oraymw's picture
By: oraymw, Oraymw
Sep 23 2016 11:00am
Login to post comments

Welcome to Ars Arcanum, the MTGO Numbers based article series. In this article, we’re going to take a numbers based approach to analyzing the Kaladesh spoiler. It’s been a while since I wrote one of these articles. I used to be that WotC had MTGO replays available for viewing, and I would gather data on those, but that is long dead now, and I’ve been going through a sort of processing period as I tried to figure out how I wanted to write about Magic going into the future. This article represents a new direction for me and my writing. Over the past year, I’ve been doing a lot of work to come up with new ways to look at and analyze Magic sets, and to get a better understanding of what makes them tick. This project culminated in a post called History of Creatures in Limited over on my own website. Once I did that, I was able to look at Magic sets with a little bit more accuracy, and that will help inform the way I look at sets in the future. (I would recommend looking over that post, since it gives a more detailed explanation of what all the different indicators mean, which forms a lot of the terminology I’m going to use in this analysis.)


This series will be based on a few key principles. First, the most important part of understanding and limited format is to analyze the creatures. They form the biggest chunk of a format’s cards, and they are the main way that limited decks win. Second, the goal of these articles is more about analysis and exploration than about prediction. I find it more productive to try to raise questions and look at a range of possible answers than to try to perfectly predict a format that no one has yet played. Third, the foundation of these articles is the concept of rarity weight; when I run the numbers, I base them on how frequently a card will show up in a draft. This means that commons have a much more important impact on a format than rares.


With formalities out of the way, let’s look at the numbers!



Creature CMCs



Our first place to start is at converted mana cost. No part of a Magic card has a bigger influence one what the card can do, so it makes sense to start in the upper-right corner. The above chart shows the average creature CMC across various sets, after being adjusted for rarity weight. We can see a slight trend upwards in creature CMCs, but nothing tremendous. The main thing to notice is that we had a glut of high average CMCs since about Avacyn Restored, but that the past three sets have seen a marked decrease in average CMCs. Kaladesh has the most similarities to sets like Theros, Gatecrash, or original Innistrad, though it pops in at the same place as Coldsnap and Scars of Mirrodin and Conflux.


It’s easy to look at low converted mana costs and assume that this heavily influences a set to be faster, but there simply isn’t a strong connection between CMC and format speed. Having lots of low cost creatures can mean that everyone is forced to act earlier in the game, but if many of those creatures are of a defensive nature, then the format is going to slow down overall. Avacyn Restored had a much higher average CMC than Eldritch Moon, for example, but those sets are pretty far from each other in the context of modern limited sets. Likewise, having expensive creatures may mean that you are trying to get to where you can cast those cards, but it might also mean that your aggressive creatures are filling an important power vacuum and become ever stronger. Simply looking at this part of the set doesn’t give you a good sense of how fast the format will be.


The average CMC in the set comes in at about 3.3, but it’s worth remembering that these tend to be heavily skewed to the right, and the median CMC is 3. This means that more than half of all the creatures in the set fall below the arithmetic mean; the vast bulk of creatures in the set cost either two or three mana, with the next highest number coming in at four mana. In your typical draft there will be 8.6 six mana creatures on average, or a little less than 1.1 per drafter. The two and three mana slots are the most important. If the set ends up being faster than average, then it will be important to fill out your curve with as many of those cards as possible since everyone else will have plenty. If the format ends up slower than average, then the competition won’t be very fierce for two drops, so you’ll only want to focus your energy on taking the very best ones highly.


There are two other important consequences for having such a glut of creatures at 2 and 3 mana. The first is that it makes five and six drops a little bit worse, because they are competing against players that are able to just play two creatures on those turns. Even in a slow format, it can be powerful to get play out your creatures quickly and then restock your hand with a draw spell, which actually becomes even easier when you powerful 2 and 3 drops, since you could draw cards and still play something that has a meaningful impact on the board. The second consequence is that it makes it difficult to play cards that don’t have an immediate impact on your board state between turns two and five. Since everyone has a lot of cards to play on their board at those stages of the game, you want to be careful not to waste your time playing things that just sit off to the side. The obvious offenders here are vehicles; it’s important not to rely too heavily on vehicles in any deck, but especially as the bulk of your low curve, since you’ll need to be playing cards that impact the board immediately on turn w and forward. Instead, it’s useful to have these non-combat cards come down a little bit later and fulfill a stallbreaker sort of function.



Creature Size



Combined Average Power and Toughness



Efficiency: Comparison of Power/CMC and Toughness/CMC


While converted mana cost is a great starting point for looking at creatures, the indicator that has the biggest impact on any limited format is the size of the creatures. The two charts above show two different perspectives on creatures; the first shows the average total power and toughness of creatures in the set, while the second shows the power against the toughness, but divided by converted mana cost. The first gives us an idea of the absolute size of the creatures, while the second gives us an idea about the efficiency of the creatures and how quickly power and toughness can be added to the board.


In the first, we see that Kaladesh has the third biggest creatures of all these sets, coming in just behind Rise of the Eldrazi and Mirrodin Besieged and just ahead of Battle for Zendikar. I should note that this includes vehicles as creatures in the calculations, which puts this number slightly higher, but even without vehicles it would still be in the top five. Kaladesh has big creatures, and a lot of these creatures are at common; the big difference with these other sets is that their big creatures came at high mana costs. They were sets that were designed to slow their formats down so that you could play out big creatures. Kaladesh doesn’t do that; the big creatures in this set come down earlier and are easier to cast, and they’re in a format that seems to give few ways to really slow things down. The kinds of creatures driving these numbers are cards like Glint-Sleeve Artisan, Gearseeker Serpent, Ambitious Aetherborn, Thriving Rhino, or Peema Outrider. All of these creatures are pushing their size for their mana cost within their particular colors.


The really key chart is that second one. In that chart, we see that Kaladesh clocks in with the highest power per converted mana cost of any other set. Even without considering vehicles, Kaladesh still comes in as the highest power per CMC set, ahead of sets like Gatecrash or Zendikar. It also comes in as the highest toughness per CMC, but not by such a wide margin. There is no regular magic expansion in history for which you could buy more power and toughness for such a low price. To really drive this home, take a look at the following chart:



In this chart, we see a scatterplot with Absolute Size on the Y-Axis and Efficiency on the X-Axis. It shows a clear delineation between old magic sets and new magic sets right about at the Conflux/Alara Reborn line, marking what I call the two eras of creatures in limited. Kaladesh is a set unto itself. It is as much different from contemporary limited creatures as those sets are from the previous era. If sets were to continue to be made with the Kaladesh formula, it would mark a third era of creatures in limited. The size of creatures in Kaladesh is unprecedented, and it’s impossible for me to accurately predict how this will affect the format. The closest comparisons are Khans of Tarkir, Journey into Nyx, or Gatecrash, but even those sets are just simply not in the same class.


I cannot drive this home enough. The creatures in Kaladesh are shocking.


One of the goals for these analyses has always been to try to get an idea of the speed of the format, and I used to use Power/Toughness Differential as an indicator for that, but now that I’ve had a chance to get more complete data and run some regressions, I’ve found a few more accurate indicators. The problem with P/T Differential is that it doesn’t hold up very well at predicting the speed of formats when we look at old sets, so I’m going to go ahead and retire that metric. But I’ve found a very strong indicator to replace it, which is Power divided by Converted Mana Cost (or P/CMC). The following chart shows that relationship:



Of note, the way that I measured speed was by just ranking the sets subjectively. I don’t have a good objective method of format speed, though I’d love to have one to make this more accurate. The problem, of course, is that with replays no longer available on MTGO, there’s not a good way to get an objective measure. Feel free to adjust your opinions on this chart based on your own speed determinations.


No other indicator shows a stronger relationship to the speed of the format than Power/CMC. This chart shows that even a small increase in Power efficiency leads to a large likely increase in format speed. It’s not a perfect predictor, and so it’s better to look at this indicator along a range of possible speeds. The following chart shows a red line that estimates the range of speed ranking that Kaladesh is likely to have:



Based on this indicator, it seems that Kaladesh would rank somewhere in the 8 fastest limited formats, along the lines of Zendikar, Gatecrash, or Avacyn Restored, but possibly near Innistrad at the low end. I should note that I haven’t had a chance to test the predictive power of this indicator, since this is the first set for which I’m rolling out the comparison. I’m excited to see where Kaladesh ends up in comparison to other formats. It’s also worth mentioning that there are other indicators that correlate with speed, and Kaladesh doesn’t score as high on those indicators, but it’s still worth keeping in mind that what we are about to see with the release of Kaladesh in unprecedented in normal Magic set releases.



Creature Utility



Creatures aren’t just power and toughness, and the key mechanics and abilities in the set have a major impact on the way a format develops. Of all the creature indicators, utility is the squishiest. Its goal is to try to capture this sense of the power of the words inside a creature’s textbox, but those things are so intangible and difficult to quantify. Utility is a sort of method to boil that down to a metric that can be compared across multiple cards, but its inherent weakness is that different kinds of creature utility are not truly comparable.


One important detail out of the utility score is that the sets that tend to score highest on Creature Utility are the ones with a lot of synergy among their creatures. Sets like Theros, Scars of Mirrodin, or Lorwyn all scored very high on utility because the cards in those sets combine very particularly and multiply their force together. Lorywn had the tribal theme, so each creature made each other stronger with a high density of the right creature types. Scars had infect and metalcraft, and Theros had heroic and devotion, and all of these are examples of mechanics that lead to a lot of decks that are focused around particular synergies. These kinds of high synergy sets tend to be a little bit faster than other sets, so there’s a small connection between the utility score and the speed of the format. It’s not nearly as strong as what we saw with P/CMC, but it’s still a better predictor than any of the other indicators.


The above chart shows that Kaladesh doesn’t score particularly high on creature utility. This comes down to two major factors. The first is that a lot of the creatures in Kaladesh have huge bodies, but they don’t come with a lot of non-combat abilities. But another key part is that it is hard to get a good read on the utility scores for mechanics like Energy and Fabricate. I like to score on the conservative side until I have a better understanding of the format, and I anticipate that once the format has been out for a few months, that I’m going to have to go back and do a heavy revision. It’s possible that entering the battlefield with a couple of energy is a huge bonus and that you’ll be trying to jam every card that can generate energy into your deck that you can find. But it’s also possible that you only really want the best energy producers. If these mechanics end up having more than anticipated synergy and strength, then Kaladesh will move up on the utility scale, and probably end up being an even faster format. My suspicion is that Energy and Fabricate will end up being more modular than focused, which would lead to more open-ended decks and gameplay rather than pushing people down a very tight path of strictly aggressive Energy and Fabricate decks.



Evasion, Card Advantage, and Drawbacks





These three make up the remaining key indicators. The other indicators are more important; they have a bigger impact on the overall strength of the creatures in the format, but all three of these indicators are important and contribute to the overall picture of how a format plays out. Fascinatingly, Kaladesh doesn’t score particularly high in Evasion or Card Advantage, though it ends up scoring very well in the Drawback category. The creatures in Kaladesh are big, but that mostly defines them.


In evasion, Kaladesh scores fairly low for a modern set, coming in around the range of sets like Innistrad, Dark Ascension, or Morningtide. The biggest question I have about that is how servos will interact with a format that has relatively low evasion. The creatures are big, but many of them can just be chump blocked, and there are going to be a fairly high number of chump blockers running around. It’s also possible that this leads to complicated board stalls, especially in sealed.


Kaladesh also doesn’t score particularly high on Card Advantage, following below the trendline for contemporary sets. This is coming after Shadows over Innistrad, which had the highest all time score for Card Advantage, mostly on the back of clues. Most of the card advantage for this set is centered in servo tokens, but I can’t really count those as a full card, so that drags down the numbers a bit. It’s also possible that energy counters play like some portion of a card, but I couldn’t really count them that way in the analysis.


Kaladesh scores particularly low in the drawback category, and that’s even with Vehicles coming attached with a tremendous drawback score. Without vehicles, Kaladesh would score near the bottom all time, only being beaten by sets like New Phyrexia or Eventide, which had some of the easiest to cast creatures of all time. There are two major kinds of drawbacks; the kind that make a card difficult to play until later in the game but allow big payoffs, or the kind that come on things that are easier to use early, but lose value in the late game. Multicolor is a good example of the first, but a mechanic like Vanishing is a good example of the second.


One of the things I’ve tried to do with the creature analysis is develop a model that assigns a weighted rank to the creatures in each set. It’s weighted in favor of the creature size stats, but is informed by all of the key indicators. I don’t mean for the ranking to be taken very seriously; it’s just a vague stab at how creatures might rank, and gives a general idea of where a set falls historically. Here is the rank chart with Kaladesh included:



Kaladesh comes in at 5th place all time for creatures, mostly because of its low scores in the Utility, Evasion, and Card Advantage categories, but this chart still demonstrates a clear trend towards better and better creatures that hasn’t stopped in the past couple of years. I doubt we’ve reached the peak of creatures in limited yet, especially if Kaladesh indicates a new standard for creature size.





The other most important category for limited is removal. Before I put together the history of creatures in limited, I’d already done a similar project for removal, so that framework is already in place. You can find the article where I broke down those categories at The History of Removal.



The most surprising thing to see when counting the percentage of removal by rarity weight is to see that that number has basically stayed consistent for over a decade. Kaladesh comes in exactly within the normal fluctuation around the average, so we’ll have about the same percentage of removal spells that we’ve seen for over a decade. There’s going to be around 3 removal spells per pack, but the quality of the removal is what really makes a big impact on the set.



The last four magic sets have seen a dramatic reversal in the trend for removal. We had seen removal becoming more and more expensive over time, but with the release of Oath of the Gatewatch, we saw a sudden and dramatic turnaround of the cost of removal. The cost of removal spells in Kaladesh looks something more like Shards of Alara than what we came to expect from sets like Khans of Tarkir, Battle for Zendikar, RTR block, or Innistrad block. I know there’s going to be a lot of people that find this number particularly comforting considering the strength of the creatures in Kaladesh; surprising, the average cost of removal spells in the set comes in a little bit lower than the average cost of creatures. This means that you’ll be able to sometimes deal with something and gain a mana advantage.  



Here we see another major shift in the kinds of removal we’re seeing. We had seen a consistent and dramatic increase in the restrictions on removal spells for years, but now we are seeing what seems to be a new era of removal spells that is scaling restrictions back to what we saw in the days of the best removal. Kaladesh comes in with the 5th fewest restrictions on its removal spells all time, which means that you’ll generally be able to use your removal on the creatures that need to be killed, and you won’t have to worry as much about fumbling around to try to find the right situational spell for the situation. Sure, we’re in a world where every creature has to be dealt with or it will kill you post-haste, but it seems that there will be options for interacting.



Effectiveness basically explains how dead a creature is when you kill it. Kaladesh scores right in the middle of this category; it has several pacifism style effects that incapacitate a creature but give opportunities for them to return, as well as some bounce spells that let the creature come back, but it’s also got several things that just kill a creature dead and get it off the board, sometimes even at instant speed. This is especially useful in a time when many of the bombs in a format are creatures, and now there are ways to make sure you don’t just lose to those.



This chart tries to get an estimate of the amount of card advantage that you get from your removal spells. We had seen a big long term trend towards having less and less advantage from killing creatures, sometimes even getting sets where you had to take a major resource blow to try and kill something, but since Battle for Zendikar, we’ve seen a little bit of a reversal of that trend. There still isn’t a huge amount of extra advantage from removal, but at least removal is coming in on a score that lets it be in competition with the creatures. Giving a little bit of advantage means that you can play removal spells that fit into your game plan, rather than just needing them to make sure you don’t get killed by a bomb.



Again, this ranked list isn’t meant to be a definitive placement of every set. Instead, it gives a model that can help contextualize where a set is falling in relation to other sets, once you account for all the different indicators. According to this model, Kaladesh comes in as the 10th highest ranked set for removal, and we can see how the most recent sets have seen a marked reversal in the trends for removal. We haven’t seen a set with removal this good since Morningtide, and that’s a little more than 8 years ago.


The strangest thing is Kaladesh is a contemporary outlier for both creatures and for removal. I’ve heard a lot of people talking about Kaladesh and noticing what they call power creep, and it seems like this analysis confirms that concept. The following chart shows a scatter plot of the model’s ranking of each set by both creature ranking and removal ranking:



This set shows that Kaladesh is in a category of its own. The most comparable sets are Shadows over Innistrad, and Eldritch Moon, and those three sets seem to be forming an entirely new category, or a new era of limited design. Who knows if these trends will continue, but if they do we could be on the cusp of having to radically rethink the way we analyze limited sets in the same vein of what we had to do after Shards of Alara ushered in the format style and philosophy that we’ve been drafting for such a long time.



Conclusions and Recommendations


This new format for spoiler analysis takes a much more holistic approach, which means that it’s harder to come to very specific recommendations about the set, but this analysis makes a few very important points. First, Kaladesh has creatures unlike anything we’ve ever seen, and you’ll want to be able to deal with big creatures coming down on efficient curves. This is likely to make the format a bit faster than average. Second, the removal is relatively strong compared to what we’ve seen in recent years. With the creatures being so strong, it’s possible that the removal is still falling behind in comparison, but it seems like this format will be a little different from what we are used to. This seems like a set that is more modular than synergistic, meaning that a lot of the cards can fit into a variety of decks, allowing you to twist, experiment, and invent on the fly to come up with new solutions to different problems.


People always want to know how good each color is, and I haven’t done an in depth analysis of the commons or uncommons of each color individually, but I figure that it can be a little bit helpful to give my first impressions. On twitter, I tend to joke that blue and green always look good to me, but in this setting, I like to look at things more seriously. On first impression, I think that black is the strongest color, though red also looks quite good. Green seems deep and it has huge creatures, as well as a lot of powerful synergies with energy, +1/+1 counters, and even artifacts with fabricate, so I think that green will at least be in contention with red and black. Both blue and white look weak to me, relative to the other colors, so I would be looking to settle in the Jund colors at the beginning.


As always, thanks for sticking through to the end of the article. If you want to find more information from me, you can check out my blog at oraymw.wordpress.com. You can also follow me on twitter @oraymw for updates about when articles go live, as well as daily discussion about Magic and limited. Finally, feel free to leave any comments below, or drop me a line at my blog or on twitter. See you next time!



Ars Arcanum Archive



Small announcement to go with by oraymw at Fri, 09/23/2016 - 20:04
oraymw's picture

Small announcement to go with this. I've made a public folder in which I can dump all of the data for articles like this on Google Drive. I explain this in a teensy bit more detail on my blog, along with a persistent link, so feel free to peruse the data to your heart's content:


I missed you Matt, thanks for by JXClaytor at Sun, 09/25/2016 - 21:25
JXClaytor's picture

I missed you Matt, thanks for the piece!