Gaming
esports
RNG Did Not Win the Hearthstone World Championship
Pavel's flashy plays turned heads, but the raw numbers reveal it was planned all along.
"Pavel is one with Babbling Book."
"Pavel is a literal god."
"Pavel is manipulating reality to his will this tournament."
"I'm never going to count Pavel out of a game of Hearthstone in my life."
Those are just a few of the many ways that announcers at the Hearthstone World Championship used verbal gymnastics to avoid saying that they think Pavel got lucky.
And at first glance, it looks like he did. The Russian underdog quickly rose through the ranks of Hearthstone's biggest tournament of the year at BlizzCon last week, and many of his wins were capped off with big, flashy plays that used random card effects to gain an advantage.
The online community, in particular, was quick to criticize Pavel, who knocked out NA's darling Amnesiac in the quarterfinals. "Pavel" is being used as a verb synonymous with getting a lucky result and "Babbling Book" is now referred to as "Paveling Book" after his exceptional results with the card in the tournament.
The consensus is in. It's been widely accepted that Pavel got extremely lucky and won the tournament because of his RNG advantage. But are we sure that's what happened? Let's take a look at the cold, hard numbers and see if Pavel really has RNG to thank for the nice, new trophy sitting in his room.
This is Pavel's happy face
© Blizzard Entertainment
How we calculated
Before we start, a quick note on our calculations: The stats in this article cover cover all 19 games that Pavel played in the Bracket stage of the Hearthstone World Championship (against Amnesiac, JasonZhou and DrHippi).
A "Random Event" is defined as any action caused by a card played by that person with an outcome that is affected by randomness.
A quick example: Playing Sylvanas onto the board is not a Random Event. But Sylvanas's Deathrattle effect is counted as a Random Event for the player who played Sylvanas. However, if Sylvanas only has one target to steal when she dies, it does not count as a Random Event since there was no randomness in the actual outcome of that event.
Card draws are not tracked as Random Events and are not reflected in these stats, as it is very difficult to objectively evaluate whether each card draw is lucky or unlucky at its time. We will, however, address a few particularly important card draws directly in our Case Studies below.
Pavel's win was as legitimate as they come
© Blizzard Entertainment
The raw stats
Let's take a look at some global stats:
- Pavel played 19 games in the Bracket Stage.
- In 2 games, Pavel had no Random Events.
- In 3 games, his opponent had no Random Events.
- In 1 game, there were no Random Events for either player.
- 52 of Pavel's total 294 Events were Random Events (18.3%).
- 53 of his opponents' total 315 Events were Random Events (16.8%).
Right off the bat, it doesn't look like Pavel had an unusual reliance on Random Events in his playstyle or decks. He's essentially identical to his opponents in that regard.
Next, let's take a big-picture look at just how lucky or unlucky those individual players were in their matches against Pavel. It's very difficult, if not impossible, to objectively sort Events into specific categories of luck. There's some subjective evaluation required, so do not take this next set of stats as specifically literal. Instead, use them as a broad indicator of trends. For instance, we actually had to invent that category mid-research after two of Pavel's plays had such good outcomes that it'd be unfair to lump it in with the existing "Very Good Events." There's nothing quite like getting Leeroy Jenkins off of a Firelands Portal to secure lethal damage.
You can look at how we categorized each Random Event in this (unpolished) spreadsheet, and make your own adjustments if you disagree.
In Pavel's 52 Random Events, the outcomes were:
- 9 Very Bad
- 6 Bad
- 10 Average
- 13 Good
- 12 Very Good
- 2 Very Very Good
Total: 61.5% non-average outcomes, with 27 good, 15 bad (+64.2%)
Pavel's fans cheer him on
© Blizzard Entertainment
Amnesiac's 15 Random Events vs. Pavel:
- 0 Very bad
- 4 Bad
- 1 Average
- 4 Good
- 6 Very Good
- 0 Very Very Good
Total: 93.3% non-average outcomes, with 10 good, 4 bad (+71.4%).
JasonZhou's 12 Random Events vs. Pavel:
- 0 Very Bad
- 1 Bad
- 7 Average
- 2 Good
- 2 Very Good
- 0 Very Very Good
Total: 41.6% non-average outcomes, with 4 good, 1 bad (+80%).
DrHippi's 19 Random Events vs. Pavel:
- 1 Very bad
- 1 Bad
- 9 Average
- 4 Good
- 4 Very Good
- 0 Very Very Good
Total: 52.7% non-average outcomes, with 8 good, 2 bad (+80%).
There are two big takeaways from this batch of data. First, none of Pavel's opponents had any Random Events that qualified for "Very Very Good."
"The second big takeaway is that Pavel's average luck was slightly worse than all of his opponents. Pavel still had significantly more good outcomes than bad, but it was closer to 50/50 than any of his opponents in these matches.
And while Pavel had higher highs with his good luck Events, he also had lower lows with his bad (9 Very Bad outcomes compared to only 1 Very Bad outcome for his combined opponents).
The tempting conclusion here is that Pavel's extreme variance in luck — particularly his higher highs — contributed heavily to his ability to win. But let's dig into a few of his most dramatic Random Events and see what the actual probabilities were.
Case Study No. 1: Polymorph onto Malygos (vs. Amnesiac)
We've got to start with the biggest upset of them all: Pavel's miracle draw into Babbling Book to get Polymorph just in time to kill Malygos. And then drawing a second Babbling Book into Firelands Portal to kill Emperor Thaurissian the very next turn! (This turn of events starts at 1:13:30.)
This series of plays drove the announcers into a frenzy and is easily the most popular example of Pavel's critics. So let's dig into it.
- Pavel had a 15.3% chance to draw Babbling Book on his first turn after Malygos.
- Pavel had a 8.3% chance to draw Babbling Book on the next turn, after the first was drawn.
- Pavel had a 29.5% chance of drawing at least one Babbling Book across those two turns
- Pavel had a 1.2% chance to draw both Babbling Books in a row.
So drawing both of those Babbling Books was absolutely a string of very, very good luck. Getting one Babbling Book to answer Malygos wasn't too unlikely, but getting both back to back is incredible. However, getting the books is only half the battle. Let's look at the odds that Babbling Book would give spells that could answer the threats Pavel was facing.
Babbling Book can give 29 different spells, and while Polymorph is the obvious favorite to remove Malygos immediately, there are other options for clearing the Malygos in two turns. Forbidden Flame and Pyroblast would've easily done it. Shatter might've been able to combo with Water Elemental (in Pavel's hand next turn). Vaporize and Cabalist's Tome are not guaranteed to work, but might have. Counterspell or Spellbender could also limit Malygos's impact. For these stats, let's assume Pavel would be reasonably happy with 5 of those 7 alternative options (a fairly big assumption).
- Pavel had a 6.9% chance of getting either Polymorph from the first Babbling Book.
- Pavel had a 13.3% chance of getting either Polymorph from either Babbling Book.
- Pavel had a 24.1% chance of getting an acceptable Malygos answer from the first Babbling Book.
- Pavel had a 42.4% chance of getting an acceptable Malygos answer from either Babbling Book.
- Pavel had a 27.5% chance of getting an immediate Emperor Thaurissian answer from the second Babbling Book.
Even with our fairly big assumption, Pavel's odds never even get close to the 50/50 point. He clearly beat all of the odds in this Random Event to immediately find not one, but two perfect answers to threats on the board.
The question is then did it ultimately matter that he got lucky here? To answer that, we have to look at the game state when these Random Events happened.
Pavel runs a minion-heavy Mage deck and had spent most of the early and mid game forcing Amnesiac to use his damage spells clearing the board. Thanks to that tactic, Pavel had made sure that, by this point in the game, Amnesiac only had 3 Malygos-buffing spells left in his entire deck! Amnesiac had already used both Raven Idols and could only hope to draw 1 Wrath, 1 Swipe and 1 Living Roots.
To make matters worse, Amnesiac never drew his Swipe that game. His available cards for the rest of the game were (in order, separated by turns): Wrath, Innervate // Arcane Giant // Nourish, Feral Rage, Living Roots, Mire Keeper // Fandrel.
Pavel had emptied Amnesiac's deck of threats long before Malygos could show up, so it's questionable if Malygos sticking on the board longer than one turn would've definitely caused Pavel to lose that match. It's possible that Malygos's biggest impact might've just been the 4 damage that he punched with every turn, in which case Pavel would've had many turns to find an answer before dying.
So while it is incredibly good luck that Pavel found the answers immediately after the threat was played, it's not entirely clear that he needed to answer the threats as quickly as he did in order to win. It's very possible that he would've still won without his extreme good luck.
Case Study No. 2: Sapping Sylvanas (vs. Amnesiac)
This is the other back-breaking play that drove Pavel's opponent bonkers: drawing into Sap the turn after Sylvanas is played on the board. (Fast-forward to 1:25:30 to see this play.)
At first glance, it looks like a miracle draw, but Pavel's Rogue deck cycles heavily, constantly drawing cards to make sure that he always has the tools that he needs. So let's look at the math.
- Pavel had 20 cards left in his deck at the start of his turn after Sylvanas was played.
- Pavel drew 3 cards during this turn.
- Pavel had a 28.4% chance of drawing Sap during that turn.
- Pavel had an 11.1% chance of drawing Sap specifically on his last draw.
At first, that looks like pretty good luck again. Pavel beat the odds by a decent margin, right? Well, not if you put those stats into context.
By the end of that turn, Pavel had been offered over half of his deck (3 mulligans at start, and 13 draws during the game). There's a 79 percent chance that Pavel would have been offered one of his two Saps by then. So, the odds are actually very high that Pavel should have had a Sap in his hand by the time Sylvanas was played.
In fact, you could argue that Pavel was already playing in the fairly unlucky (21 percent likelihood) scenario where he didn't already have Sap in hand, ready to go. You might say that he was due to draw it.
Of course, that line of logic gets messy quickly. That change would have impacted all of his previous decisions, and the decisions of his opponent, and who knows what would've happened. Even still, it's difficult to call this scenario a particularly lucky Random Event in the grand scheme of the match — it's probably better described as Pavel's earlier bad luck (which led to him not already having Sap) correcting itself and approaching balance.
Case Study No. 3: The Game Without Random Events (vs. JasonZhou)
As a fun bonus, RNG haters should definitely check out Pavel's match against China's smile-prone pro, JasonZhou. It's Pavel's Malygos Druid taking on JasonZhou's Dragon Warrior, and not a single Random Event happened in the entire match!
And here's a big fat spoiler for you: Pavel won the only game with no RNG in it. That doesn't necessarily mean anything important for our analysis, since it's an isolated incident, but it's an interesting tangent.
Case Study No. 4: C'Thun's Victory Stomp (vs. JasonZhou)
In their sixth and final match, JasonZhou had Pavel's C'Thun Warrior deck on the ropes — JasonZhou had 10 over lethal damage for his next turn, with 11 of the damage coming from minions on the board. But then C'Thun came. (This moment happens at 1:15:00.)
Pavel's 14/14 C'Thun had to clear the entire board if he was going to survive another turn. Let's look at what happened.
C'Thun dealt 11 damage to minions, and 3 damage to face.
All three minions (4 health, 4 health and 3 health) were killed.
At first, that sounds lucky, but let's break it down. Each of the four potential targets (three minions and face), took 3-4 damage. That's actually as average of a result as you could make. There are no outliers there — everything looks perfectly reasonable.
In fact, two of the minions were killed very early, which lowered the chances of hitting any minion in the later turns (but, obviously, it worked out in the end). That early bad luck required the last 4 shots from C'Thun to have good luck: Namely getting at least 1 damage onto the 3/3 so Pavel's own 2/3 could trade into it.
Once again, Pavel got a little more luck than he needed and C'Thun outright killed the minion (a 25 percent outcome). A nice upside, but unnecessary for Pavel's survival. Even that small bit of luck is hard to pin as a black mark on Pavel. By then, he already had a 93.7 percent chance of surviving and winning — the only outcome that lets JasonZhou win (his 3/3 takes no damage from 4 shots) has the significantly lower probability (6.3 percent).
Pavel holding up his well-earned trophy
© Blizzard Entertainment
Conclusions
As a general verdict, across all of his games, Pavel beat the odds — his Random Events helped more than hurt. But, that's true for all of his opponents. The thing to remember about the Random-Event cards that the pros choose to run is that they're, first and foremost, just great cards.
Pros use them because the Random Events that they generate are more likely to have good outcomes than bad. They wouldn't include them in their decks otherwise. So when you look at the raw luck of pro players, every single one of them appears to be "lucky" because the cards they're using aren't pure coin-flips. They function more like rigged dice guaranteed to roll 6s more than 1s — and that's why pros like them.
By many metrics, Pavel's overall luck actually appears to be worse than his opponents. For every Babbling Book, there's a Brawl that kills 1/1s to leave up a 9/9 Taunt Medivh. The biggest difference seems to be that Pavel got his biggest upsets at very flashy, dramatic moments. His few miracle Random Events came on key turns and were completely unparalleled by his opponents. His highs were higher.
But those miracle outcomes also tended to come at moments when Pavel didn't actually need them — he was, in most cases, already guaranteed or likely to win without them.
For example, I didn't bother to dissect Pavel's Leeroy Jenkins off of Firelands Portal against JasonZhou in Game 4, because Pavel already had lethal damage in his hand ready for next turn, and Jason had no way to stop it. That didn't stop the crowd from losing their collective minds when it happened, though. It was incredible hype, but ultimately unnecessary.
Likewise, I ignored Pavel's C'Thun game-ender against DrHippi in Game 4 because it was a guaranteed victory for Pavel no matter where the damage landed. Clearing the board was fun and exciting, but DrHippi was already doomed with no cards left in deck and an only an unplayable Baron Geddon in hand. Pavel had once again carefully extracted his opponent's threats and eliminated them one by one until he had the advantage. C'Thun was just the flavor for the kill.
So why did he win?
Based on his interviews and some of the unique deckbuilding choices he made, it doesn't seem that Pavel has any particular preference for random-effect cards. In fact, it might be the opposite. All but one of his decks were exceptionally low in Random Events. His Druid deck had 2 games with no Random Events at all and his Warrior deck created only 11.5 percent of his total Random Events.
But his Mage deck has a lot of Random Event cards in it. In fact, it's where 36.5 percent of his total Random Events came from. But that's just the way Mage is right now. Some of Mage's best cards have random effects: Flamewaker, Arcane Missiles, Firelands Portal, etc. If you want to build the best Mage deck, you simply have to embrace the random effects — all the other pros' Mage decks did too.
This stark contrast between Mage and the rest of Pavel's decks points to his biggest strength, and the real reason why he won the tournament: he's very flexible. No matter what came his way — from his deck or his opponent's — during the tournament, Pavel assessed it, adapted to it and then made it work for him.
It's easy to focus on the few incredibly lucky Random Events that benefited Pavel, but you have to balance those with the 15 "bad" or "very bad" Random Events that happened to him too. He got hit with more than double the number of negative Random Events than his combined opponents (15 to 7), and still came out the winner!
More than anything else, Pavel proved he's a master at rolling with the punches at this tournament. He likes the guaranteed cards, but he can win without them. His exceptional ability to adapt on the fly and identify obscure, risky outs kept him alive through the hard times so he could make it to the good times.
RNG didn't win the Hearthstone World Championships. Pavel did.
Josh Augustine is a freelance contributor. He currently works as a game designer at Daybreak Games.
For more Hearthstone features, follow @redbullesports on Twitter and Like us on Facebook.
