LEC 2021 Summer Playoffs Prediction

LEC 2021 Summer Playoffs Prediction

With the Summer Split of the LEC ending last Sunday, it's time to look back at the season, see how reality lined up with our predictions, and wildly speculate (i.e., run some models/simulations) on the playoff tournament that starts in a week and a half!


Pre-Season Predictions

At the start of the split, we had generated a power ranking for the competing LEC teams, complete with our expectations of their win/loss records. We had a major caveat with those expectations: they were made given the information we had at the time, before a single game had been played in this season. As the season unfolded, new data was incorporated into the model, and we were able to generate updated predictions for each match (which is what our suggested bets are based on - always using as up-to-date data as possible). So, let's take a look at our initial pre-season predictions, our later predictions, and estimations extrapolated from market odds, and see how they stacked up with reality:

Team Name Acronym # Matches Wins Predicted (Initial) Wins Predicted (Final) Wins Predicted (Market) Actual Wins
G2 Esports G2 18 11.87 11.68 11.51 12
MAD Lions MAD 18 11.32 11.18 11.52 12
Rogue RGE 18 11.27 11.34 11.89 13
Fnatic FNC 18 9.38 10.64 10.26 11
FC Schalke 04 Esports S04 18 9.12 7.57 7.58 3
Team Vitality VIT 18 8.95 8.06 9.02 8
Misfits Gaming MSF 18 8.32 9.44 9.66 12
SK Gaming SK 18 7.22 6.05 5.8 5
Astralis AST 18 6.67 6.77 6.38 7
Excel Esports XL 18 5.87 7.28 6.38 7

We had stated that the season was anyone's to win among the top three (G2, MAD, RGE) - this turned out to be true, with Rogue taking the highest record and the others just a single win behind them. There was, however, an unexpected contender for this top echelon: Misfits Gaming. They handily outperformed our initial estimate, with a three-way-tie-for-2nd-place 12 wins! As they continued to beat expectations, our model adjusted, as evident by our higher final win predictions, but their performance still managed to surprise both us and the market.

On the flip side, FC Schalke 04 massively flopped, with just 3 wins across their 18 matches. Our initial estimate was again adjusted by our model as their performance dwindled, but not even their biggest rivals could genuinely have claimed to expect this.

A quick aside on the market's predictions: from the odds offered on the market right before each match starts, we can infer an implied win probability for each team for a given match. From then, as with our own calculations, the total season wins are just the sum of their individual match win probabilities. These odds are only offered fairly close to when matches occur, so in terms of data access they are contemporaneous with our "final" predictions and thus are fair to compare against those. No oddsmaker offers odds for every match at the start of the season, or even just odds on total wins, so there's no good point of comparison for our initial pre-season estimates.

For this split in particular, we suggested bets on only 13 of 90 matches, so we were generally very much aligned with the market. Hence, our final win estimates are very similar to the market-implied win estimates. We would argue, however, that ours are actually closer to reality: our total deviation across all teams is 11.9 wins, compared to the market's 12.8.


Summer Playoff Bracket

Next step, playoffs!

The top 3 teams of this best-of-5, double-elimination bracket qualify to Worlds, so the stakes are turned up. Among the 10 LEC teams, the top 4 (Rogue, MAD Lions, G2, Misfits) meet each other in the winners' bracket, while the 5th and 6th place holders (Fnatic and Team Vitality) start in the losers' bracket:

source: official LEC twitter (@LEC)

This is a particularly brutal bracket for the two losers' bracket starters: since they face each other in round 1, rather than simply facing losers of the upper bracket, they each have an incredibly long and arduous journey to the finals. In fact, to actually take 1st place in the tournament, they would have to consecutively beat literally every other team in the competition. That's about as uphill as a battle can get. Nothing is impossible, and you know we (and our model) love rooting for underdogs, but things are looking pretty stacked against Fnatic and Vitality.

Not sure what else you could have possibly expected of us, but just in case this is a surprise: yep, we used our model to simulate the outcomes of this tournament 10000 times (this is called Monte Carlo Simulation, if you're curious).

Each outcome is path-dependent, unlike the regular season (in the regular season, each team plays each other team twice, regardless of how any matches turn out. In a tournament, obviously, winning or losing determines what happens next). There's no straight-forward way to show every likely possibility, so instead we aggregate these 10000 simulations in a few ways:

First, the probability that each team wins the tournament overall, and their median placement:

Team Name % 1st Place Median Place
G2 Esports 34.62% 2nd
Rogue 30.78% 2nd
MAD Lions 27.55% 2nd
Misfits Gaming 5.29% 4th
Fnatic 1.70% 5th
Team Vitality 0.06% 6th

G2 Esports is the most likely to win the tournament outright. Rogue and MAD Lions are pretty hot on their tail though, each with a reasonable shot at taking the crown. All three of those teams are fairly likely to take at least 2nd place as well (obviously, they can't all do that, but we would be surprised if the top two aren't from among those three). Misfits, in spite of a pretty strong summer record, has an incredibly difficult first matchup against G2 which hampers its overall chances. Fnatic and Vitality, as mentioned previously, are going to have quite the ordeal ahead of them.

Next, the probability that each team qualifies to Worlds, by placing anywhere in the top three spots:

Team Name % Top 3
Rogue 89.89%
G2 Esports 81.86%
MAD Lions 81.71%
Misfits Gaming 33.51%
Fnatic 11.63%
Team Vitality 1.40%

Even though G2 Esports is more likely to actually win the whole tournament, Rogue is more likely to top 3! Due to the nuances of some particular matchups, G2 basically has a more spread out distribution of outcomes. Part of this is the fact that Rogue's #1 seed guarantees at least a 4th place finish, which no other team has. Rogue has an almost 90% chance to secure a spot to Worlds! 90% is not exactly a guarantee, but would still be a major upset. Fnatic has somewhat of a shot as well, though Vitality would need almost a miracle. To get 3rd place, either of them would have to win their first three matches, which is certainly doable. In fact, the probability that one of these two underdogs does this is 13.03%! That's right, there's almost a 1 in 7 chance that we see a pretty big upset! If we include Misfits as an underdog here, the odds that the top three aren't just Rogue/G2/MAD Lions is 44.2%!

And finally, the distribution of each team's likely outcomes:

Team Vitality is by far the most likely to take 6th place, while Fnatic is favored (almost 80/20) to beat them but is likely to lose after that and secure 5th. No other team can even place in 6th by definition, and Rogue can't even place 5th due to their #1 seed. MAD Lions is fairly equally likely to place anywhere in the top three.

We can also simulate a bracket where the favorite (as predicted by our model) wins every match, giving us these final standings:

Team Name Place
Rogue 1st
MAD Lions 2nd
G2 Esports 3rd
Misfits Gaming 4th
Fnatic 5th
Team Vitality 6th

This is technically the single most likely outcome to occur (i.e. single exact bracket), but it is still quite unlikely given just how many possible brackets can occur: with our win probabilities, the chance of this exact bracket happening is 1.13%. It's still a fun exercise, though.

(Side note: there are 8 matches, each of which has 2 possible outcomes which determine the overall bracket, for a total of \(2^8=256\) possible brackets. This means that any particular bracket has, on average, a \(1/256=0.39\%\) chance of occurring, so the most probable bracket is almost 3 times more likely to occur than the average bracket. Not that this means much analytically, just interesting to point out.)


How Do I Use This Info?

There isn't a direct way to bet on which team will win this tournament overall, or each team's place, etc. As usual, however, you should have opportunities to bet on each individual match, as the tournament progresses and the two teams in each match become known. We'll keep up-to-date on this with our suggestions, so make sure you're checking back before each round!

Hopefully, this analysis still gave you a sneak peek of what to expect in the LEC going forward. Besides, speculating on tournaments for fun is sort of a national pastime (of, let's be honest, pretty much any nation).


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!