Update: Some folks over on reddit have pointed out that this analysis doesn’t quite do justice to some factions. The biggest reason for this is that a 0.0 and a 1.0 aren’t fundamentally different in terms of whether or not you’re going to play them in draft, but one hurts a faction’s score significantly more than the other. I’m going to leave this article as it is, but I’ll be posting a follow-up (will be linked here) with some more data that better shows that some factions have too high a density of bad cards.

In the last month, Dire Wolf Digital has released two balance patches for Eternal. In each patch, one of the major objectives was to tone down the dominance of Justice in draft. This was much needed, in my opinion, as it was often correct to force some sort of Justice deck, even if it wasn’t particularly open. You’d often end up splashing, or a full-on three factions, but that was fine because the multi-faction Justice cards (e.g., Awakened Student, Whirling Duo, Rooftop Vigilante, Slay) are powerful enough to make up for a shaky power-base, and Justice as a whole is not as reliant on its tribal synergies as other factions.

So, Justice needed to take a step back. The first round of nerfs didn’t quite get the job done, so DWD hit it again. It’s still a little early to tell if this will be enough to finally end Justice’s reign of terror, but were nerfs ever the right answer in the first place?

I’m not so sure. With the exception of Valkyrie Arcanist, I didn’t feel that any of the Justice cards were under-costed, and certainly not oppressively so. Arcanist needed to get hit with the nerf bat because allowing decks to play too many of them created bad play patterns by which you keep chaining 6/6 flying threats until your opponent finally can’t deal with one. None of the other nerfed cards had that kind of problem.

Will the nerfs succeed in putting Justice back in line with the rest of the factions? Maybe. But will it make the format better? That’s the question I asked myself when I began this analysis.

The Data

For the data used here, I took the draft tier list ratings produced by the folks over at rngeternal.com. You may not agree with all of their ratings. I certainly think some things should move up or down a little. Still, this represents a sampling of multiple high-level drafters that is the best available data set outside of my own biased rankings.

The cards are each rated between 0 and 5, at increments of 0.5. Thus, we expect an average card to have a rating of 2.0 or 2.5. These are the bread-and-butter type cards in draft. The majority of your deck will (hopefully) be between 2.0 and 3.0. If you’re playing any cards at 1.5 or below, you’re not feeling too great about it, usually. That’s not to say you should never play 1.5s. Sometimes you just have to fill in that last slot, or sometimes you’ve really gotten there on whatever synergy the card is asking for, or sometimes you just really need something to do before turn 4.

Ratings of 4.5 and 5.0 are generally reserved for absurd bombs like Mystic Ascendant or Ijin, Imperial Armorer. Cards of lower rarities should never be 4.5s or 5.0s because that would create a very skewed draft format toward whatever faction held that 5.0 uncommon. Indeed, no common or uncommon holds a rating higher than 4.0 in the RNGEternal spreadsheet.

The first thing I did was to hack and slash that wonderful spreadsheet, cutting out all the rares and legendaries. These just don’t make a huge difference to the health of a draft format because they show up so infrequently.

I then sorted the cards by set and faction. I plotted the number of cards at each rating, for each faction, at both common and uncommon. Multifaction cards are grouped all together in the plots, but I consider each faction pair individually in my further analysis, which includes variance and total strength of individual factions in an average pack of each set.

The Ideal Format

What does an ideal spread of ratings look like for a draft format? Or, rather, what spread do we expect a good draft format to produce? For commons, this is a fairly simple question to answer, since we expect most of them to be between 1.5 and 2.5, with a few above and a few below. Commons should almost never be 4.0s, and only a few per set should even go to 3.5 because we don’t want them to overshadow uncommons in terms of power level, and we don’t want too many of a single, powerful card showing up to warp the format. We also don’t want too many low on the scale, since having too many bad cards showing up frequently will lead to people failing to scrape together playable decks at too high a rate. A regular bell-curve type distribution entered at 2.0  is probably the “ideal,” but it obviously isn’t the only criteria for a good format.

At uncommon, things are a little less clear. The high-value uncommons should obviously be higher on the rating scale than the best commons. But the uncommon rarity also tends to introduce more complex, “out there” cards, as well as more build-around type cards. If you’re drafting competitively, and not just for the memes, you probably want to shun all-in build-around strategies because they’re inconsistent. The “out there” cards tend to be over-costed because the developers don’t want them accidentally breaking the format. Thus, at uncommon, you not only tend to have more 3.5s and 4.0s, but you also have more unplayable cards at 1.0 and below. So, we expect more of a double-peaked distribution, with maxima near 3.5 and 1.0.

I haven’t analyzed rares and legendaries here (perhaps later), but I would expect those distributions to be even more skewed.

Set 1 – The Empty Throne

I thought Set 1 draft was a great format. I always felt like I could draft whichever supported faction pair was open, and I never really felt as if one was better than the other. Combrei did get the one single 4.0-rated common in Awakened Student, but Student never felt completely oppressive, thanks to a variety of efficient answers. Playable count was rarely a problem, unless I really waffled on my factions deep into pack 2. Do the numbers bear out that assessment?

Set 1 starts us off strong. The average value of a common is 1.996, which is pretty much spot on in terms of what we’d expect. The distribution is somewhat skewed, with a large spike at 3.0, meaning there are more premium commons, but also more fringe playables, and fewer average commons. Fire appears to be weaker than the rest, with a wide spread in power level, while Primal seems to be the most evenly-distributed. Now let’s have a look at the uncommons.

Once again, Set 1 has an absolutely beautiful distribution, with the expected peaks at 3.5 and 1.0. The average uncommon rating was 2.31, a reasonable amount higher than the commons. The factions appear to be fairly balanced, with the exception of Primal, which has fewer weak cards overall, and thus will have a higher average.

Let’s look at some more detailed numbers. The table below shows both the averages and standard deviation of the commons and uncommons, along with the expected average rating of a card in that faction that you’ll see in a pack. For those who aren’t familiar with statistics, the standard deviation is a measure of the variance in a set of data. Larger standard deviations mean that the data is more spread out; smaller standard deviations mean that the data is more clustered.

FactionAvg. Com.StD. Com.Avg. Unc.StD. Unc.
Fire1.751.132.320.97
Time1.951.112.111.36
Justice1.880.922.091.14
Primal2.030.902.611.16
Shadow1.901.002.331.09
Multi-Faction2.400.742.601.01
Neutral2.391.021.33--

 

These numbers show that Fire has the weakest commons, while Primal has the strongest by a margin of 0.08. Worth noting is the larger standard deviation of the Fire commons, where we have a couple of really premium options like Torch, along with a lot of really bad ones like Ruin.

Primal also wins at uncommon, by quite a large margin. Looking back at the distribution once more, this isn’t due to any excessively powerful uncommons. Rather, Primal just has fewer unplayable ones. As long as these uncommons aren’t oppressive–and they aren’t–having a high average score at uncommon won’t break the format.

Of course, multi-faction cards have a large impact on how strong or weak single factions are as well. For the following table, I’ve assumed that we know we are drafting each two-faction pair. With that in mind, I combine the cards from both of the factions in that pair, along with the relevant multi-faction cards to compute what kind of power level we’re likely to see, depending on our faction pair.

Faction PairAvg. Com.StD. Com.Avg. Unc.StD. Unc.
Combrei1.951.062.171.27
Elysian2.011.082.421.22
Rakano1.861.072.201.03
Stonescar1.851.142.311.05
Feln2.001.072.511.11
Praxis1.851.182.221.19
Hooru1.951.002.351.13
Argenport1.891.012.211.14
Skycrag1.891.182.471.16
Xenan1.931.092.211.26

 

These numbers bear out my experience with the format. The spread in all factions is only 0.16 at common and 0.34 at uncommon. Even the pairs that weren’t directly supported, receiving no multi-faction cards, are decently strong. The median lies at 1.91 at common and 2.27 at uncommon. Keep these numbers in mind as we compare the other sets to Set 1.

Set 2 – Omens of the Past

I enjoyed Omens draft, but I didn’t find it quite as balanced as Set 1. I almost never drafted Praxis. I did draft Skycrag sometimes, but I never had much success with it. Usually I stuck with Argenport, Xenan, or Hooru. The format was much lower-power than Set 1, both in terms of the raw power of each card, and because Set 2 rendered moot the powerful multi-faction cards found in Set 1. There were far more payoffs for drafting a critical mass of cards from a Set 2 faction, as well as a higher density of multi-faction cards at Common, so you usually found yourself heavily committed to one of the Set 2 faction pairs. This reduced the value of the multi-faction cards from Set 1, because you would often have to splash if you wanted to play them, and things like Awakened Student have much less impact when you can’t consistently land them on turn 2.

The ratings here for Set 2 have been taken in the context of Set 3, so many of the synergistic cards have been lowered in value from where they once were. These numbers aren’t reflective of the health of the Set 2 format, only Set 3. Let’s have a look…

The average rating for commons is 1.88, which is significantly below that of Set 1. This distribution is also less pretty than that of Set 1. Not only is it misshapen, but the factions are also poorly spread.  Primal, Shadow, and Fire, have zero mono-faction commons that rated at 2.5. Shadow only has two commons above that; the rest are below-average cards. Time’s biggest concentrations are at 3.0 and 1.0. Fire has nearly as many cards at 0.5 or 0.0 as it does everywhere else combined. Justice is the only faction with a nearly-even distribution between 1.0 and 3.0.

The average uncommon rating is 1.92, which is well below Set 1. The distribution looks a little more like the two-peak distribution we’d expect. The higher peak is down at 3 instead of 3.5. Once again, Fire is in the doghouse, with only three uncommons at 1.5 or better.

FactionAvg. Com.StD. Com.Avg. Unc.StD. Unc.
Fire1.461.191.130.58
Time1.821.172.291.09
Justice2.040.642.131.04
Primal1.860.911.421.15
Shadow1.460.901.421.11
Multi-Faction2.130.932.650.98
Neutral3.000.00----

 

Yowza. Justice is top dog at common by a long shot, with Time and Primal being tier two. Fire and Shadow are both atrocious. Remember how I mentioned that you’re really unhappy to be playing 1.5s and below? Well, the average Fire and Shadow common is nigh-unplayable. Same goes for Primal and Shadow uncommons. Shadow and Fire are complete and utter trash. Primal has decent commons, but very little punch at the uncommon slot.  Time does pretty well, with decent commons and the strongest uncommons. Justice is the clear winner, however, with above-average commons and the second-strongest uncommons. When it comes to drafting, strength at common is often the most important thing at determining which faction is dominant, and Justice runs away with it here.

Not only does Justice have the strongest average, but it also has a very small standard deviation. This means that the ratings are tightly clustered around the average of 2.04, resulting in fewer stone-unplayable cards. When you see a Justice card, most of the time it will be a 1.5 or above, meaning that, at the end of the draft, if you’re in Justice, you’re simply more likely to have more playables than a player who isn’t Justice. Somehow Fire’s average uncommon is 30% weaker than its average common .

The average multi-faction card is much stronger than the mono-faction cards (as they should be). Are they strong enough to pick up the slack in the other factions?

FactionAvg. Com.StD. Com.Avg. Unc.StD. Unc.
Combrei1.930.992.211.07
Elysian1.841.051.851.19
Rakano1.751.001.630.98
Stonescar1.461.071.270.92
Feln1.660.961.421.13
Praxis1.731.181.661.13
Hooru1.950.781.961.16
Argenport1.810.882.001.22
Skycrag1.771.161.361.01
Xenan1.631.122.041.21

 

The short answer is no. The Fire and Shadow factions place much lower than the factions that contain neither. Argenport and Rakano are somewhat carried by their Justice half, as Xenan and Praxis are somewhat carried by Time. The latter two are hampered by the fact that they have weak multi-faction options. Praxis’s were never too good to begin with, with such hits as Brilliant Discovery and Sandglass Sentinel. Xenan’s multi-faction cards used to be fantastic, but the complete lack of lifegain support in Set 3 has knocked them down several pegs.

The median rating is 1.76 at common and 1.76 again at uncommon. This is well below the median power level of the factions in Set 1, particularly at uncommon.

Set 3 – The Dusk Road

Much maligned for relying too much on synergies that don’t come together, Dusk Road has been the primary target of the nerfs. Certainly, Justice is by far my most drafted faction, followed by Shadow and then Fire (usually because of Whirling Duo).

Set 3 has a pretty reasonably-shaped distribution. The only problem is that it peaks a little too far to the right. The average rating for a common is 1.87, on par with Set 2’s low rating. Set 3’s distribution is better-shaped, but it really lacks premium commons. Set 2, which is of a similar size, had 28 commons at 3.0 or above. Set 3 has only 12, none of which are in Primal.

Uncommon is quite a bit different from the previous two sets. The distribution is pretty much flat between 1.0 and 3.0, which isn’t really what we want from our uncommons. There are about a third fewer cards at 3.5+ when compared to Set 2, which itself was considerably weaker than Set 1. Furthermore, two-thirds of those powerful uncommons are multi-faction, compared to about 40% in Set 2. That makes it even more unlikely that you will be able to first-pick a desirable uncommon out of your Set 3 pack, since taking multi-faction cards first overall is a major risk that is usually best avoided in favor of a slightly weaker, but mono-faction card.

Speaking of mono-faction cards, let’s have a look at some numbers.

FactionAvg. Com.StD. Com.Avg. Unc.StD. Unc.
Fire1.940.752.080.78
Time1.880.821.961.15
Justice2.190.771.570.84
Primal1.560.752.000.84
Shdow1.780.711.810.90
Multi-Faction----3.100.80
Neutral----1.0--

 

How do the faction pairs stack up? Justice is once again the clear winner at common, but it’s much weaker at uncommon. Fire is somewhat redeemed here, boasting strong averages at both common and uncommon. Time is a clear third, followed by Shadow, and then by Primal. Though Primal has relatively strong uncommons, it’s so much weaker at common than the other factions that you won’t often see Primal cards that keep you invested, even if you do first pick a solid card.

Faction PairAvg. Com.StD. Com.Avg. Unc.StD. Unc.
Combrei2.030.831.841.09
Elysian1.720.852.001.01
Rakano2.060.791.880.89
Stonescar1.860.742.000.85
Feln1.670.781.940.86
Praxis1.900.792.000.97
Hooru1.880.821.790.87
Argenport1.980.801.770.93
Skycrag1.750.892.200.97
Xenan1.830.781.941.03

 

At common, Justice takes four of the top five. That’s it, case closed, Justice is OP!

Is it, though? Look at the actual scores. Only Rakano and Combrei exceed the 2.0 threshold, and not by too big a margin. Several factions, most notably those containing Primal, are well below 2.0.

At uncommon, the Justice pairs are actually the lowest. Primal factions get a bit of a bump here, but because of the inconsistency of seeing any given uncommon, it’s not enough to help them overcome their major deficit at common. The power level of the uncommons as a whole is much weaker than that of Set 1.

The median ratings are 1.87 at common and 1.94 at uncommon. The common power level is close to that of Set 1, but uncommon is still well below the bar.

Bringing it All Together

In order to get a clear picture of how this all shakes out, let’s look at each set’s average scores side-by-side, starting at common. Since we get two packs of Set 3, I’ve included a weighted average that counts Set 3 twice.

FactionSet 1Set 2 Set 3Weighted Avg.
Fire1.751.461.941.77
Time1.951.821.881.88
Justice1.882.042.192.08
Primal2.031.861.561.75
Shadow1.901.461.781.73

 

In Sets 2 and 3, which make up 3/4 of the draft, Justice is the strongest faction, but it isn’t exceeding the 2.0 mark by that much. The more glaring departure is in how low some of the other factions are. Particularly in Set 2: Fire and Primal miss the 2.0 mark by 0.54! That is a massive margin. In Set 3, Primal misses by 0.44, and Shadow by 0.22. Both of those margins are greater than the margin that Justice is exceeding 2.0.

Can the multi-faction cards make up these deficits?

FactionSet 1Set 2Set 3Weighted Avg.
Combrei1.951.932.031.99
Elysian2.011.841.711.82
Rakano1.861.752.061.93
Stonescar1.851.461.861.76
Feln2.001.661.671.75
Praxis1.851.731.911.85
Hooru1.951.951.881.92
Argenport1.891.811.981.92
Skycrag1.891.771.751.79
Xenan1.931.631.831.81

 

This reveals what I believe to be the crux of the problem. The Justice pairs are the only ones that don’t have major drop-offs at common after Set 1. Heck, Set 2 is supposed to be THE set for Praxis, Skycrag, and Xenan, while Set 1 isn’t supposed to support those combinations. Yet, the average scores at common for those three factions take a major dip in Set 2.

Set 3, on the surface, seems as if it is supposed to support all factions. However, in my experience, Hooru, Feln, and Xenan lack any sort of identity in Set 3. Hooru seems to want to be “things with two battle skills,” but all of the payoffs are at uncommon or higher, and the cards aren’t playable on their own if you don’t get there (a 1-cost 1/1 and a 4-cost 1/5, for example).  Xenan is…I don’t even know. Dinosaurs? Nightfall? Dinos in the Dark? Feln wants to be the Unseen curse deck, but none of the payoffs are worth playing some of the terrible curses in the set. As a result, the Primal factions, Feln in particular, wind up with a major deficit in Set 3. Elysian’s score is higher in Set 2, a set which did not support that combination at all.

Remember the median scores from Set 1? At common, the median faction pair score was 1.91. Only the Justice faction pairs meet that threshold, and only Combrei really exceeds it by a decent margin.

Every non-Justice pair falls short of the mark. Praxis gets the closest, but Xenan, Skycrag, Feln, and Stonescar all miss by a larger margin than Combrei exceeds.

At uncommon, Justice is actually one of the weakest factions. However, the overall power level is quite a bit lower than we might like:

FactionSet 1Set 2 Set 3Weighted Avg.
Fire2.311.132.081.90
Time2.112.291.962.08
Justice2.092.131.571.84
Primal2.611.422.002.01
Shadow2.331.421.811.84
FactionSet 1 Set 2Set 3Weighted Avg.
Combrei2.172.211.842.01
Elysian2.421.852.002.06
Rakano2.201.631.881.90
Stonescar2.311.272.001.90
Feln2.511.421.941.95
Praxis2.221.662.001.97
Hooru2.351.961.791.97
Argenport2.212.001.771.94
Skycrag2.471.362.202.06
Xenan2.212.041.942.03

 

The median average sits at 1.97. This means that the average uncommon across the three sets is weaker than the average common should be. Time has the best uncommons at 2.08…but that is the same average as Justice gets at common! Of course Justice is going to dominate the format when its commons are just as good as the top uncommons of all factions. Again, though, this is a symptom of non-Justice factions being too weak rather than Justice itself being too strong.

Conclusions

While Justice is clearly the strongest single faction across the draft format, this analysis does not find that it is excessively strong. Rather, it finds that other factions are excessively weak, particularly at uncommon. Nerfing Justice cards will bring balance to the factions, but it will not create a more fun drafting environment. Consider two matchups, one where both players have a deck which averages a 3.0, and one where both players have a deck which averages a 1.8. In the first match, both players get to do powerful, exciting things. Even if you lose, you say to yourself, my deck was great, but my opponent’s was better. What a game! If you win, you say to yourself, awesome! I beat my opponent even though they had a really good deck!

Now consider the other case. Even if you win, you are left with the feeling that you got lucky that your dumpster fire burned a little hotter than your opponent’s dumpster fire. And if you lose, you’re left with the feeling that your deck is complete trash and you suck at drafting. Your opponent’s deck was terrible, but it still beat you.

Is that the kind of experience we want to have, especially for new players? I consider myself a skilled drafter, able to read signals and have the discipline to switch gears and abandon my first picks, and I still find myself scrambling to put together an actual deck quite often. I’ve played some real stinkers as my last few cards. It must be even worse for those who aren’t as serious about drafting.

From a design standpoint, it actually makes sense to overshoot the power level a little bit. Rather than 2.0, aim for, say, 2.2. High enough that players can build consistent, fun decks without crossing their fingers and hoping they get to 28 actual playable cards.  As long as the various factions stay balanced, that is a fine place to be. New players won’t find themselves playing complete garbage cards, and skilled drafters will be able to really put together some sweet decks.

With that in mind, Justice is fine where it is. What we need is to elevate the other factions to that level, not lower Justice to theirs. The few buffs we’ve received thus far are a good start. We need more of that. In particular, Primal, Fire, and Shadow all need significant help in Sets 2 and 3.

At the end of the day, these numbers don’t necessarily tell the whole story. After all, you need to consider power curves and synergies as well. There will be times when you take a 2.5-rated 2-drop over the 4.0-rated Pillar of Amar because you already have a few 6+ drops and really need a cheap play. Plus, it doesn’t matter if a faction is stacked with 3.5-level commons if those commons all cost 6+ power, as you won’t be able to play too many, even if you can draft them. I’ll be taking a deeper look into those things with my next little project.

Until then, for all the complaining I’ve been doing here, I’ll be drafting because that’s still the best thing to do in any card game, even if I believe the format could be improved.