Requiem for an Edition Part 3 of 4

performance

Eric Lofgren from Variance Hammer is back with part 3 of his article series! Check the Tactics Corner for more great articles!

And here we have the big one – an army by army breakdown of the performance of the LVO entrants. Who did well, who…did not so well?

First up, a look at some specific army combinations, to see if some very similar armies have markedly different performance.

Inquisitorial Allies:

The person who won the LVO won it with an Imperial Guard and Inquisition list Note: the LVO was actually won by an Eldar/Dark Eldar Seer Council/Wave Serpent List-Ed. So there’s a question – does the flexibility of the Inquisitorial army, with its tiny squads, some nice transports, cheap access to Psykers and special characters etc. put you at an advantage over non-Inquisitorial armies?

The median score of the 19 armies with Inquisitorial allies was 2028, while the median score for the 174 armies without the aid of the Inquisition was 2518.5. This difference though wasn’t statistically significant (p = 0.88), and the visualization of the scores bears that out – the distributions are pretty similar, there’s no clear “cliff” that either army doesn’t seem to get over (as there was for unallied armies generally), and both types of armies were also at the bottom.

Chaos Armies:

So it’s been established that straight on Chaos Space Marines are somewhat out of favor with players, but does this bear out in performance as well?

The median score for a CSM army was 2021, for a Chaos Daemons army 2017 and for a combined CSM/Daemons allied list, 2516. That seems like a pretty hefty advantage for the combined powers of the Chaos army lists. It is however, not a significant one (p = 0.49). Partially, I suspect this is because of the small number of CSM armies, and partially, if you look at the plots, looking at the median is a flawed approach:

If you’re an average player, what kind of Chaos army you play doesn’t matter. Break out your pure Chaos Space Marines lists! But there’s a pretty clear cliff, again, where neither pure Daemon nor pure Chaos Space Marine armies climb. Above about 4000 total score, there doesn’t seem to be much hope – if you want to win, again a combined list is probably the best bet, presumably because of synergies between the two lists.

Eldar:

How do the Craftworld Eldar perform versus their dark kin? Is it better just to ally the two, for hilarious Farseer and Venom hijinks (my personal, entirely non-competitive answer: Yes)?

The median Eldar army score was 3024, the median Dark Eldar army 2273, and the median allied army was 3532. Like the Chaos armies, thats a pretty hefty step, and while it’s still not statistically significant (p = 0.11) it’s getting close. Looking at the plot, there’s some clear performance breaks:

There’s a bit of a floor for the Eldar, beneath which no Eldar list fell – and it’s about 1000 points higher than the Dark Eldar. Similarly, the Dark Eldar’s best performers hit their cliff at about the same range as middling-fair performing Eldar armies. And both are pretty clearly dominated by Eldar/Dark Eldar combined armies, who have both a higher median, a higher minimum, and a higher maximum (notably, 3rd and 4th place in the tournament).

Whether or not this is a book-level synergy, or just a manifestation of the Seer Star lists performing very well is something I can’t answer without access to the full army lists, but it’s clear the Eldar book is performing well, and the Dark Eldar enhance it, but on their own, the Dark Eldar, with a very out of date book, are struggling.

Tau:

The last breakdown of small groups of related armies we have is the Tau: comparing Tau lists (either with no allies, or unconventional allies like IG or SM), the feared Taudar lists, or the Tau/Tau double lists I mentioned in a previous post.

The median score for a straight-Tau list was 3019.6, the median score for the Taudar 3532, and the median score for the double-Tau lists was 2771. As you can probably guess by now, this difference is not significant (p = 0.55). Visualizing the distribution of scores between the three armies paints an interesting picture:

First, I’m not sold on the utility of using two Tau detachments. There’s a pretty serious floor for both of them, and both armies also perform within the top tiers. The centers of the distribution are pretty much the same. Basically, while it feels like list building shenanigans, it’s not manifesting as a clear tournament advantage.

On the other hand, the Taudar lists…while rare, are rightly feared. The widest part of the distribution is pretty high, and the floor under which no Taudar lists performed was quite high. It’s a reliable high performing list. Now that the Tau and Eldar aren’t Battle Brothers however, it will be interesting to see if this remains true. It should also be noted however that, while a solid performer, there are a number of armies that had higher tournament performances.

5th vs. 6th Edition Armies:

How are the older codexes doing? The Dark Eldar, Necrons, Grey Knights, Blood Angels, Space Wolves, Orks and Imperial Guard were all using 5th Edition codexes when the LVO took place. How did these armies do in comparison to those with modern, 6th Edition codexes?

Obsolete, 5th Edition codexes had a median score of 2522, and 6th Edition codexes had a median score of 2514.5. Again, this difference wasn’t statistically significant (p = 0.55). Looking at the scores visually, we see some interesting things – the “hump” of the Fifth Edition distribution is higher, but there’s a wider base – there are still well performing 5th Edition books, and some poorly performing 5th Edition books (as we’ll see, Orks and Space Wolves), but there is nothing inherently tournament losing about having a codex an edition out of date – a well written 5th Edition codex will beat a poorly written 6th Edition one.

Overall:

Lets have a look at the overall picture. This is the performance of all armies, usually by their primary detachment (except the few I’ve broken out here):

  • Blood Angels (BA): 2017
  • Combined Chaos (CC): 2516
  • Chaos Space Marines (CSM): 2021
  • Dark Angels (DA): 2019
  • Chaos Daemons (CD): 2017
  • Dark Eldar (DE): 2273.5
  • Double-Tau (DT): 2771
  • Eldar (E): 3024
  • Eldar/Dark Eldar (E/DE): 3532.5
  • Grey Knights (GK): 2017
  • Imperial Guard (IG): 2770
  • Necrons (NE): 3022.5
  • Orks (ORK): 1020
  • Space Marines (SM): 2023
  • Sisters of Battle (SoB): 2028
  • Space Wolves (SW): 1266
  • Tau (Tau): 3019.5
  • Taudar (Tau/E): 3532
  • Tyranids (NID): 3022
  • Overall Median: 2516

So which army reigns supreme? By medians, the alliance of Eldar and Dark Eldar barely beats out Taudar as the highest performing lists. Interestingly, neither one of these armies won the tournament Note: the LVO was actually won by an Eldar/Dark Eldar Seer Council/Wave Serpent List-Ed. They’re just very solid performers. Other take-away messages: Ork players should be super excited about their new codex, and based on the performances in this tournament, the Space Wolves could use some improving as well in terms of a primary detachment. I weep a little bit for my Dark Angel and Sister of Battle armies as well – though in terms of a median performance, both are actually fairly close to the Space Marines. All of these comparisons are (you guessed it!) not statistically significant (p = 0.07), but once again, it’s fairly close, and see the caveat in my first post about how underpowered this all likely is.

Lets look at the picture:

This shows the distribution of every army, with the red dashed line indicating the overall median score in the tournament. Interestingly, only the Space Wolves and Dark Angels had entirely below average performance within the tournament. The Space Wolves especially surprised me, as people still look at them as an “Old, but still good” army list, when at least for the LVO, this is actively not true. On the other end, the standout armies were the combined Daemons/CSM armies, Eldar/Dark Eldar armies, the Imperial Guard, the Necrons and the Space Marines. The Tau, despite being “OP”, were solid performers, but not particularly winning armies.

Next up: Closing Thoughts

Tags:

About Reecius

The fearless leader of the intrepid group of gamers gone retailers at Frontline Gaming!

10 Responses to “Requiem for an Edition Part 3 of 4”

  1. Brent June 23, 2014 12:30 am
    #

    Eric, this continues to be a fascinating series; I’m looking forward to your closing thoughts.

  2. dvs1 June 23, 2014 12:51 am
    #

    Does it matter that the guard/IG player actually lost to a DEldar beaststar in the quarter finals? I believe Alex Fennel won it with his own pink and flowery DEer Council.

  3. fluger June 23, 2014 7:27 am
    #

    Just a quick thought on this fascinating statistical discussion.

    The top end players will naturally gravitate towards books that they perceive to be powerful (often rightly), and it seems like older books tend to get stale as well for players seeking new tactical challenges. I wonder if there’s just a natural migration by the better players away from older books because of these things and it tends to make certain armies fare poorer.

    I wonder what the stats would be like if solid players got behind the wheel of some of the older dexes.

    Anyway, great series.

    • Gordy June 23, 2014 9:52 am
      #

      There are a lot of top players who stick with certain armies, though. Blackmoor pretty much plays GKs, GTA plays Eldar, etc. Some players bandwgoning shouldn’t skew the results nearly as much as everyone bandwagoning, since the older armies still have some skilled representation.

    • AbusePuppy June 24, 2014 4:22 am
      #

      I think it depends on the codex- some codices, like GK, ‘Crons, and IG, still have legs under them and can perform quite well when piloted by a good general. Others, like BA, seem to be something of a lost cause- it doesn’t matter how good your general is if the codex itself doesn’t have the right tools.

  4. Dash2021 June 23, 2014 8:28 am
    #

    Fantastic discussion. Always a fan of statistical break downs. Though your sample sizes are to small for significance, you are getting some pretty strong trends.

    As taudar is effectively dead it will be interesting to see the representation of deldar in the future. The two books complement each other fantastically, and very little about that dynamic changed. I stopped playing jet council b/c it was broken and neither my opponents or I were having fun, and I’m glad I did. There are a lot of very strong options when combining these two dexs, and I’m going to guess that it will continue to trend higher and higher in tournies.

  5. Nuln-Oil June 23, 2014 9:23 am
    #

    Makes sense and it confirms what we already knew.

  6. Eric June 23, 2014 9:26 am
    #

    “Does it matter that the guard/IG player actually lost to a DEldar beaststar in the quarter finals? I believe Alex Fennel won it with his own pink and flowery DEer Council.”

    Not particularly? The way the plots end up smoothing things out somewhat damp out the impact of one game result.

    “The top end players will naturally gravitate towards books that they perceive to be powerful (often rightly), and it seems like older books tend to get stale as well for players seeking new tactical challenges. I wonder if there’s just a natural migration by the better players away from older books because of these things and it tends to make certain armies fare poorer.”

    This is possibly an explanation. It would be *very* hard to parse this out, outside having a tournament where what army you play is randomized.

    “As taudar is effectively dead it will be interesting to see the representation of deldar in the future. The two books complement each other fantastically, and very little about that dynamic changed. I stopped playing jet council b/c it was broken and neither my opponents or I were having fun, and I’m glad I did. There are a lot of very strong options when combining these two dexs, and I’m going to guess that it will continue to trend higher and higher in tournies.”

    It will indeed be interesting to see.

  7. Reecius
    Reecius June 23, 2014 11:19 am
    #

    I have been really enjoying these article, thanks Eric! SO interesting to see a breakdown of actual performance vs. our prejudices and conceptions. There’s a million other variables that go into this of course, that are difficult, if even possible to measure, but this helps out a ton.

    I hope you continue to do more of these as they are a lot of fun to mull over. Like baseball cards!

  8. nkump June 23, 2014 12:32 pm
    #

    These are really entertaining to read! Feel very even-handed in how you present the data…which is quite the feat, especially on the internet! Thanks again!