January 10, 2018
By: Doishy

Doomsday in 2017

We shall begin with a Pixar-esque bittersweet beginning. The deck is working, people are happy, we can run maindeck Abrupt Decay yet still deploy triple basics to fight the Miracles menace and then bam: Some idiot erects a roadsign and Wizards cripples a lot of creative decks to target a single one, which they fail to kill. Suddenly it's over. So many people, so many of the deck's pilots have had the final straw. They are done, that's it. Game over.

Is it the end?

I hope not.....

2017 has been a difficult year for Doomsday as an archetype primarily due to the SDT ban. A lot of it was spent trying to work out any and all possible replacements, both card wise and playstyle wise. A lot of people switched to ANT as in their mind the deck had lost the one benefit over it's bigger sister in terms of card selection, advantage and resilience.

I talk with a lot of people about the deck but one of the people I talk most with is Nanda aka Namida. He is a Doomsday veteran who resides in Japan and has always been a helpful and insightful voice on the Stormboards, TheSource, MtGSalvation forums and on the Doomsday Facebook group. One morning I awoke to find a veritable essay from him and his feelings on where the deck was right now. We had a very big debate over that essay with some strong points and feelings being conveyed from both sides.

I will avoid posting the full debate but I do want to use his points to address a lot of the frustrations pilots, both current and old, are having with the deck and hopefully give some comfort to those still playing, maybe even enticing some people back. I am looking to address each point truthfully. Although I have stated my goals I will not sugar coat things and there are some things that cannot be changed about the deck so I am not expecting a mass conversion of people but even if I can persuade one or two, I will be happy!

For many points I will not just be talking shit out of nowhere. Back in November I started a mini project to begin collecting statistics about the deck and I have put a call out to anyone who can to provide match results for the deck. These results have a number of data points that are needed to be considered "good data". I also stressed to all submitting I wanted all reports. All wins, losses, draws. I was not looking to get a rose-tinted view but instead wanted an accurate snapshot of how the deck was performing. This mix of pilots also allowed a mix of experienced and new people to get involved hopefully allowing a wide range of data sources and to minimise pilot variance to allow the deck itself to shine through results wise. The results pool is still small at the time of writing (337 matches recorded) but it still provides a good starting snapshot to get an idea of how the deck is doing. Any results involving the mirror match have been ommitted for any statistics involving game results.

Although I am presenting these statistics this article is just using them to help answer questions. I am planning to release a full article soon to properly describe and present the data and stats collected from the data so if any explanation is missing regarding the data that is presented here, it should be addressed in this later article.

Withut further ado, and continuing on the Disney theme...

Let's get down to business:

Here are Nanda's opening points:

"Building the deck again, I'm still feeling the same issues--basically I'm
frequently getting hands where you have to choose whether or not to stall
out by keeping cantrips as combo pieces or stall out by having no cantrips
when you finally assembled BBB and Doomsday. I'm realizing more how Top set
up a lot of the quick wins by being selection without costing a card to let
you look for Doomsday or Dark Ritual, but also because a lot of my Doomsday
wins were actually off of Burning Wish, Top and LED mana. Getting BBB feels
like a huge choke point for winning quickly. Does that make sense? 
Rituals feel bad, but you need one to actually be fast with Doomsday. 
If I don't open on Ritual and Doomsday in hand I feel stuck between not 
playing cards and hoping to draw into what I need, or playing cards and hoping
that I still have enough resources to win on the same turn by the time I find 
my combo pieces."

This block actually contains two points in one as I see it.

  • The deck is inconsistent.

The idea is that you have to choose whether to use cantrips to sculpt or to hold for the combo turn. Doomsday is unique insofar that cantrips double as combo elements to either draw into the stack, exchange between the stack and your hand, draw into a kill elements or draw to win with LM.

I will admit that yes, you do have to make these choices however that is unfortunately the core of the deck and the element that makes it unique to both TES and ANT. DDFT probably has the most elements of choice and the most potential decision points per turn. This is not even taking into account actually resolving the namesake card. Use of cantrips and lands/fetches has always been the hardest element of the deck and the most crucial. Saying that though, the current iterations of the deck I think play the most number of cantrips of any. Along with the U combo triplet of Brainstorm, Ponder and Gitaxian Probe we also run a number of Preordains and / or Conjurer's Baubles both of which add to the cantrip count. We are, like any deck in a game with some level of variance, at the whimsy of fate however unlike many decks we do have the tools to minimise that variance.

To work through this let us start by looking at our levels of variance, specifically, as Nanda mentioned, our opening hands. Mulligans were tracked across all games of a match and the win percentages calculated from when mulligans occur. Below you can find two graphs: the percentage of games where mulligans occur and the percentage of games won against mulligans.

Mulligan Percentage Across Games

Mulligan Win Percentage Across Games

As you can see, despite the concerns, the deck seems to mulligan very well. There could be some variance in that people are keeping dodgy seven card hands for fear of hitting a bad six however with the scry rule in effect, going down to six is certainly not the end of the world. This is also shown to be the case in the second graph. Seven cards have a very good win rate with a heavy reduction as one mulligans to six and to five. One thing to note is how few data points there actually are for mulligans to five or less showing some severe data changes for the win percentage across these points. As a result nothing solid can be inferred from these low points but I I hope the fact that the number of these results being so low is a testament to the general consistency of the deck itself.

Yes you will have garbage hands. We run some oddball cards that our often useless outside of an active combo turn and sometimes variance can really kick you in the proverbial nuts but most of the time, at least according to the numbers, we seem to be doing fine.

  • Making BBB for a quick DD without DR is difficult.

One of the issues with the deck is that the namesake card does require BBB to cast. This is an interesting element with a deck that is primarily U. In terms of options for creating BBB we have Dark Ritual as a four of. Rain of Filth as a potential one of. Lotus Petal as a four of and generally four lands that make B (normally 2 Seas, 1 Badlands and 1 Swamp). Nanda is correct in saying that DR is the best way to make BBB quickly however it is far from the only way. Two lands and a lotus petal is often enough or even one land and two petals can occur often.

"...a lot of my Doomsday wins were actually off of Burning Wish, Top and LED mana."

We also have the option thanks to Conjurer's Bauble to play out Doomsday on turn 2 via Burning Wish and two LEDs thereby replicating the 'saved' draw effect we gained from SDT. One of the more well known recent achievements of the deck, the 5-0 result from an MTGO by g0ld_rook had him run four CB and he noted that he had under appreciated the saved draw effect gained by them during early testing with them.

This point feeds into some of Nanda's later comments:

..."I'm sure the deck is fine if the game is going long, but I feel completely
toothless in the early game. If your hand isn't perfect then you're not 
racing most fast decks, and that means you have to use your cantrips more
aggressively and I don't think that's a winning proposition when every card 
you spend to look is actually costing a full card now."

Which creates a third point:

  • Doomsday is not a fast enough combo deck.

Of the "classic" Legacy storm variants you have three main camps (though some would argue only two). ANT, DDFT and TES. In order of speed it has always been considered to order TES as the fastest, ANT as second and DDFT as the slowest. This has often been true as TES sacrifices deterministic wins for speed and DDFT sacrificed speed for resiliency and card advantage. Now that SDT is gone where does the deck stand on this scale?

For this comparison I will be using data taken from two spreadsheets linked in an article written by Bryant Cook on the TES website. A link will be provided at the end for those interested in the article. To summarise it; Bryant took game data from a TES pilot and an ANT pilot and compared a number of factors including one listed as "Active Combo turn". This turn has been tracked on a scale of T1 to T4+ with the theory that beyond T4 there is little difference for a deck concerned about being fast between say T5 and T20. As part of my data gathering I also collected combo turn statistics. This is the turn the combo occurs so if you are executing a DD Pass-the-turn pile it is the turn you cast Doomsday to make the pile, not the turn you win on. This can create some results skewing but I will address that a bit more in a moment. First let's look at the raw data:

Data collated on 03/01/2018:

DeckAverage Combo Turn
ANT3.01
DDFT2.95
TES2.61

Wait what?

Yeah, that's what I said when I first crunched the numbers. How is DDFT, traditionally the slowpoke of the Legacy Storm world being on par if not a tad faster than ANT?! There are a few reasons for this, partially to do with the data skew I mentioned before. This can include turns where we "go off" without winning on that turn. This can mean that the "combo turn" may be the turn before we actually win (potentially as many as five turns prior to winning) which could skew the data. This also includes us using something in our arsenal which mimics that which TES has: Burning Wish into Empty the Warrens. This can be an early play which, even if only making ten or so gobbos can lead to fast wins. This does mean that the data will skew towards earlier turns but also means the TES data is affected by this bias. In the interest of transparency, it's worth noting that the DDFT data was recorded with a scale of 1-5+ turns, which has been converted to correspond with the TES and ANT range of 1-4+.

When discussing this point regarding speed with Nanda he made the following comments:

"Doomsday is being carried on the back of Burning Wish at that point is
how it has been feeling. Not like Doomsday is doing anything powerful, 
but that we're playing a faster Storm deck with pet cards instead of
the cards that make it a consistently fast threat. I've been liking when
I draw Burning Wish in these versions of the deck because we have so much
fast mana. I have felt miserable anytime I had a Doomsday in hand." ... 
..."I want to be wrong about the deck not winning in a timely manner, but
I can't reconcile the idea that the mana is too awkward for Doomsday
to consistently be a Turn 2/3 threat without having to spend cards you
may not be able to afford spending. That sums up the issue for me, I think."

This gives us point 4:

  • Doomsday is now just a worse ANT/TES.

These are fair comments. It is true that cards like Past in Flames are incredibly powerful engines that in many situations are less akward and more consistent than Doomsday itself. Our data above including EtW results definately skews it faster so what if we take out those results and consider only the wins that use the deck's namesake card? We'll also have a look at the breakdown in how the deck actually goes about winning to work out how much the statement of...

"...we're playing a faster Storm deck with pet cards instead of the
cards that make it a consistently fast threat..."

...is true.

DeckAverage Combo Turn
ANT3.01
DDFT2.99
TES2.61

Focussing only on the DD winning lines we see the turn average has definately gone up but is still on par with ANT. Following on from this point again Nanda had real concerns that Doomsday as a deck was not winning using its namesake card. Let's look at the breakdown of winning game engines:

Combo Engines Across Games

I guess that paints a clear picture. The four engines measured are the primary ones of any line that uses Doomsday itself, BW into EtW, natural ToA kills and then anything else (such as weird SB plans or other alternatives like Telemin Performance). We can see that a vastly overwhelming majority is using the namesake card. Does this make the deck better than the others? No, not inherently. Does it mean that it is unique enough to not be a bad TES/ANT clone? Yes, I think so.

Nanda ended his message with the following final point to make:

Sorry for the rant earlier. It's frustrating because it feels like having to learn to
walk after losing a leg. Is no one else feeling this way? I want to stay the course
but I keep getting discouraged by how even just my goldfishing is playing out.

Which gives a final point of:

  • Post SDT ban DDFT is not the same as pre-ban DDFT.

Now I have no fancy stats or graphs for this. It's true. DDFT is a different deck now. We have only made a few changes, going from 16 lands to 15, replacing IU with AoI, upping to 4 BW and using Preordain and Conjurer's Baubles but they are noticeable ones and yes, it will take time to re-learn. Cantripping is now even more important, knowing when and how to use each card is key and the sheer number of potential tricks you can do thanks to exiling via AoI is crazy but it does take time. Many, myself included have had to go back to the basics of the deck which can really make you feel down. I don't blame anyone who has put the deck down for good as a result of this frustration. You're not alone. I will however say that DDFT is still alive, it's still relatively competitive (overall match win % at the time of writing is 56.69%) and it is still a lot of fun to play, punishing as all hell when it goes wrong and a highly rewarding deck when it goes right.

Closing Thoughts

I'm sorry for this information splurge as it is a lot to take in.

At the end of this article resolving we are left with the following five point stack having been addressed:

  1. The deck is inconsistent. No more than any combo deck
  2. Making BBB for a quick DD without DR is difficult. Potentially but often not
  3. Doomsday is not a fast enough deck. Slower than TES, as fast as ANT
  4. Doomsday is now just a worse ANT/TES. DDFT is it's own style, not just a clone
  5. Post SDT ban DDFT is not the same as pre-ban DDFT. Correct and we as pilots have to work through this

Hopefully this has helped you as the reader come to make a decision whether to put the deck down, pick the deck up or just understand what happened to the deck in general. As stated at the start; I have presented a small cross section of statistics to help answer some of these points but these are a small taster of what we have been tracking. In the near future I plan to write a more detailed article simply going through the stats and giving more information on how they are collected which should help explain any further data bias concerns people might have on what is said here.

If you wish to contribute to our community data collection for the deck you can find the link to the spreadsheet below.
If you wish to collect the data but not in a public forum then please feel free to take a copy of the spreadsheet and if you could send that link to Doishy via theSource forum or the Stormboards that would be appreciated. A couple of people currently do this for the sake of running their own trending but if we could use the data for overall trending, that would be appreciated.

Thanks for reading!

Bryant Cook's Article spreadsheets visited 03/01/2018
g0ld_rook's 5-0 list
Doomsday Stats Collection Spreadsheet

Tags: DDFT, Stats, Comparison