's ace, Effort, on Moon Glaive, which almost no Protoss player has ever been sent out on.

This accomplishment puts him solidly in first place of WL's player rankings, his record at 6-0 from another 2 wins against Air Force ACE. Owners of him in FPL must be pretty happy, for he was a huge earner this week, netting in a full 16 points, for a cheap cheap cost of only $2. Cries of "Guemchi Bonjwa!" have already been heard around the forums, and although we think it might yet be too early to conclude, the future does seem rather bright for this young Protoss player.

Another player that seems to have also been neglected thus far, is Hiya. Why we've never mentioned him in any of our updates, we are unable to fathom either, but here he is with this week's Game of the Week.

GOTW - HiyA vs HoeJJa on Roadrunner

By Pangshai

Some of you might disagree, but Hiya is largely in the same category as Guemchi. He is generally a pretty good player, and has had some notable achievements, such as taking out Stork and Jangbi, on the same day, in the set against Khan last round. Possibly overshadowed by his teammate and best Zerg alive, Jaedong, Hiya has bore the burden of surviving Hwaseung's Terran lineage silently. We tried to figure out why we never ever spoke better of him, and we couldn't really come to a conclusion, but perhaps, it was because of games like this.

The match opened on Roadrunner, and Hiya landed his CC, in a cheery teal colour, at the 2 o'clock position. Things got feisty quickly when Hiya moved out with an early SCV, proxying a 7 (or 8?) rax in the centre of the map. In the meantime, Hoejja, brown Zerg in the bottom right, was oblivious to the incoming bunker rush, happy to get things going with a standard 12 hatch at his nat.

Luckily for him though, his overlord had gone the right way, and soon discovered the absence of a barracks in Hiya's main. A drone scout then confirmed the presence of a proxy rax, and the Zerg player immediately pulled drones to stop the Terran aggression. Unfortunately for him, Hiya had pulled all the stops in his attack, sending a team of SCVs down to support his couple of marines.

Hoejja, for his part, was no Jaedong either. His drones failed to stop rallied rines from heading across the double bridges to his nat, and ended up stuck in the middle of the map, prevented from resuming their mining duties by the now completed bunker at his nat. It seemed the game was surely lost, and Hiya was going to progress to his third win of the day.

The hatchery fell, inevitably, and Hiya even managed to block Hoejja's attempt to run by with his lings. The previously mentioned drones, however, had found their way to the Terran main. There, they met the remainder of Hiya's SCVs, and the sight of only 5 SCVs must have brought a smile to Hoejja's face. Outnumbering them with 7 drones, Hoejja forced the SCVs to temporarily cease mining in favour of chasing his drones around.

Concluding that his work at Hoejja's nat was done, Hiya retreated his SCVs there to deal with the situation in his main. This provided an opening for Hoejja's lings to exit his main, and they ran past the bunker to join a battle at Hiya's ramp. Lings and drones battled SCVs and a lone marine, with the SCVs holding the ramp while the marine fired from above. The Zerg forces were driven off eventually, but the SCV losses were not insignificant.

To add to the Terran economic losses, during this confrontation, Hiya had had only 1 SCV mining. It was unsurprising that he decided to send the SCVs holding his ramp back to mine, and Hoejja, ever the opportunitist, once again, hopped onto the opening, sending his forces up the ramp and taking out the solo marine. More SCV losses followed, but Hiya's defense was soon to be bolstered by 4 returning marines that had abandoned their bunker.

Upon their return, they put an end to Hoejja's remaining 2 drones and 1 ling, and returned to the ramp to defend. Unfortunately for Hiya, he made the exact same mistake that he had previously - not leaving an SCV on the ramp for defense. Hoejja's lings ran up, clawed their way through the marines, and Hiya saw his victory just go up in smoke.

The most fitting description of this game would be, hilarious. It was a comedy of errors from both players, with Hoejja eventually coming up on top because he knew how to make zerglings, and he timed his attacks well. It's sad that the first time we mention Hiya in an update, we feature him in one of his worst games, but heh, chum, you've only got yourself to blame.

And that game is just another example of why we love WL. Unprepared for a certain matchup on a certain map? Don't worry, just cheese! When players face an unexpected opponent on an unfamiliar map, cheese seems to be the strategy of choice, and it spices things up a whole lot. The number of mirrors played is also greatly reduced because of the way WL works, and this is something to rejoice about as fans. Just look at Moon Glaive. Ever catch one of the 6 PvZ games on it? In the first week of WL, we've already seen 2 PvZs on it, and no doubt, we will continue to see more non TvTs on this map.

More interesting news this week, Flash met Jaedong again in the ace match of KT Rolster vs Oz, and found revenge for his loss in the MSL finals. Jaedong opened with his 3 hatch before pool build (ZvT revolution, anyone?), and Flash as usual, fast expanded. Jaedong then got some early upgrades going, and Flash looked like he was behind after his first push got taken out by mutaling, and his tech and upgrades were late to get started. Jaedong even kept Flash in his main for ages with muta harass, but his 3rd gas, had been delayed by Flash's first marine medic group. What changed the whole game was Flash's push. Jaedong's hive tech had probably been done for a while already, but for some reason, his defilers weren't out, and when they did eventually spawn, they didn't have consume. No one knows if he forgot to research consume, or if he simply got his defiler mound late, but needless to say, Flash cleaned up and won the game.

This game pointed out the fragility of an advantage. Not going to dwell on it for too long [note: and please refrain from bringing comments on the MSL finals into this thread], as it's been discussed to death, but If the lights had gone out a couple of minutes before the end of the game, a regame would've been the most likely scenario. A simple mistake like Jaedong's that caused the game to be reversed, really demonstrated how easy it is for an advantage to be lost.

In case you haven't figured out, we love numbers

In other news, FBH chose an opportune moment to start playing well again, when he faced off against SKT's star studded line up and nearly took them all down. His game versus Protoss was tested, not once, but twice, against two of the very best Protosses in the business. FBH overcame, and in the last set, he almost achieved an all kill of SKT with 3 proxied factories against Fantasy. Unfortunately for him, he was probably too focused on the ceremony he was going to perform after winning, pissing his advantage away, and letting the game slip between his fingers. No huge loss though, Stork cleaned up two games later with carriers on Fighting Spirit and brought the win home for KHAN.

Thats all folks, that is all we have for the premier week of round 3. This is merely a taste of things to come, of kills to kill. Be sure to join us next time with another exciting recap of the games and happenings!

This Proleague update was brought to you by Team Liquid's PL 2009-10 coverage team - riptide, heyoka and pangshai. Thanks to keit as usual for the stellar graphics!

[WL] In Search of an All Kill

[WL] In Search of an All Kill

Completed

Ongoing

Upcoming