'Ard Boyz Prelim Results - Some Statistics

I noticed a neat graph over at Danny Internet's Bald and Screaming (he assures me the malaware is gone) and am going to nick it for a moment going to get more data and expand it. I used the same initial source Danny used, the Dakka forums, as well as blogs and friends over in the States. Yes, I cross-checked data/stores to make sure I wasn't counting things twice. Now, as we recall using tournaments as proof of an army doing well has a lot of issues. Messanger of Death has started looking at these for me but suffice to say there are just too many variables one cannot account for. This makes determining the validity of tournament results very difficult. Not only do we not always have all the match-up data but we don't always know exact army lists, terrain (and what armies were matched up on what terrain), dice, player ability, personal factors such as fatigue, etc. TO rulings outside of the ruleset (they happen), number of people using certain armies, not to mention soft scores, battle points and so on.

Now this graph is only based on a certain sample of 'Ard Boyz results and does not have the number of total armies represented. For all we know only seven Tau players could have showed up and all of them placed. Unlikely but a statistical possibility. We can make some assumptions though. For example:

1. Grey Knights very likely had the highest number of participants - flavor of the month and all that. If not the highest certainly one of the highest.
2. 5th edition books are more likely to have higher number of players - fit the new ruleset better, new models, rules, etc.
3. Older books very likely had less participants due to older rules, less models, more expensive armies (i.e. all metal), etc.

These are of course assumptions so let's look at both graphs. 

Kirby's graph!

Danny Internet's graph

I've included both my (118 sample locations) and Danny's (87 sample locations) graphs - his of course looks prettier. So what do we see?

1. Grey Knights have an edge on everyone else. I would put this down to more people playing them, not everyone is used to them and thus could be beaten by the unknown army factor and that they are a good army. Would I say they are the best and worthy of smashing everyone in these rankings? No and I think these factors contribute to their higher placing total.
2. If we assume Grey Knights got several extra placings from being new, there appears to be a clear divide between having a good codex and a not so good codex. From Space Wolves across to Imperial Guard there is no significant difference in terms of placings from one book to the next. This includes, Space Wolves, Dark Eldar, Blood Angels, Space Marines, Imperial Guard.
3. We can further this by saying there appears to be three 'tiers' - good armies, okay armies and crap armies. I would hesitate towards this though as some armies I would consider good, such as Black Templars and Tau, are more likely to be influenced by external factors - i.e. older book so less players, 2500 points is harder for older books, etc. That being said, it appears that Orks, Eldar, Tyranids and Dark Angels are 'middle of the run' armies who probably do well against the worse books but are on the back foot versus the good books.
4. There is comparability between the two graphs. Only Black Templars move up in terms of placings compared to other armies (jumping Tau and CSM) and there does seem to be a group of "better" armies. In my graph with more samples, Space Wolves clears the pack just a little bit more and Orks fall behind in that initial good army bunch. Those 5th edition armies, excluding Tyranids, all seem to do relatively well compared to each other across graphs.
5. Some armies appear to be much worse than others. We know Tau, Black Templars and Immo-spam Witchunters are all pretty good but don't appear to place well in these rankings. Tau suffer at 2500 as we know which is a good explanation for why they don't place highly here and the rest of the low-end result armies you don't actually see often and when played, are often by "old school" types who are more into backstory than gaming (this is a generalisation). We can tentatively conclude though that many of these books suffer against their 5th edition brethren even if they do have some perks.

Now these results are hard to actually extract information from in terms of statistical validity - just look at the list of problems mentioned in the first paragraph. It does hint towards some things though.

1. 5th edition is actually pretty damn good in terms of balance and Games Workshop has done a very good job with their army releases in terms of external balance.
2. Tyranids were the 5th edition flop.
3. 5th edition books > 4th edition books. This we already knew so we have some concurrent validity going on here.
4. Minimal evidence of codex creep. I know people love to say that each new codex is the best and whilst Grey Knights have significantly more placings than the other armies, this result is more than likely from a higher number of participants. If we look past Grey Knights however, we appear to see clumping rather than a linear progression of armies based loosely on release date indicating as in point 1, that the books in 5th edition are quite comparable to each other.

We cannot of course use this data as basis for this but it is concurrent with what a lot of us have been saying for a while - 5th edition and the books they produce are quite balanced. This is at a points level uncommonly seen in normal play (2500 points) with many of the initial factors discussed unaccounted for (specifically number of army entrants and lists) so we need to look at these with conclusions with pretty severe scepticism. I certainly wouldn't suggest anyone use them to say army A or B is worse/better than the other or that Army A or B is good/bad. Rather, they show us a potential level of balance in 5th edition where most 4th edition books (including Orks and Daemons) suffer.

All in all some interesting data with the Semis hopefully offering more. Again though, it's not going to give a definitive answer on anything particularly when we take the 2500 points level, varying missions and all the other factors initially outlined into play.