LCK 2021 Summer Playoffs Prediction

LCK 2021 Summer Playoffs Prediction

Last of the four major regional leagues we're going to cover this summer, the LCK playoffs have begun! You know the drill by now: season performance, pre-season predictions, and finally a simulation of the playoffs. Worlds qualification is on the menu, so let's dive right in!


Pre-Season Predictions

Here's our pre-season power ranking for the LCK. Broken record time: these predictions were all done before the season with the information we had at the time; throughout the season the model updated itself with information from new matches and was able to give us more up-to-date predictions (as well as us being able to keep track of roster changes), which feeds into our suggested bets.

The outcome?

Team Name Acronym # Matches Wins Predicted (Initial) Wins Predicted (Final) Wins Predicted (Market) Actual Wins
DWG KIA DK 18 15.39 14.88 13.37 12
Gen.G GEN 18 14 13.19 12.08 12
T1 T1 18 12.36 11.05 11.48 11
Hanwha Life Esports HLE 18 10.76 8.59 8.62 7
Nongshim Red Force NS 18 8.18 8.68 9.5 12
Liiv SANDBOX LSB 18 7.42 8.35 7.99 11
Afreeca Freecs AF 18 7.37 8.24 7.88 11
KT Rolster KT 18 5.59 7 7.55 7
DRX DRX 18 5.56 4.93 5.49 2
Fredit BRION BRO 18 3.37 5.1 6.04 5

For the most part, teams were much closer to each other in performance than we had expected. Nobody broke 12 wins, even though the top two teams (DWG KIA and Gen.G) both had a strong shot at it. Nongshim Red Force, Liiv SANDBOX, and Afreeca Freecs all overperformed, joining T1 and the top two teams in the playoff bracket.

Hanwha Life Esports, on the other hand, failed to stay up to par and took home only 7 wins. Their strong spring performance (12-6 during the season, and 3rd place in the playoffs) had set them up to expect a similar run in the summer, but no such luck. However, that strong performance also left them with 50 Championship Points, which is enough to guarantee them a spot in the Regional Finals even without participating in the summer playoffs!

Finally, DRX also had a weak showing; we had predicted them to be near the bottom of the lineup, but 2-16 is a particularly painful record.


Bets

Across 90 summer split matches, we found 32 of them worth betting on - slightly over one in three. This is fairly close to our LPL rate (34%), and ahead of LCS/LEC (each around 25%). We're still largely in agreement with the market (if we disagreed on every game, it would almost certainly mean that our model is fundamentally flawed; the market is usually pretty reasonable), but have enough room to make some money.

Overall, we would be up $373.50 on $9053.36 of total wagers (not all at the same time, though):

A series of unfortunate and outsized losses kept our earnings down: in particular, Afreeca Freecs had huge upsets against DWG KIA and Gen.G (the top two teams!!) while we had bet on the favorite, but also managed to drop a match to DRX (who only won two matches all season) when we put money on them not to, so let's just say that Afreeca Freecs is not our lucky charm. This is, of course, the nature of variance, and in the long run we expect such instances of luck to average out and not sting so much.

Just like the LPL summer season, the LCK was rife with upsets. We've talked before about how simply always betting on the underdog or favorite should be equally bad, and equally unprofitable (due to the vig). Yet here, betting on the underdog in every single match would have made you money:

This is largely propped up by the absolutely massive $1016 payout you could have gotten betting $100 on KT Rolster against DWG KIA in week 1. That match was paying out 11.16x, which implies about an 8.4% chance of winning. Across so many matches, a huge upset to the tune of 8.4% had a good shot of happening at least once, but it's still fortunate that it happened on such a massive disparity and payout.


Playoff Bracket

By whatever fortuitous (or expected) path that got them there, six teams now find themselves in the playoffs, vying for Worlds qualification!

The bracket is a single-elimination, best-of-5, 5-matches-in-total affair. In the second round, DWG KIA gets to choose which round 1 winner it will face, with Gen.G facing the other round 1 winner:

source: official LCK twitter (@lck_global)

In rounds 1 and 2, the relative placement of the losers is determined by their seed going in; this is important because those placements determine Championship points, and placing 3rd vs 4th could be the difference between qualifying to Worlds or not.

DWG KIA is in a particularly good spot here; its 1st round bye as well as #1 seed going in guarantee it at least 3rd place, which paired with its strong spring season performance should almost certainly give it enough Championship points to qualify for Worlds (we'll talk about the one scenario where they don't in a bit).

Gen.G is almost in as good a spot - the only way they take 4th place is if both they and DWG KIA lose their first matches. Otherwise, GEN either takes at least 2nd by winning, or 3rd by being a higher seed than any round 1 loser other than DK.

Simulating the Bracket

Running through this bracket 10000 times (Monte Carlo Simulation) yields the following probabilities for each team to win the whole thing:

Team Name % 1st Place Median Place
DWG KIA 60.02% 1st
Gen.G 33.62% 2nd
T1 2.45% 4th
Afreeca Freecs 1.74% 6th
Liiv SANDBOX 1.14% 5th
Nongshim Red Force 1.03% 4th

(One note here: because DK gets to choose which round 1 winner they will face in round 2, and there's no way to know their subjective preferences, we have assumed that they will always choose to face the lower-seed team.)

DWG KIA is a pretty big favorite to just take the #1 spot. They're bolstered by the fact that their biggest competitor, Gen.G, is still fairly unfavored in their potential and quite likely finals matchup: DK is favored 62/38 over GEN in our model.

So, naturally, Gen.G is in a similar spot - they're very likely to make it to the finals, but have a tough battle there.

Everyone else has a very low probability of winning, though of course anything can happen.

Point me in the right direction

However, this bracket isn't all-or-nothing. The winner gets a direct qualification to Worlds, but everyone else gets Championship points, which stack with their points from the spring season. The team with the highest point total also qualifies.

Here is each team's point total going in:

team_name points_prior
DWG KIA 90
Gen.G 70
Hanwha Life Esports 50
T1 30
DRX 20
Nongshim Red Force 10
KT Rolster 0
Afreeca Freecs 0
Liiv SANDBOX 0
Fredit BRION 0

We've included teams not in this bracket; in particular, Hanwha Life Esports had done well enough in the spring that while they can't get a direct qualification to Worlds here, they're guaranteed a spot in the Regional Finals which will determine Korea's 3rd and 4th Worlds slots.

Points are awarded based on summer playoff standings as such (AQ = Automatic Qualification):

Place Points Earned
1st AQ
2nd 100
3rd 80
4th 50
5th 30
6th 10

Here are the odds that each team qualifies by either path:

Team Name % Qualify
DWG KIA 97.17%
Gen.G 96.47%
T1 2.45%
Afreeca Freecs 1.74%
Liiv SANDBOX 1.14%
Nongshim Red Force 1.03%

DWG KIA and Gen.G are all but guaranteed to go to worlds.

DWG KIA is guaranteed at least 3rd place, due to their first round bye and #1 seed. Realistically, the only way they don't qualify is if they lose immediately, take 3rd, and then Gen.G wins their first match but loses in the finals. Then, both DK and GEN will have 170 points, with the tiebreaker going to GEN for having more points in the summer split. Apparently, this will occur roughly 2.83% of the time.

Gen.G's got a similar setup: to not qualify, they would have to lose in the first round while DWG KIA wins their first match but loses in the finals. Then, DK would have the most points by a significant margin, while neither team directly qualifies.

You'll notice that for the bottom four teams, their probability of qualifying to Worlds is exactly equal to their probability of winning the playoffs! This is not, in fact, an error: they're so far behind in points going in, compared to DK and GEN, that they can't possibly qualify via points; their only path forward is to get the direct qualification by winning outright. The odds of this happening for any one of them is 6.36%, which is around 1 in 16. Or, conversely, the odds of the two possible qualifications just going to the top two teams is 95.64%!

Of course, placing well here would still allow the underdogs to get to Worlds via the Regional Finals, so while it may seem that this bracket is almost a formality, actual placement does matter. We'll cover our predictions for those once the playoffs are over and the contestants are set.

Finally, here's the distribution of each team's outcomes:

Because of the quirk of how 3rd vs 4th and 5th vs 6th places are determined, it's actually impossible for Afreeca Freecs to get either 3rd or 5th place. Nongshim Red Force can't place 6th, and places 4th only if they win against Afreeca Freecs, lose to Gen.G, but then DWG KIA loses their first match - so not very likely.

Favorite Flavors

If every match was simply won by the favorite, we'd end up with these standings:

Team Name Place Points
DWG KIA 1st AQ
Gen.G 2nd 170
Nongshim Red Force 3rd 90
T1 4th 80
Liiv SANDBOX 5th 30
Afreeca Freecs 6th 10

Surprise, this is just their seed going in! Not really much room for disparity. Worth pointing out, though: if LSB and AF faced each other in a lower bracket or whatever to determine which of them takes 5th vs 6th place, it would be almost a coinflip, with a slight edge determined by which team gets to start on blue (and normalizing for that, AF would have a slight edge).

This is actually a remarkably likely scenario to end up in: this happens 14.04% of the time in our simulations! So, 1 in 7 times, this playoff series plays out exactly as expected. This is partially because this bracket only has 5 matches in it, which gives 25 = 32 possible brackets, so the average bracket has a 1/32 = 3.125% chance of occurring. Given that, it's not unreasonable that the most likely outcome, in a tournament full of fairly lopsided matches, is almost 5 times as likely as the average.


The B Word

And again, betting on the overall outcome doesn't really have a market. Each individual match will have its own odds though, so make sure you keep up with our suggested bets to stay on top of things!

We hope you still found this analysis interesting as an illustration of how this tournament could play out. And even though this is one of the more lopsided brackets, with an incredibly high likelihood of the standings just being equal to the teams' starting seeds, "very likely" isn't actually synonymous with "always going to happen", so we'll just have to wait and see!


For more in-depth match analysis, discussion, and predictions, as well as quantitative betting strategy tips, subscribe to stay in the loop! You can also follow us on Twitter or Facebook for quick prediction snapshots of upcoming matches. Alacrity is your fastest path from simply gambling to beating the market. Stay ahead of the game!