Friday, July 8, 2011

Fascinating Nids Discussion on Nikephoros' blog

So, Nikephoros has been doing some really interesting work over on his blog. I mean - REALLY interesting work. He is attempting to create a simple and quantifiable system for determining the "competitiveness" of a list.

Even trying to take on a project like that is really laudable. Whether it's truly possible or not is another question.

But, as a good scientist (and it's been clear to me for some time that Nikephoros IS a good scientist), he was not satisfied to simply create the system and call it a day. He started examining lists using his system and looked for correlations between the scores he got and the results those lists were getting in tournament play. I encourage you to browse through his June and July archives for some examples.

And while you're there, follow his blog. It's a shame he hasn't broken 100 followers yet. His blog is quality. Follow it.


Now at some point Nike's efforts inspired Lyracian to do some analysis on some Tyranids lists. One of the lists happens to be my current tournament list. How my list got included in a "Fex Star list" analysis, I'm not sure, but hey I'll roll with it.

Nike followed up on Lyracian's efforts with his own post and a few thoughts.

So first, I wanted to mention that it has become very clear to me that I am considered to be "an authority" on Nids. Some have even become to view Best Overall as a "Nids Blog." I'm not sure that I'm really any great authority on an army that I have not yet taken to a tournament at the 2k level, but apparently if you talk about a subject long enough and thoughtfully enough people start thinking of you in that light.

It's interesting to see some of the track-backs and mentions of my blog out there. Almost all of them are associated with Nids and people pointing others here for Nids advice.

All well and good of course, but with one major caveat. I am learning just right along with everyone else. I don't claim to be an authority, I'm just doing my best to figure it out.

Anyways, looking at the numbers from Lyracian's analysis, I have to admit I was a bit self-conscious. I'm thinking "man, my list sucks compared to those others!"

But then you look a bit closer and see where the numbers are really coming from, and I feel a bit better.

For one thing, I think that the value of the melee damage to vehicles is overestimated. Let's take as an example Hulksmash's list versus my own.

Hulksmash has a "Dead Land Raiders Per Game (DLPG) score of 23.61. This is in contrast to my own DLRPG score of 18.52. On paper, therefore, my ability to kill Land Raiders is significantly inferior to his.

All fine, well and good. But a closer look shows that all of Hulk's AV 14 killing comes in the form of melee.

There are two main problems with this approach. One problem is that close combat relies upon getting your critters over to, well, close combat. I consider the likelihood of Genestealers being able to assault a parking lot to be quite high. On the other hand, the odds of a Carnifex assaulting a Land Raider that don't want to be assaulted are quite low.

As another point, it appears from the analysis performed that the number of penetrating hits is assuming that the vehicle moved at combat speed and so is being hit on a 4+. When you take into account the possibility (likelihood?) that the Land Raiders moved at cruising speed and the raw number for Hulk's list goes from 23 down to a little over 7.

Now keep in mind that this metric is calculating the beating that your army can do to an opponent over five full turns. I can conceive of my T-Fexes shooting at my opponent for five full turns. It happens all the time, actually. But the possibility of Carnifexes being in close combat against armor for five full turns is... low.

So having said all that, I recognize that you can ALWAYS come up with criticisms for a metric which is inherently reductionist. You have to be reductionist in order to be useful. Too much detail and your analysis becomes a big pile of formless goop.

But do the details in this case reveal an underlying flaw in incorporating melee attacks, undiscounted, into the same metrics as the shooting statistics? I certainly think they do. 40k inherently favors shooting for the purposes of killing vehicles. While you certainly CAN kill vehicles in close combat, I think it ought to be recognized that doing so is inherently more difficult to accomplish.

At the end of the day, I stand convinced that Nike's metrics are very useful. And that is all, at the end of the day, we can ever expect from any such analysis. I think that having scores within the ranges he outlines are a NECESSARY but not SUFFICIENT condition for success. Furthermore, I don't think that a list which has a higher score in a given area than another list should be automatically assumed to be better at that task, especially since the system cannot account for ranges, maneuverability, etc. I am certainly not going to go trying to tweak my list to 'boost my scores.' (Whether I tweak it because I found changes through playtesting is another matter).

My natural list building efforts resulted in an army well within the range of what is "normal" for competitive tournament lists. From there, it's a matter of playtesting and strategy. And those are things difficult to accomplish in a spreadsheet.