On the 16th January 2021, the KeyForge Premier League hosted two qualifiers for their Season 2 League. The format for both qualifiers was Archon Solo with no restrictions - players could bring a deck from whatever set they wanted, as long as they could verify ownership or reasonable access to the deck.
The purpose of this article is to look at the overall field of each event, how those fields narrowed down to the top 16 decks, and examine the meta in these events (with some caveats).
First, a definition of what ‘meta’ means in this context. The metagame is the ‘game outside the game’. So, while we’re all playing KeyForge, and thinking about how to win games, we’re also thinking about the social strategy: what other people are likely to bring, and what choices we can make to put ourselves in a better position before the game starts. An example of this would be, for example, “lots of people are using Infurnace, so bringing a pip-heavy deck or a deck which relies a lot upon recursion is risky.”
We’re going to be looking at the overall fields for both the European-friendly (Qualifier 1) and Americas-friendly (Qualifier 2) Archon Solo events. The size of the field in each event makes this a somewhat limited sample of data (59 and 75 decks in each respectively.) In addition, there were 20 decks which were played in both events, which is going to affect how similar the statistics are across both data sets. As a result, this analysis has the potential to be very inaccurate as a representation of the meta in the wider global community.
Using Decks of KeyForge’s fabulous tag system and analysis tool, I’ll be looking at deck statistics, house and set breakdown, most common cards, SAS score ranges and more.
So, with that out of the way, let’s have a deep dive into each event…
Qualifier 1 (9AM GMT/4AM EST) - General Field
The first qualifier predominantly featured Mass Mutation (MM) decks - out of a total of 59 decks, MM was 28 of them, making up just short of half the field. Age of Ascension (A0A) made a particularly poor showing here, with only 2 players choosing to bring AoA decks with them. The rest of the field was evenly split between Call of the Archons (CotA) and Worlds Collide (WC).
With MM and CotA making up two thirds of the field, it is unsurprising that the most common Houses, as well as House combination, was Dis, Logos and Shadows (in that order). 9 of the 59 total decks fielded in this event were Dis/Logos/Shadows, making up 15.25% of the total decks. Interestingly, a close second was Logos/Saurian/Star Alliance decks at 8 total decks (13.6%), consisting of 2 MM and 6 WC decks.
The SAS range for this qualifier was between 70 - 91, and the inter-quartile range was 78 - 85. The median score was 81, which has a SAS percentile of 99.57%. This means that, on average, the decks used were within the top 0.5% of decks in existence (as evaluated by SAS). That is insane!
Looking at other metrics;
- The average AERC Score was 66.3
- The average META score was 0.93
Next, we're going to look at the top 10 cards in this deck pool. (There's 11 as there was a tie.)
The top cards come largely from Dis, Logos and Saurian.
Infurnace takes first place as the most common card brought. As a combination of aember control and disruption through purging cards, its popularity really comes at no surprise.
Largely, these cards can be split into three categories:
Disruption tools: Infurnace and Mark of Dis narrow your opponent's options.
So, what decks from this field made it through to the Top 16?
Qualifier 1 (9AM GMT/4AM EST) - Top 16
The set breakdown of the decks which qualified were pretty much proportional to the set data from the general field, with MM taking up over half of the decks in the top 16. Sadly neither of the AoA decks made it through.
The house breakdown is almost identical in terms of proportions, too. Combination-wise, Dis/Logos/Shadows was the most common, present in 4 decks (25%) - 1 CotA deck, and 3 MM decks. Every MM deck in the top 16 contained Dis as one of its houses.
I thought it would be interesting to look at the lists of the 3 MM Dis/Logos/Shadows decks to check for similarities, so, here are the lists (beautifully presented by SkyJedi's Archon Matrix bot):
In terms of placement, Сварливый Стас Перерва made top 8, and the other two decks made top 16. They all have quite similar lists, with multiple cards shared across two or even all three of the decks. Let's have a look at some:
Every deck has Infurnace, and two have multiple copies. As mentioned previously, this is a strong card for disruption and aember control.
Two decks have a copy of Etan's Jar. This is a very versatile artifact with huge disruptive power, especially if the opponent has little to no artifact control.
Ultra Gravitron. DOUBLE Ultra Gravitron! This giant creature provides huge levels of efficiency alongside creature control.
Two decks have multiple copies of Bo Nithing. As the steal effect increases with the number of keys forged by the opponent, Bo Nithing becomes very punishing in the late game and provides an incredible catch-up mechanic.
"Borrow" is in all three decks as artifact control. Taking control of artifacts is generally extremely effective, and adds to your advantage. Unfortunately it's not much use against artifacts like Etan's Jar or Quixxle Stone.
Unfortunately none of these decks were able to make it the top 3 spots for KFPL qualification,
Moving on to the SAS statistics:
The SAS range of the top 16 is somewhat reduced from the general field, but still pretty wide at 75 - 89. The inter-quartile range is 79.5 - 86, with a median of 82.5, which is slightly increased compared to the general pool, but not massively different.
In terms of other statistics:
The average AERC was 69.1
The average META score was 0.8.
These are pretty much in line with the general pool - not a huge amount to see here.
Next, the card pool:
Unsurprisingly, Infurnace is still the card which appears the most in the top 16 decks. However, there's a few interesting differences between the lists. Firstly, there isn't a single CotA symbol on the set list here - all the most popular cards appear exclusively in the newest sets, WC and MM, despite CotA cards being reused in both sets. This might imply some power creep in the newer sets, but again, this is a very limited sample.
Sinder gets to make the cut this time at the second-highest frequency - looks like taunt in MM did some work!
The cards in this pool seem to be more disruption-focused than the general pool - Infurnace (with Stirring Grave for recursion) and Mark of Dis still rank highly. What's missing here compared to before is the board control cards - Axiom of Grisk and Gateway to Dis have dropped off, and that was definitely reflected in the extremely board-heavy top cut matches.
Finally, let's have a look at the top 3 decks, the pilots of which received KFPL Season 2 invitations:
A tentative glance reveals more of the same - high-efficiency MM Logos and control-heavy MM Dis. Interestingly, Shadows doesn't feature in any of these decks despite being the third-most popular house in the event.
A more in-depth analysis can be found here from Screech Bob-Omb Commentary:
In summary, players in the earlier qualifier favored MM decks with high-efficiency Logos and disruptive Dis, which dominated the top cut and was key to the top 3 decks.
Qualifier 2 (5pm GMT/12PM EST) - General Field
The set distribution of decks entered in the second qualifier is largely proportional to the first event, with a few exceptions. AoA was a much more common choice, and while MM was still the most popular choice, its share of the field is was bit less than the earlier qualifier (38.7% in this event versus 47.5% in the previous event.)
The house breakdown was also largely proportional to the first event, with Dis, Logos and Shadows being the most popular choices. However, Untamed was a much more popular choice in this event, while Saurian numbers are reduced. The most popular house combination by a huge margin was Dis/Logos/Shadows, making up 16% of the field.
The SAS range was much larger in this event, stretching from 66 - 95 SAS. However, the inter-quartile range is quite similar to the first qualifier's, spanning from 76 - 84 SAS with a median value of 79.5. While this is affected by the two events sharing 20 decks (26.6% of the deck pool in qualifier 2), it does suggest that people in both events brought decks of similar quality.
Other statistics from this are:
The average AERC was 66.7
The average META score was 0.81
These are largely the same as the previous event.
The most popular cards in the event mirror the previous, with a few exceptions. CotA cards make up more of the top picks, and all but one of them are in Dis. Infurnace still takes the top spot, but with a smaller overall presence (there were 16 more decks in this event, but fewer Infurnaces). The reduced number of Saurian decks in the event means that there are no Saurian cards in the top spots - in fact, every card is either Dis or Logos. This implies a much more limited pool in this event than the previous. Again, these cards can be widely categorised into Disruption and Efficiency cards, with only one board wipe (Gateway to Dis) appearing very frequently.
Let's review the top 16 decks from this event:
Qualifier 2 (5pm GMT/12PM EST) - Top 16
The top cut here is quite different to the previous event, with CotA having the largest representation. This is disproportionate to the entries - while the general field was 26.67% CotA, in the top cut CotA makes up 43.75% of decks. A tempting conclusion is that CotA decks took up the space taken by AoA decks in the general field, but this isn't entirely correct - the proportion of WC decks in the event was also reduced. CotA dominated the field here.
The houses represented are largely in line with the proportions in the general pool. Dis, Logos and Shadows continue to be the most prevalent houses. Shockingly, this combination was massively over-represented in the top cut, making up 31.25% of all decks in the pool. Of the 5 Dis/Logos/Shadows decks which made the cut, 4 of them were CotA decks. Let's have a look at their lists:
In terms of placement, both Odin E. L'Aggressivo and Shrewd "Anteater" Ice placed in the top 8, and the other two in the top 16. With even a cursory glance, it's impossible not to see the similarities in these lists. Let's look at each house and see which cards are most prevalent:
Control the Weak (CtW): between these 4 decks, there are 8 copies of CtW. CtW is an extremely powerful control card due to its ability to make opponents effectively skip their turns, providing the user with a significant card advantage, as well as locking out counter-plays.
3 of the 4 decks feature a Timetraveller. These decks can be more robust against Infurnace decks, as you can shuffle your discard back into your deck to protect your cards.
Lights Out - featuring in most decks, this bounce effect is particularly strong against exalt/capture-heavy WC and MM line-ups.
Miasma: Making your opponent skip a key step is pretty good, it turns out.
The SAS range for the top 16 is slightly tighter than the general pool, but still very wide at the high and low extremes, ranging from 72 - 88 SAS. The inter-quartile range is 79 - 84.5 SAS, with a median of 81. This is somewhat tighter than the first event's top 16, with a huge concentration between 79 - 81 SAS. Fun fact: this quartile is entirely CotA decks, with two of them being Dis/Logos/Shadows.
Other statistics for this group are:
The average AERC was 68.5
The average META score was 0.94
The top 16 cards for Qualifier 2 are drastically different to those in the top 16 in the first qualifier - the list is dominated by CotA Dis cards, providing a huge amount of immediate disruption and control. Infurnace makes an honorable appearance in this top list, but not nearly at the frequency as the previous event. Logos efficiency tools (Daughter, Wild Wormhole) also make the list, in line with their representation in the general pool.
Let's have a look at the top decks:
Like the first event, MM decks have placed incredibly well with the high efficiency levels brought by MM Logos. The diversity is a bit more obvious here, with MM Star Alliance making a surprise appearance at the 11th hour.
A more in-depth analysis of these decks is presented by Cull the Weak here:
In summary, although decks with a limited pool of houses (featuring 2 or more of Dis, Logos or Shadows) made up most of the Top 16, the decks whose pilots were awarded KFPL Season 2 places showed a somewhat larger variety of decks and different styles of play than the general pool.
Is there a Meta?
Given the trends identified in the data, I think it is clear that there was a meta for these events. Players tended towards Dis and Logos decks with high efficiency ratings and powerful disruption tools. 5 out of 6 decks in the combined top 3 were MM decks, demonstrating its popularity and relative strength. Something which I haven't yet touched on is that 3 of these decks contain MM Saurian - these decks tend to have a large amount of aember control coming from Amphora Captura, Faust the Great and similar tools. This leads to a capture-heavy board state which is vulnerable to bounce effects like Hysteria, or full board clears like Gateway to Dis. However, the ability to reliably increase your opponent's key cost is a huge advantage over the traditional CotA control decks, which have a reduced focus on maintaining a board and potentially lower aember generation in comparison as a result.
Was the Meta Different Between Events?
I don't think there was a significant enough difference in the deck pools to say there are independent metas in the Americas-friendly (75 players) and European-friendly events (59 players), especially with the 20 deck overlap between the two datasets. Similar choices in sets and houses were made with minor divergences, and the decks which made the top 3 of each event had overlapping houses and toolsets. What is safe to say is that MM has become a competitively dominant set, and has also provided a more diverse competitive environment through a larger variety of deck strategies and tools.
Here is a breakdown of the stats for each event, examining the SAS, AERC and META Scores, the prevalence of Dis/Logos/Shadows decks in the total pool, and the representation of each set in each pool. Bolded values are the highest values across the 4 groups.
The only major differences between the statistics is the set breakdown between events and their cuts. The disproportionate representation of CotA in the Americas-friendly top cut may suggest that players in that region were better at identifying decks which were strong against the types of threats posed by later sets. We identified some of these features earlier in the article, so this would make sense. Conversely, the opposite could be true - that players in the European-friendly qualifier were better at selecting decks which would counter the dominant CotA archetypes. Either way, there does seem to be a small difference between the two metas, but not enough to say they are distinct.
Deep Dive into Dis
In both events and their top cuts, Dis was the most common house (with the exception of the Americas-friendly top 16, where Logos was slightly more popular). The popularity of Dis was overwhelmingly in CotA and MM decks.
Below is a chart showing how common Dis was as a house in each qualifier within the pool of decks from each set. For example, Qualifier 1's general pool had 2 AoA decks, 1 of which contained Dis, rendering the share 50%.
What's interesting here is the efficacy of each brand of Dis in each of the events. The more successful MM decks in Qualifier 1 contained Dis, whereas the CotA decks were more diverse. The reverse is true for Qualifier 2, where every CotA deck that reached the top cut contained Dis, but the representation in MM Dis was proportionally smaller.
One reason for this could be that each brand of Dis performed better within the individual event metas. Let's look back at the most used cards in each of the top 16s which were Dis, and look what each card brings to environment:
The cards in Qualifier 1 are purge-focused, especially upon Infurnace and its recursion with Stirring Grave. This provides a large amount of aember control, as well as threat elimination. Mark of Dis can force your opponent into sub-optimal turns and prevent the opponent from answering your threats. The rest of the list is made up of common creatures, with Sinder (high power, taunt) providing protection for the battle-line.
Qualifier 2's cards are more focused on reducing the impact of the opponent's turns - locking them into sub-optimal choices or even turn-skips with Control the Weak, eliminating the board with Gateway to Dis, and reducing the cards an opponent can use through Ember Imp. These disruption cards generate an immediate card advantage if they are used correctly. Dust Imp provides aember gain without providing a juicy Infurnace target. Infurnace is still in the top cards, but at a much lower level than in Qualifier 1, and without the recursion.
So, why did these different flavours of Dis enjoy different degrees of success in the qualifiers? It may be down to how quickly each of the card pools has an impact on the game. Infurnace with recursion has a more punishing effect the longer a game goes on, as the impact of losing a high value card tends to have its largest effect upon a reshuffle. CotA Dis cards change the pace at which a player progresses through their deck on the first run-through, and gives the player an early card advantage which hopefully carries them through the game. If the meta is largely decks with a long game plan, then Infurnace will have a more punishing effect, but a meta with quicker decks which don't expect to reshuffle, and don't rely on recursion, will be less affected.
Extra data would be needed to investigate this - the amount of turns in each match, with reference to the sets and houses in play, would shine further light on this.
What About the Stats?
I thought I'd take a section of this article to address the relevance of checking the SAS, AERC and META scores of the decks in the article, given there's not a huge amount of difference between them. The META Score has interested me since it was introduced by Decks of KeyForge as an analysis tool. It is described by the Decks of KeyForge site as "represent[ing] how well a deck's AERC is distributed." The AERC score is a measurement of key metrics, such as Aember Control, Expected Aember and Artifact Control. A further description of what AERC measures is available on the Decks of KeyForge About page.
As described on Decks of Keyforge's About page, currently, the META score consists of:
A (Aember Control) from less than 5 to less than 2 gives -1 to -4 META.
C (Creature Control) from less than 6 to less than 2 gives -1 to -3 META.
R (Artifact Control) one or more hard R gives 1 META.
3+ trait strength worth of board clears is worth 1 META.
Although it's not mentioned, decks also gain a META point if they have scaling aember control, such as Too Much To Protect or Bring Low. Added together, the maximum META score attainable is 3, while the minimum is -7.
The consistency of a positive META Score in the event decks of between 0.81 - 0.97 shows the success of it as an evaluation tool for competitive decks - on average, the decks people chose to bring had a reasonable level of the accepted tools for success.
I think there is a question here of chicken or egg - did players choose to bring decks based on their strong statistical scores, or do the scores accurately reflect the quality level of the deck? The large SAS ranges, especially within the top cuts, seem to suggest that there are still many decks whose contents are underrated, or indeed overrated, by SAS and AERC as evaluation tools. Of course, there are a huge variety of different factors in play - the amount of reps a player has put in with a deck, how the player feels on the day, technical issues and sudden emergencies can all impact on performance in an event. I think it is safe to say, though, that big numbers don't automatically result in a win.