Cloud9 - LCS Week 6

If you've been following along, you'll have seen that while we agree with the market on the probable outcomes of almost every single LCS game this weekend, we're quite a bit more bullish on Cloud9. They're already favored in all three of their matchups, but we think that the odds market doesn't accurately capture just how favored they really are against Dignitas and FlyQuest (we're roughly in agreement on their Evil Geniuses matchup).

While there's no simple way to really show why a team should be an 80% favorite rather than a 70% favorite, we're going to walk through some of the statistics we look at to determine our probabilities, and thus why we think Cloud9 is going to be so dominant going into this weekend.


Overall Probabilities

metric Cloud9 Dignitas
Market Odds 1.32 (-312) 3.2 (+220)
Implied Win Probability 75.76% 31.25%
Calculated Win Probability 82.42% 17.58%
Expected Value 0.022 -0.5976
Bet Size $275.03 $0
metric Evil Geniuses Cloud9
Market Odds 2.33 (+133) 1.56 (-179)
Implied Win Probability 42.92% 64.10%
Calculated Win Probability 37.71% 62.29%
Expected Value -0.2379 -0.1063
Bet Size $0 $0
metric Cloud9 FlyQuest
Market Odds 1.27 (-370) 3.54 (+254)
Implied Win Probability 78.74% 28.25%
Calculated Win Probability 85.20% 14.80%
Expected Value 0.0185 -0.6531
Bet Size $303.83 $0

In summary: we Cloud9 is ~7% more likely to win against both Dignitas and FlyQuest than the odds offered would imply. Against EG, we actually just split the vig (our calculated probability for each team is lower than the odds-implied probabilities, which is possible because the vig makes them add up to more than 100%)


Cloud9 v Dignitas

Simply put, the individual players of Cloud9 are just better than those of Dignitas. This is fairly common knowledge, and everyone agrees Cloud9 is favored here, but to put it in perspective with their pro-game winrates (weighted to count recent games more heavily):

player_name winrate team_name
Zven 61.68% Cloud9
Fudge 59.81% Cloud9
Perkz 59.25% Cloud9
Blaber 59.07% Cloud9
Vulcan 58.07% Cloud9
FakeGod 46.50% Dignitas
Neo 46.28% Dignitas
Akaadian 44.60% Dignitas
Aphromoo 44.34% Dignitas
Yusui 40.37% Dignitas

With Zven coming back, Cloud9 regains its strongest individual player, though their entire roster is incredibly strong (for comparison, LeBron James has something like a 66% winrate in his career).

Cloud9 also plays better as a team; here, we show each player's relative winrate differential when playing with each of their teammates, versus when playing without that teammate:

For example, Zven's already impressive 61.68% winrate goes up by 2.36% when playing with Blaber.

Dignitas's players, on the other hand, seem to play worse together than they do on their own (though this could be attributed to their recent struggles as a team).

And, finally, past clashes between these players have also favored Cloud9:

Akaadian's effectiveness against Cloud9's roster is fairly muted simply due to not having very much data there, but the rest of the Dignitas roster performs poorer against any of Cloud9's players than they typically do.


Cloud9 v FlyQuest

The FlyQuest matchup is pretty tricky, now: on July 6th, FlyQuest announced quite the roster shakeup for this week, whereas we had generated our prediction before this news on the assumption that they would field their regular squad.

This is an unfortunate risk of wanting to place earlier bets; however, on average, placing early bets still tends to favor us pretty substantially!

In addition, the odds market had access to the same information as us when they set their lines! When we placed our bet, the odds we were offered would have also made the same roster assumption that we did. What does this mean? Well, we aren't inherently favored or unfavored with this shift, for one thing. Cloud9 is just as likely to be more favored than before as it is to be less favored (in fact, we'd expect them to be even more favored, given that they're playing FlyQuest's academy roster). Turns out the market agrees: odds have barely moved; the payout for Cloud9 hasn't moved at all, while FlyQuest has shifted from 3.54 to 3.6, so they're slightly less favored to win and paying out slightly better.

We would actually believe this new roster to have a better chance against Cloud9, though that's largely driven by relatively little precedent of these players facing each other before. Models are impossible to fit to every scenario perfectly (especially without overfitting), so we can apply some qualitative logic here and disagree with our model on this one. We should never actually place bets by this method, but since we'd already placed this bet a week ago, it should help give us some mental relief that things aren't too far off (and with the new roster, we would have simply suggested that there's no edge in betting on either team).

Since the roster change has flipped things around so much, we won't break out all of our metrics as we did for the Dignitas game, but suffice it to say that Cloud9 still looks pretty strong here!


Big Money

On top of our (relatively) unusual stance of betting on the favorite (as always, we just bet where there's value, it just so happens that odds are more likely to underestimate the underdogs), we've also recommended placing unusually high bets on these matches, relative to the absolute difference between our calculated probabilities and those of the market. We've had plenty of matches where we think an underdog is ~10% more likely to win than the odds would imply, and yet suggest fairly conservative bets. What gives?

Well, our bet sizing methodology is asymmetric in this regard - the short version is that we're willing to make higher-value bets on more likely outcomes, because while the EV may be the same, our risk (and variance) on that EV is lower.

For a given difference in calculated probabilities relative to the market (the four lines below), here's how our bet size changes based on the probability implied by the market:

For example, focusing on the purple line: if our calculated probability for a given team winning is 10% higher than the market, we'd be willing to bet ~13% of our unit bet size if we believe them to be a 35% rather than 25% winner, but ~38% of our unit bet size if we believe they're an 85% rather than 75% winner.

Our disagreement with the market is the same, but because it's a more likely outcome we're willing to stake more on it.

Thus, we're willing to place fairly high amounts on Cloud9 simply because as the favorite, we're much more likely to see our desired outcome.

We'll go into much more detail on some of the math and nuances behind this in a future article covering everything related to bet sizing, but wanted to highlight this here to whet your appetite for such analysis!


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!