LoL Worlds 2021 - Finals Prediction and Bet Suggestions

LoL Worlds 2021 - Finals Prediction and Bet Suggestions

It's been an exciting month as teams from around the world clashed in the League of Legends 2021 World Championship, and now we're down to the final two teams. South Korean incumbent champions DWG KIA face off against the Chinese Edward Gaming.

Everyone seems to be expecting a dominant DK showing, but EDG have proven themselves to be quite a contender as well. Let's see where the market and my predictions shake out.

In addition to just talking about the odds on the overall winner, I'll go over some other available bets to see if any of them are worth a wager.

Overall winner

Let's start with the main attraction: the new World Champion! At the end of the day, this is the only one that truly matters, and is what both teams will strive for. It doesn't matter if you win 3-0, 3-1, or 3-2, the Summoner's Cup is still yours.

Alacrity's LoL Worlds 2021 Final: DWG KIA v EDward Gaming Prediction
Alacrity's LoL Worlds 2021 Final: DWG KIA v EDward Gaming Prediction
metric DWG KIA EDward Gaming
Market Odds 1.29 (-345) 3.61 (+261)
Implied Win Probability 77.52% 27.70%
Implied Win (Normalized) 73.67% 26.33%
Calculated Win Probability 59.38% 40.62%
Expected Value -0.2985 0.2859
Bet Size Bet @ odds > -119 Bet 17.87%

The odds market puts DK at a dominant 74/26 advantage. This is a best-of-five, so individual-game advantage is a bit lower (the more games in a series, the more exaggerated probabilities get). In a simple model, ignoring momentum and assuming the same game odds for every single game, this would equate to a roughly 64% single-game advantage.

Just to put in perspective how absurd a even a 64% edge is against a team of EDG's caliber: DK ended their LCK summer season with a 29-14 record, for a 67% win rate. They did about as well against the sum total of all nine other Korean teams as they're expected to do against the LPL's No. 1 seed and Worlds finalist.

If we add in their 33-10 spring season, that brings them up to an overall 72% win rate in their region, so sure, maybe the summer was a bit of a fluke for them. Still though, EDG have similarly dominated the LPL in that timeframe. Their 53-18 record comes out to a 75% win rate. I'd make the same claim as before... a 72% win in the historically strong LCK is incredibly impressive, but it shouldn't translate to a 64/36 matchup against the best (and similarly dominant) team out of an almost equally strong region.

Across a few other metrics, as well as region strength, DK do pick up an edge in my model, so they're still favored, but I think the market's a bit too bullish on them. A relatively small bet on EDG with a 41% chance to win the finals will pick up some expected value, as the payout is pretty generous.

Other match-level bets

You can find odds on a variety of other bets having to do with the overall match:


Exact score

One of the most straight-forward additional bets is on the exact final score. And hey, would you look at that, I happen to always simulate the probabilities of each outcome anyway:

Alacrity's LoL Worlds 2021 Final: DWG KIA v EDward Gaming Outcomes
Alacrity's LoL Worlds 2021 Final: DWG KIA v EDward Gaming Outcomes

These take into account all of the different paths a match can take to get to each final score. In addition, there's a momentum effect at play. Winning a game, especially for an underdog, tends to increase a team's probability of winning the next game in a series.

If we match these probabilities up against the odds market:

DK - EDG Odds Implied Probability Calculated Probability Expected Value
0-3 13.50 6.29% 10.90% 0.4715
1-3 8.25 10.29% 14.60% 0.2045
2-3 7.75 10.96% 15.11% 0.1710
3-2 4.80 17.69% 18.01% -0.1355
3-1 3.10 27.39% 21.11% -0.3456
3-0 3.10 27.39% 20.26% -0.3719

My model believes EDG have a better chance to win than the market does, so naturally I'm going to skew towards EDG winning more games as well. Any EDG-winning scores are going to be more likely than the market is pricing in, so there's some expected value to find there.

A note of warning: a typical, head-to-head bet with two outcomes typically carries a vig of around 6% (the overall winner bet clocks in at around 5%, which is surprisingly low, but there's going to be more volume on the market to support that for a high-profile game). Here, it's 15.1%, so the oddsmaker is taking a higher than normal rake to compensate for the harder-to-predict nature of this line.

Total games

Markets also offer an over/under on the total number of games. Since there are guaranteed to be at least three games in a best-of-five, the lines are simply set at 3.5 and 4.5 (in other words, betting on whether or not there will be at least four games for the first one, and whether or not there will be five games for the second).

Luckily for us, we just generated the probability of each possible outcome, so we can extract the total number of games from each of those scenarios. For example, the probability of fewer than 3.5 games means a 3-0 or 0-3 outcome, so we can just add up those probabilities. Compared to the market:

Number of Games Odds Implied Probability Calculated Probability Expected Value
>3.5 1.44 64.62% 68.84% -0.0087
<3.5 2.63 35.38% 31.16% -0.1805
>4.5 3.19 29.11% 33.12% 0.0565
<4.5 1.31 70.89% 66.88% -0.1239

Same logic as above; higher chances for EDG to win, and a more even matchup in team strength, means that the series is likely to go longer than the market believes. However, this difference isn't as big as I would have expected, which is somewhat surprising. The market is pricing in a 29% chance of this going the full five games, which seems a bit high for what is nominally a 74/26 matchup.

The vig is relatively reasonable on these bets, 6.95% and 7.14% respectively. A bit higher than a typical line, but lower than many other prop bets, so I wouldn't necessarily feel bad about taking any of these from that perspective. However, the vig coupled with the relative agreement between my model and the market means that the vig erodes most of the expected value.

Match handicap

Finally, markets offer handicapped odds on the winner of the match. For example, DWG KIA @-1.5 in the odds line above pays out if DK win by at least 1.5 games, so 3-0 or 3-1. EDG @ +1.5 is the opposite, and for example a 2-3 loss by EDG would still pay out here (with the handicap, it comes out to 3.5-3).

This isn't actually a distinct bet, however. As you can see in our examples above, these bets just pay out if any of a few specific exact match score outcomes occur. You can synthetically create any of these bets by just betting on multiple exact scores. The vig on these bets is slightly better, however, since each bet only has two sides, so that might be worth some extra value. With my predictions for each possible outcome, you should be able to combine them to get the probabilities of any of these handicaps, so I'll leave that as an exercise for the reader.

Game-specific bets

In addition to bets on the overall match, there are dozens of bets that you can place on each individual game. These include which team scores more kills, first team to various kill scores, first blood, total number of kills, total number of towers, and more esoteric things like whether or not both teams will kill Baron.

I won't offer any specific predictions here, sorry. The data for these exists, obviously, and it's possible to try to put together some sort of model for some of these, but most of these honestly seem pretty random and are going to have higher than normal vigs anyway.

In addition, a lot of these simply boil down to game length. The longer a game goes, the more kills there will be, so when you're betting on the total number of kills you're mostly just betting on how long the game is (which is already a bet you can make). You're really just adding another layer of uncertainty on top of an already arbitrary thing to predict. The same problem exists for total towers, total dragons, kill handicaps (it's hard to win by 10 kills if the game ends early), and really anything else involving totals.

A final caveat: there's an issue of incentive alignment here for the teams. Pretty much any prediction relies on the assumption that each team will be doing their best to make that outcome happen. For the overall winner, or even winner in each individual game, this is trivial. Each team wants to win, ultimately that's what matters, so each team will do everything in their power to win. All of these other things are tangential at best. They're all correlated with winning - getting more kills or towers or whatever is usually a good sign towards winning a game - but teams aren't going to explicitly target them in the same way. If a team's in a leading position, maybe they don't go for a fourth dragon when they can just end things, and if you'd bet on the total dragon kills you'll end up disappointed.

Of course, all of these are things that are just fun to watch and track, and therefore fun to bet on. If you enjoy being on the edge of your seat watching the game unfold with a bit of extra money on the line, who am I to stop you?

You can find this bet, and many others, on the suggested bets page. I cover every single professional League of Legends match (that I can get data on), so you can extract maximum value from the market.

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!