TESPA Regional Playoff Spring 2019 Report
Welcome back to the TESPA Hearthstone Meta Report! It's been a while since we shared one of these, but we took a break due to lack of interest. With the new expansion fresh and few tournament results with the new cards to look at, we decided to share an analysis of how decks are doing in the new meta. One of our goals in this report is to identify strategies that could be relevant in the upcoming Hearthstone world championship and future rounds of TESPA, which are also in the conquest format. Some of the advantages of the data in the Regional Playoff are that most of the teams play at a reasonably high level (teams get to confer and nearly all have at least one legend level player), and are close in skill (since weak teams were eliminated and no pros are present). This should hopefully be a somewhat fair representation of deck power.
In order to analyze deck power, we matched the deck codes submitted by each team to the team's match results. This was repeated for each round of swiss and the double elimination bracket at the end. If a team didn't show up for a match due to dropping out or didn't play for a round, the team's decks and the opponents' decks were not collected as data in that round. Data was only collected on the Open Division, because teams in the Varsity division played too few matches for the sample to be much use in analysis, and the matches they did play were in small groups where teams may face the same opponent more than once.
So lets dive into the data. While relying on one metric to assess a deck's strength in conquest is a bad idea, I believe that the best indicator of a deck's strength is its popularity in top 8 compared to its popularity in the beginning of the tournament. This is a good indicator because it can reveal the power of a deck when it is used as intended in conquest: to take advantage of the ban or to counter a deck since all 4 must win.
Tier List: By Top 8 Percentage
Tier List: By Matches Won
In the charts above, Popularity represents the number of teams that brought an archetype to TESPA. Archetypes that were played for less than 20 matches in total were omitted from the tier list due to insufficient sample size. Percentage of the Field w/ Deck represents the percentage of teams that brought the deck as one of their four options. Percentage of Matches Won is the proportion of matches that teams using the deck won. The more popular decks are forced to have win rates closer to 50% because the odds of two lineups with the deck facing each other is very high. In such instances, one team must lose and one must win, driving the win rate to 50%. Finally, Popularity in Top 8 is the number of teams with a deck that got to the top 8 stage in their respective regions. 32 teams made it to top 8, since there are four regions.
Percentage Growth to Top 8 is the percentage of teams with the deck in top 8 divided by the percentage of teams with the deck at the beginning of the tournament. Numbers above 100 signify that a deck was able to make it to top 8 more often than competing decks, while numbers below 100 signify that a deck was perhaps responsible for an early end to teams that used it.
While top 8 percentile is in my opinion the best way to represent deck power, there are some issues with using it as the only metric. Therefore, we also included the tier list sorted by the percentage of matches won by teams with the deck. It's immediately obvious that there are some discrepancies in how decks performed across the two metrics used here. In one tier list, Mech Paladin is far above the others, and in the other it is dead last. This is why I'm choosing to call it tier X. It's hard to say exactly how good a deck is when some individuals do really well with the deck but the average team does poorly. Freeze Mage and Midrange Hunter also had respectable win rates, but didn't have amazing showings when looking at top 8 performance.
One possible reason for the struggles of some Mech Paladin, Mech Hunter, and Conjurer Mage players is that many individuals with these decks didn't bring all three of them. When inspecting lineups as combinations of decks, one common trend emerged: those with all three of the above decks did very well while those with only one of them did poorly. When comparing this trend to stats on HSreplay, an explanation emerges: all three of these decks beat warrior very very reliably, but have very poor win rates vs popular decks like token druid, rogues, and zoo. Even though the mage, hunter, and paladin are unlikely to beat the common aggro decks, the vast majority of teams brought either bomb or control warrior. It would actually be an enormous struggle for warriors to dodge sweeps from this lineup if not for the lack of a another strong anti-warrior option. Several teams just opted for a warrior of their own with greedy techs, hoping that they could win the mirror or beat aggro first to get a win and deny warrior winning from the opponents to take the match. If world championship contenders can find the right 4th deck for this lineup, they could do really well by farming the control warriors that will almost certainly be around. If they try to pair druid or rogue with one of these, they could let the warrior through and lose the match.
Since there are so many new decks and interesting tech choices in the TESPA field, we will now look more closely at the tools each class has, and if they were any good this past weekend.
Mage
On the graph to the left is a depiction of how a few decks' prevalence changed over the first 5 rounds of the swiss and how it compares to the top 8 prevalence (the last point in the line). Round 6 was cut from these charts because not all regions had a round 6.
Over the course of the swiss rounds and going into top 8, the field size of remaining teams held pretty steady for the first three rounds, then rapidly started to fall. Midrange Hunter is a good example of how much a deck can change with a rapidly shrinking field, many teams with the deck dropped going into round 4.
In the case of Mage, this graph is interesting because of how flat it is. Neither of these Mage decks had a large presence in the TESPA field, but their presence in the tournament remained pretty stable throughout the swiss, with small bumps in prevalence towards the end. This suggests that most players to bring the decks at least did decently well, perhaps because of how well the mage toolkit can target certain strategies popular in the meta. Conjurer mage's ability to slam huge minions turn after turn makes it a solid option vs warriors. Freeze Mage's aoe effects make it potent against the zoo and token druid decks popular in the tournament, and it can close games effectively against rogue and zoo because the classes rely on using their health aggressively as a resource and often enter burn range as mage approaches 10 mana. This is less true against token druid, but that deck has no answer to nova+doomsayer besides soul of the forest into it, and the mage doesn't have to kill all the boards like shaman and warrior, it just has to stall until a burn setup.
Warlock
If you look either of the tier lists above, you can see that Zoo Warlock was the best performing deck in TESPA. The likely reason for this is that warrior is one of the best decks, so banning it makes sense if you want to bring other good decks. Zoo is held back on ladder by the warrior match up, but is also the only reliable counter right now to token druid, which over 80% of teams in TESPA brought. Therefore, Zoo slots really nicely into ban warrior lineups that want to bring the three great classes: rogue, warrior, and druid.
Unfortunately, Zoo seems like the only good way to play warlock at the moment, a couple teams tried control but had no success.
As a long time Hearthstone player, however, one of the first thoughts I had upon seeing the Zoo decklists was where are the soulfires? Soulfire has been a staple in the deck since its creation, but about half the TESPA teams decided they were better off without the card. Therefore, I statistically analyzed the success of both builds. In the chart above, you can see that the builds without soulfire experienced more growth as teams were eliminated than builds with the card. Additionally, I found that teams without Soulfire in their Zoos won 55.56% of their matches while teams with the card won 54.21%. A possibility is that Soulfire is a weak card against token druid 1/1s and 2/2s, and token is the deck zoo lineups are targeting. The card may be better in a ladder environment.
Priest
Priest had a really bad time in TESPA Regional Playoffs. The biggest issue for the class is that no one played it, although doing so was probably a good decision. Only 1 team had silence priest, but the team went 3-2 so there may be some hope for the deck. However, 3 teams brought Resurrect Priest, and they all suffered. On average the Resurrect Priest teams won 30% of their matches.
Druid
While the class didn't have as much diversity as Rogue and Warrior in TESPA, Token Druid was more popular than either of these classes in Regional Playoffs. And this popularity was a good idea. The win rate of the deck is 50%, which tells us nothing since almost every team had it. However, the deck did have above average representation in the top 8.
One question I have seen come up a lot on the Competitive Hearthstone subreddit is if Crystalsong Portal is actually a good card in the deck. So I used the stats from TESPA to try and determine an answer.
Just like for Soulfire in Zoo, I tracked the survival of lineups with and without crystalsong portal. While lineups with crystalsong portal stayed relatively constant in meta share, lineups without the card steadily increased in meta share over the course of the swiss stage. The win rates of lineups with and without the card were pretty much the same. Yet by top 8, the meta share of portal token druid was the same as it was at the start of the swiss, and no-portals token druid had grown in meta share by 30%! It seems that token druid has plenty of gas without crystalsong portal, and that the deck is better off with a sticky bomb like Eccentric Scribe or more 1 drops to win the board early.
Rogue
While nearly every team in the tournament brought a rogue of some kind, there was a pretty even divide between two different approaches. One idea was to play a small burgle package to utilize Vendetta, and rely on the burgled cards for gas into the late game. The other popular idea was to use Myra's unstable element as gas and try to kill the opponent as quickly as possible, without wasting time and energy on burgle effects.
Overall, lineups with the burgle package had a higher win rate than lineups with the Myra burn plan. Burgle teams won 53.4% of their matches while Myra decks only won 49%.
However, while Myracle struggled a lot in the early rounds of swiss, it dominated in the later rounds, with a 110% growth into the top 8. Burgle actually had a lower meta share in top 8 at only 89.42% of the meta share at the start of swiss.
Myra's meta share drop from round 1 to round 2 actually looks like the teams dropped curve for rounds 1 to 2 (shown on the right). It seems like many of the teams to get knocked down to 0-2 early were running Myracle Rogue. Perhaps the deck is more punishing of bad play than the burgle build of rogue.
Or perhaps Myracle is just more punishing of bad tech choices.
While nearly everyone running the burgle package had at least one shadowstep in their rogue deck, there was a divide among myracle rogue teams on the card. However, it seems like it was not the best choice in Myracle. 42.6% of teams running shadowstep in rogue won in the first round of swiss. This is a really low win rate considering almost half the field had a rogue with shadowstep. While burgle decks with shadowstep didn't struggle much round one, many Myra decks did and dropped out before round 2. This early dropout suggests that shadowstep Myracle lists were significantly deflating stats for Myracle, and that once they were gone the deck did very well, as can be seen in the rising meta share into the top 8.
Shaman
While it got some good stuff in the new set, Shaman didn't do very well in TESPA playoffs. It's possible that the best way to use the new tools hasn't been discovered, but looking at HSreplay none of the class' archetypes seem amazing vs warrior, druid, and rogue, the three leaders in the current meta.
While only used by two teams each, Malygos Shaman, Control Shaman, and Big Shaman all appeared to do ok. One control shaman actually got top 8, suggesting that the archetype could at least be a tool for countering rogue even if druid and warrior are very challenging. Lineups with one of these three decks won about half of matches, so there is certainly potential to explore slower shamans.
Of the slower shaman decks, those running Shudderwock experienced the greatest difficulty. It was a fairly popular archetype, with 10 teams bringing Shudderwock control shamans. However, those teams on average were only winning 43.75% of their matches, and only 2 teams to bring it got top 8 meaning the deck fell in popularity as rounds progressed.
Murloc Shaman, however, looks like it is one of the most overrated decks of the new set. Both overload builds and builds without overload synergy struggled, experiencing a decline in meta share even early in the swiss. While neither deck looks good, the overload package may make the deck slightly less bad. Lineups with no overload only won 40% of their matches, while lineups with overload won 46.7% of their matches.
Hunter
While it lost one of the best cards in Hearthstone history to rotation, Deathstalker Rexxar, Hunter seems to still have a place in the meta due to its ability to contest warrior effectively.
Midrange Hunter seems like a decent deck that is only held back on ladder by its horrible match up against rogue. When rogue is banned away in conquest, midrange hunter can do decently well. In TESPA, Midrange Hunter had a really strong match win rate at 54%, but it didn't have much representation in top 8. Midrange isn't a counter to token druid and warrior, but it has a good chance against them, making it an alright 4th deck for conquest lineups banning rogue. Yet teams in the top 8 had more coherent strategies, and opted for decks with more favorable match ups rather than a solid lineup filler like midrange hunter.
Mech Hunter, according to the stats in our tier list, is a terrible deck. However, much of its problems are due to its inclusion in lineups not targeting warrior. Mech hunter has quest rogue level match ups against aggro, but 4 teams still took it to top 8 by using it as intended with decks like conjurer mage and mech paladin to defeat the warriors nearly every team brought.
Paladin
As most TESPA teams realized early in the expansion, secret paladin has a problem. The deck runs out of gas too quickly without cards like divine favor and call to arms to maintain the great tempo turns 2-4. The rotation of good control tools like steed and Tarim also mean that paladin has to be proactive on the board, and can't pass the early turns. However, Mech paladin is able to use the secret package as good early game before transitioning into mech tempo plays starting on turn 5. Secrets like redemption are also very strong with mechs, since they can have powerful deathrattles.
Mech Paladin was placed into tier X in our tier list because it's hard to evaluate when some teams do terribly with the deck and other do really well. However, I believe that the deck's slight edge against token druid and frustrating minions in the warrior match up make it a really good way to target warrior in lineups that ban aggro. The only issue for the deck is the zoo match up. If the best lineups start to include both rogue and zoo, the deck could run into trouble because it can only ban away one of them.
Warrior
The vast majority of TESPA teams brought a warrior of some kind last weekend, but there was a pretty even divide between control and bomb strategies. Bomb warrior is much better at taking a proactive game plan against slow combo decks, since the bombs can deal lots of damage against draw heavy strategies, and Dr. Boom can push tons of damage. However, teams with control warrior did substantially better overall in the tournament. Bomb warrior certainly doesn't seem like a terrible deck with a win rate of 48.2%, but it had far lower representation in the top 8. Since most teams are probably banning warrior due to its anti-aggro potential regardless of bombs, this could be a sign that control warrior does better against some of its counters. Control has more room for tech cards like hecklebot and silence that could be useful.
Or perhaps control warrior's top 8 percent growth is due to teams not banning it when they should, leaving the warrior an easy win. Control warrior is a harder counter to token druid and rogues than bomb warrior, and these were two of the most popular decks around.
Closing Thoughts
It seems like Zoo has a powerful role to play in conquest as the token druid counter. If Zoo, rogue, token druid, and control warrior is the default lineup at worlds, perhaps strategies like control shaman or freeze mage could abuse the deck like Fr0zen's big spell mage at the last world championship. Zoo certainly has its weaknesses.
Or maybe lineups countering warrior could find success, the idea of mech paladin, mech hunter, and conjurer mage is a great start that only needs a fourth member that beats warrior.
If you are wondering what tech cards to use on ladder, try cutting soulfire in zoo and crystalsong portal in token druid for more board development. Aoe is at its weakest in a long time, so exploiting that could be effective even by cutting cards that seem powerful.
Also give freeze mage and control shaman a try. These decks seem like they could have potential, but there isn't much to look at for stats on HSreplay. Even the stats that exist could be wrong since there are so many competing builds with substantial differences.
Thanks for reading the whole article, I know it was long. Let us know what you enjoyed and what we could improve in future write ups. We might take one last look at conquest once worlds are over.