Forum adverts like this one are shown to any user who is not logged in. Join us by filling out a tiny 3 field form and you will get your own, free, dakka user account which gives a good range of benefits to you: No adverts like this in the forums anymore. Times and dates in your local timezone. Full tracking of what you have read so you can skip to your first unread post, easily see what has changed since you last logged in, and easily see what is new at a glance. Email notifications for threads you want to watch closely. Being a part of the oldest wargaming community on the net. If you are already a member then feel free to login now.

This series of posts is meant to be a serious philosophical and mathematical look at the way in which players discuss and think about playing the game of 40k in a competitive fashion. This is not meant for tournament advice, but to examine in ideal the ideologies and strategic thought paths that lead us to decisions of the power of units, unit combinations, and army lists, as well as our strategic and tactical choices in the game itself. The idea is to develop a meta-language to better examine the game of 40k in the purest form, and as such to come to a better understanding of the game itself. This is a lofty goal to be sure, but it is one that seems long overdue for warhammer.

I would like to ask all posters in this thread to please consider that this is meant to be of great seriousness, and to please make sure that your thoughts are of sufficient intellectual rigor before posting. That is not to say that humor is not welcome, but please incorporate it with relevant posts.

For the first topic, I would like to bring up an idea that has been bouncing around my head the last few months, after seeing Therions Mech Eldar list. The idea is Platforms.

Simply put, the platform is an examination of units not as the unit itself, but as a collection of statistics arrayed and thought of as a singular platform. Most players already so this. The Las/Plas marine squad is purely a platform, examined for its 3 wounds before non-scoring (the fact that it is 3 marines lost is much less important than it being 3 T4 3+ wounds) and the platforms ability to bring competent firepower to bear against a number of targets that are considered to within the role of other units (tanks/Meqs, which many codex seem to suggest are the territory of Elites/HS).

The platform concept is meant to take this examination to the logical (or perhaps absurd) extreme, which is to strip all guyness (see funny words) from the unit, and examine it purely as a singular blob, with a certain area is takes up. Shooting is examined via the relevant shots it emits, and the statistical chance of these shots effecting target units, instead of looking at the maximal shot output. This is not to say we disregard any shots, but we look at weaponry output in a role by role and range by range basis. We also proceed to examine the number of wounds to reduce the units effectiveness, either to remove scoring, or remove a relevant weapon. With regards to these wounds, we also want to examine the difficulty in removing them. A guard squad has 5 wounds before its effectiveness is threatened, but these 5 wounds are not as valuable by far as the two wounds of a Carnifex. Additionally, the movement value of said platform is important. Few platforms deviate from the norms of don't move unless forced, or 6" a turn, and those that do should be considered important exceptions.

Importantly for the concept of platforms is that we are looking at average performances. When we examine whether a specific platform is good as a whole, or at its role, we are not looking at maximums or minimums, but at the average performance that will be expected. To look at anything else is to remove any sense of objectivity from the examination, as certain platforms are exceedingly good in the hands of lucky players. This is not to say we should ignore that such maximums and mins to occur, entirely the opposite, we should put properly proportioned mindfulness into this. Such mindfulness is why we inherently find most rending platforms to be good, regardless of many platform supporting statistics, as the outlier outcomes of rending are so favorable.

Please be aware that this is a very rough outline of the concept, and is very ready to be developed. Development of such language should go hand in hand with development of a better understanding of the game itself.




Finally, to give an example of what concept I am attempting to convey, I will give a quick rundown of the platform that so intrigued me.

3 Eldar Guardian Jetbikes, one with Shurican Cannon + Warlock on Jetbike with Singing Spear and Destructor. This totals to under 150 points. This squad has nearly unsurpassable mobility, as it can move 12+ inches every turn. Even more importantly, it has the JSJ ability, which allows it to break normal turn order of movement (breaking rules is seemingly immensely important in 40k, a topic for late), as well as allowing it a further 6" every turn. It has reasonable durability, as it can survive 2 wounds at MeQ toughness. It also has the additional durability that very high movement brings, of avoidance of shots (100% damage resistance to shots that don't get fired). Most relevantly, as a platform it can reasonably engage any target with a weapon or weapons that can have significant impact. Against hordes it has a heavy flamer, the cannon and the TL cats. Against Medium and light vehicles, the cannon and spear. The Spear can even threaten heavy vehicles, as well as the unit utilizing it mobility for side or rear shots. Against MeQ it has a heavy flamer, and the ability to roll a large number of dice. Its worst role is against monstrous creatures, who have the toughness and armour to survive the cannon/SS and are singular and thus unaffected by the HF. However, it should be viewed that as a singular cheap platform, such a deficiency isn't bad.

Return on investment (ROI) is a good term for this discussion then. The concept being that for any given role that you assign a unit, it has an effectiveness per point at performing that task. You can then compare this effectiveness to other units given the same task.

The common way to do this is first find the chance that your unit will destroy its target in a given turn, then multiply by the cost of the target, multiply by the number of turns you anticipate your unit will devote to attacking, and finally divide by the cost of your unit.

An easy example is the Imperial guard autocannon team vs. the lascannon team attacking a Falcon. The guard player wants to know if the easier penetration of Str9 outweighs the extra shots of the autocannons vs. this target.

Autocannon team: 6 shots * .5 hit * .33 glance * .11 chance to down a falcon = .1089 chance to destroy a falcon per turn.
Lascannon team: 3 shots * .5 hit *.66 glance * .11 chance to down a falcon = .1089 chance to destroy a falcon per turn.

Equal! When you run into something like this, you can be sure the designers had this in mind when designing these units. However, lets add cost into the equation.

Autocannon .1089 success * 6 turns expected attack * 210 cost of kitted falcon / 95 cost of unit = 1.44 ROI. Meaning that if an autocannon team fired only at kitted falcons for 6 turns, it can be expected to earn 1.44 times its own worth in falcons destroyed (1.44 * 95 = 136.8 points of falcon, or about half a falcon)

Lascannon .1089 success * 6 turns expected attack * 210 cost of kitted falcon / 110 cost of unit = 1.27 ROI.

Thus the autocannon is clearly the superior choice when focusing on the dreaded falcon. The decision of what to bring has just started though, as lascannons are good at many things that autocannons are not and vice versa. It's up to the list builder to weigh these options with ROI in mind at all times when making these decisions.

I love this type of dicussion, If only my brain was able to crunch out such things.

... Plain and simple, when a person says "lucky".. sometimes they are right. I try in almost any list to eliminate the real issues of chance, (if a am shooting, I shoot alot, if I am taking saves, I am taking very few)I never really know the exact numbers of times I should have actually taken out X or Y but it seems this is where the thread is going. I like it.

Statistical Fallacies:
Unfortunatly people don't entrensically understand statistics. Because of this you see people make subjective decisions about how the feel a unit should perform. My first observation is the luck factor in 40k.

Statistical Clumping: We've all had that moment of glory or utter dissappoint. We've seen people roll 8 1's for their terminator saves of watched a sanction psycher force weapon a lord of change to death. We first turn to mystisicm. We blame god budda or satan for not

living up to their end of an agrement. I know people who throw dice a way that underperform expectations. So if luck isn't responsible , what is? The term statistical clumping explains these phenomena. Basically clumping is the idea that random results wil clump together but even out over time. A good example is the coin flip. A statistics professor I had gave my class a challenge. She asked half of to make up 100 coin flips and right down the results. She asked the other group to flip a coin 100 time and record the results. She left the room while we fliped coins and coind flips. When she returned we gave her our results and she immediatly told figured out which one was fake. because in this case there is a 50% chance of either result the fake flips looked like this:

U D U U D D U D U D D U D D U U D D

While the real results appeared clumped

U U U U U U U U U D D U D U U U U U D D D D D U D U D

Over a long enough time line the clumping would of evened out. This is essentialy the luck of 40k. I've seen people with solid lists and good tactics lose just because they couldn't wound or hit anything. This brings me to my next fallacy.

Anecdotal Evidence.
I have a math challenged friend who had a amazing experience with some fire warriors. 24 rapid fire shots killed 8 terminators. To this day he believes his FW's are a great way to kill heavy armored troops. The sad truth is this.
24 shots = 12 hits
12 hits = 8 wounds
8 wounds = 2 dead MEQs 1 dead terminator.

Killing 8 terminator with small arms fire is an amazingly 'Lucky' event. But when we examine the expected outcome we see that this is most likly a rare event. However

I've met too many people who saw extraordany events occur and then expected that as their norm. Now that said 40k is a game of chance. So if there is a remote possibility

of an event occuring, then it can occur.

My last fallacy is not a statistical one. I nicknamed it the vacuum fallacy.

People tend to asses a units worth in a vacuum. By this I mean they fail to take all kinds of extra data in account when they are trying to determine the value of thier models. it's usually expressed as.
"my gaunts can kill your fire wariors" This statment is made while ignoring the hammerhead hovering gently behind said fire wariors, along with the ruined building that said fire warriors are hiding in.

Gdurant's post transitions into another mantra I hold dear for 40k, and that is 'safety in volume'. We grind out the probabilities for each event to find the best unit for X, but this is a dice game, and unless you are looking at huge numbers of rolls the exact probabilities will rarely pay out. The easy way around this is to always err on the side of taking more shots, more wounds, more rerolls when you encounter comparable efficiency to a unit with less of any of that. The higher the volume, the closer you can expect it to perform to the calculated probabilities.
This sells product too of course, so don't expect GW to be nerfing it anytime soon.

gdurant - I really have to disagree with your clumping theory. It's possible yes, but actually no theory or anything else than the effact of tossing a statistically relevent coin (And those are probably not the ones you pay with as the sides arn't equal.) you will come up with a 50/50 chance for each single try.
Now it can occur that you get clumped resulsts, but it also could be even results with only a few throws.

btw statistic isn't done by experiment. It's pure theory. 100%!

Maybe "vacuum fallacy" happens to less experianced players and twelve year olds. Probably combining both.


@Moz:
This is the reason why 'roll as many dice as you can' and 'force a lot of saves' is a very good tactic.

vogelfrei wrote:gdurant - I really have to disagree with your clumping theory. It's possible yes, but actually no theory or anything else than the effact of tossing a statistically relevent coin (And those are probably not the ones you pay with as the sides arn't equal.) you will come up with a 50/50 chance for each single try.
Now it can occur that you get clumped resulsts, but it also could be even results with only a few throws.

btw statistic isn't done by experiment. It's pure theory. 100%!

Maybe "vacuum fallacy" happens to less experianced players and twelve year olds. Probably combining both.


@Moz:
This is the reason why 'roll as many dice as you can' and 'force a lot of saves' is a very good tactic.

Statistics aren't theory. You can hyposis based on statistics. Such as "if I roll six dice I should roll one six" I may or may not roll that six. Now if I roll over thousand times and look at the averages. You'll find that over that time you might of rolled all sixes in a couple of rolls. you'll also notice that you'll going long periods of time without rolling certain numbers. every time you roll a dice you have an equal chance of rolling any given number. Or as we do in 40k you can add a success rate to the roll. Since a 4+ roll is a succes and anything less than that is a failure, you could flip a coin when shooting BS3 units instead of rolling. Keep in mind though. just because something might happen 50% of the time doesn't mean it HAS to happen. It just means that it is more than likly to happen and will probably happen 50% of the time.
Feel free to disagree with the idea of statistical clumping. However if you glance though any text book on statistics you'll find it right there.
If you want to check out a text that isn't as dry I'de suggest the cannon. The author dedicates a whole chapter to statistics. It's actually a very fun book (even if it is about scinece)

A statistic is a data collection experiment, not a theory. For example, say I collect a sample of people and ask them there favorite ice cream flavor.
50% chocolate
40% vanilla
10% strawberry
Now I could make a hypothesis that I were to go and check local icecream sales that I'de see about 50% of people bought chocolate. But your right you don't experiment to get statistics.

edit: I rewrote it for some more clarity, sorry if it's still giberish I'm under the weather and a little wasted on dayquill.

Not to be prudish about this, but discussion of statistics seems too basic for this, as an understanding of statistics is being assumed. Perhaps another thread would be a good place to discuss statistics if you wish.

DarkHellion wrote:This series of posts is meant to be a serious philosophical and mathematical look at the way in which players discuss and think about playing the game of 40k in a competitive fashion.

Sorry, it jsut looked lik part of your basic premise. Besides, a lot of players don't understand basic statistics. I just wanted to address some of the pitfalls. Alot of people might understand the math. How to calculate percentages and the like. But they don't really know how to apply it. Statistics are just part of the epistemology.

I'll let it die though. Your post your rules.

gdurant wrote:Statistical Fallacies:
Unfortunatly people don't entrensically understand statistics. Because of this you see people make subjective decisions about how the feel a unit should perform. My first observation is the luck factor in 40k.

Statistical Clumping: We've all had that moment of glory or utter dissappoint. We've seen people roll 8 1's for their terminator saves of watched a sanction psycher force weapon a lord of change to death. We first turn to mystisicm. We blame god budda or satan for not

living up to their end of an agrement. I know people who throw dice a way that underperform expectations. So if luck isn't responsible , what is? The term statistical clumping explains these phenomena. Basically clumping is the idea that random results wil clump together but even out over time. A good example is the coin flip. A statistics professor I had gave my class a challenge. She asked half of to make up 100 coin flips and right down the results. She asked the other group to flip a coin 100 time and record the results. She left the room while we fliped coins and coind flips. When she returned we gave her our results and she immediatly told figured out which one was fake. because in this case there is a 50% chance of either result the fake flips looked like this:

U D U U D D U D U D D U D D U U D D

While the real results appeared clumped

U U U U U U U U U D D U D U U U U U D D D D D U D U D

Over a long enough time line the clumping would of evened out. This is essentialy the luck of 40k. I've seen people with solid lists and good tactics lose just because they couldn't wound or hit anything. This brings me to my next fallacy.

Sort of, but not quite. A basic building block of statistics is the assumption of independence, that a previous result will have no baring on the next result. You flip a coin, 50% it’ll be a head… then you flip the next coin and 50% it’ll be a head, regardless of the previous result.

Now, I think your professor was getting at the idea of results ‘stringing’ together a lot more than people realize, that is to say that over 100 dice rolls you’ll likely get a few run-ons of 4 heads or 5 tails or similar. People underestimate that likelihood and when falsifying data will record a more even spread of results than a real test will demonstrate.

Similarly, in 40K people often underestimate the standard deviation from one test to the next. Firing 10 bolter rounds at some IG can expect around 4 and half kills, and most people will then think that 4 kills or 5 should result, maybe sometimes getting 3 or 6 kills for an anomaly. Truth is 3 kills or 6 kills are pretty likely, and 2 kills or 7 kills are something a lot of players will talk about for a long time afterwards… but are likely enough that you’ll see reasonably often in games.



I think the most basic concept in building an army is in developing an idea of what an army has to achieve to win. This is almost always excluded from discussion on list design.

There are a few general categories you’ll need to fill out to build a solid list. You need to be able to kill GEQ. You need to be able to kill MEQ. You need to be able to kill light, medium and heavy ground vehicles. You need to be able to kill light and medium skimmers. Very importantly, you need to able to take objectives, either by taking and holding or stealing them late with high mobility units.

The exact make up of what you will need to win will vary based on your playing conditions. Iif everyone around you plays marines your need to kill GEQ will be less important. If you play on terrain heavy boards then objective grabbing becomes more important, and killing power less important, but the opposite is true if you play on planet bowling ball. Lastly, the type of game you play is also pretty important, your requirements will be significantly different if you play random missions, compared to pre-selected missions, or custom written ones.

Different units will excel at different tasks, and plenty can meet the requirements of more than one requirement. This can make some units very useful despite poor mathhammer performances, just because nothing else in the list fills that role. But an army, no matter how many MEQ it can kill in a turn, will lose most games against skilled players unless it can meet all of the above requirements to a satisfactory level.

My submission is a bit less crunchy than ROI, statistics, etc. A set of concepts that we've been bandying about on the B&C is elemental archetypes, which can be applied to armies, to units, and to groups of units. These concepts were adapted from the Asian masters such as Sun-Tzu and those guys, but Silent Requiem and described over in this thread.

He goes into great detail, particularly about tactics for Water armies, but I'll summarize.

Earth armies win by outlasting their opponent. A marine SAFH is a good example of an army. A devastator squad is a good example of a unit.

Fire armies win by overwhelming their opponent, often (though not always) in assault. A stealer shock army is a good example army. Harlequins would be a good example unt.

Air armies which by outmaneuvering their opponent to create localized concentrations of force. A Ravenwing army is a good example army. A land speeder is a good example of an Air unit.

Water armies win by remaining flexible at all times, retaining the ability to attack or retreat as needed, firing all the way, or plunging into assault. It is a reactive force in many ways, prepared to deal with any of the other elements. A Grey Knight army is the example Silent Requiem uses, and in his hands it is indeed a pure Water army. A default-armed dreadnought is an example of a Water unit.

@Pariah Press, That is the guy with what seemed like the greatest win/loss record for a Grey Knight army ever. That is until you looked at his 1000 point list and saw the 2 land raiders in the list. The amazing thing is that his opponents at the store he played at did not adjust their lists over time to his.

Darrian

Does he meditate beneath a waterfall?

I don't really like the elemental archetype concept, personally. I read his post, and it seems to overly complicate things.

ALL armies need to be reactive to some degree. You can design an "air" army all you want, but when you run into a more mobile army, you need to recognize that fact and play more like a static army. You can run an assault army, but you have to realize when your opponent's assault units outclass yours and switch your tactics.

I think he tries too hard to force a "four-element" viewpoint. I believe that the truth is that, when analyzing armies, there are two parameters that you can plot on a graph opposite each other. One is Assault/Shoot and one is Mobility/Durability. This allows you to plot static-shooty forces (Guard, SAFH Marines), mobile-shooty forces (Mech-Tau), mobile-assault forces (KoS, Wych Cult), and static assault (Footslogging orks), each in separate quadrants of a graph.

The armies in the extremes of each quadrant are obviously better at what they're setting themselves up to be able to do, but less able to adjust on the fly, whereas the armies that are plotted closer to the center are less focused, but more able to react.

Slient Requiem's "water" army is simply an example of an army that plots in the center: equally mobile and durable, equally shooty and assaulty. That means when he faces a very shooty army, he can go to assault, and when he faces a very mobile army, he can bunker, and so on.

A good player is the one who can not only understand their position on this graph, and their opponents, but knows how to adjust their game to account for the differences in position.

Looking at the current "tier 1" lists, they're both plotting close to the center. Both Godzilla Nids and Mech Eldar exhibit good durability, mobility, assault, and shooting capability. That gives them the ability to pick the best approach to deal with any given opponent or mission.

Mech Tau will never be as all-around strong as Mech Eldar, and it isn't just because of the differences in their skimmers. Mech Tau doesn't have the ability to take assault to an enemy that presents a stronger shooting capability and has to resort to hiding and VP denial, whereas against a static shooting army, Mech Eldar can deploy a troupe or two of harlequins and run through their ranks.

Like so...

For typical Trifalcon eldar, Mech-tau, SAFH marine and Zilla nid.

Edit: and please don't post 500 follow-ups arguing with the numbers. I just punched those in over a matter of seconds without much thought. The idea is to give an example of the quadrant layout with examples we are all aware of.

Yeah, those are kind of cool.

So, if you take two graphs, and superimpose them, the general who is best able to exploit the regions that their graph covers and their opponent's doesn't stands the best chance of winning.

Obviously, an army that has more total area is then more flexible in the types of approach that they can take. A SAFH army is going to do the best if they're able to play their game, engaging all their shooting against opponent's shooting, as they're better at that. If they get dragged into a game of hide-the-skimmer, they're going to lose.

Well, you have to make sure you get the scale right; durability, for example, may end up being more important than, say, assault.

Overall, I really like this idea. Trying to come up with all the definitions through freestyle debate will be VERY tedious, though; as the poster says: "Meetings: None of us is as dumb as all of us". I think it'd be more effective for each poster to come up with a set of definitions and guidelines as completely as possible, and we can analyze each set from there.

Kudos for a really good start on this, DarkHellion!

Nice, I'm taking a break from work, and then this thread looks exactly like some of the graphs I was working on :-P

Here's one thing I'd like to throw out there. The concept of a "star unit"

Everyone has one of these, where they usually perform much better than expected, because the player is personally attached to it, and is far more careful with its use. You know, the unit you come across that you writeoff as points inefficient and useless, and then they end up turning the game around. While it's hard to quantify, I think that definately has an impact.

These are the kind of threads that make Dakka the best 40K site around.

Thanks Darkhellion.

Darrian

Thanks Moz, I was thinking about graphs to chart what we were talking about.

You present some interesting strategic concepts.

We talked a lot about this a long time ago in a Thousand Sons thread. Since most armies are either categorized by being either shooty or choppy, but the Thousand Sons did not really fit into either category, but were very durable. The same can be said for Necrons.

Redbeard wrote:
So, if you take two graphs, and superimpose them, the general who is best able to exploit the regions that their graph covers and their opponent's doesn't stands the best chance of winning.

That is about it. The part of the graph that only you are in and your opponent is not in is your strength. The area that you are not in, and your opponent is, is your weakness.


It is interesting to note at some successful armies just load up with one aspect of the continuum, and ignore the others.

For example durability:
Necrons are heavy into durability.
Thousand Sons
Death Guard
Orks now will be here with a horde approach.

Shooting:
Imperial Guard

Mobility:
Dark Eldar

Assault:
Stealer Shock
(Also the ex-demon bomb)

Nice stuff.

Heres another concept, build on the premise that every unit is a compromise between shootyness, assaultyness, mobility and durablity. I will generalize here and there to keep things legible. For simplicitys sake lets assume that units are either shooty or assaulty and either mobile or durable.

1. When an assault unit faces a shooty unit, mobility will win the game, because the mobile contender diktates the distance of the battle.

When same types meet, as in assault vs assault or shooty vs shooty, mobility doesnt help. Both armys want to fight at the same range, so theres nothing to gain from a certain distance for anyone. This means that investments in mobility are mostly wasted, so overall the mobile army gains little and suffers from a relative lack of durablity against a force that didnt waste resources. In short:

2. When same types meet, mobility loses the game.


Do you know "rock, paper, scissors", where every type of unit beats another type in a never ending chain?

The chain for 40k goes like this:

Durable assault > Mobile assault > Durable shooty > Mobile shooty > Durable assault > etc.

Now lets apply this to armylist development and metagaming. Lets first analyse why Godzillanids are such a well rounded army:

Dakkafexes are undercosted, so the list uses these as its core and follows the theme with more TMCs. The TMCs can be considered "Durable shooty". Durable shooty armys main weakness are "Mobile assault" opponents, so the smart Nid player fills the rest of his list with a counter to "Mobile assault": Genestealers ducking behind TMCs, who in this role are "Durable assault".

When you check the above chain you will see that not only do Stealers cover the weaknes of TMCs, which is "Mobile assault, but also do TMCs cover the weaknes of Stealers which is "Mobile shooty".

Yeah those graphs are an excellent way to design an army or visualize one. Very nice!

- G

Expansion of the platform idea, and more discussion: The platform seeks to expel what I believe is the erroneous view that units are composed of large numbers of individual troopers who are differently armed. While this is true of the construction of the unit, it does not reflect upon how one should or does utilize the unit in the game itself. Within the game, units function far differently than either the sum or the product of the whole. Instead, we must create a calculus of unit evaluation, designed to examine the integrated value that said individuals give when utilized as a whole within game. This is theoretical and arbitrary, as it exists within a vacuum while in game the units will be used in conjunction with other units, however this does not detract from the lack of language to express how a singular unit interacts with the game state.

This is the type of thought I intend to expand upon in further discussion. This is to be a rigourous look at all angles of the game from a basic foundational level. I will confess that I am not the greatest player of 40K, however I believe we can develop a system of thought that allows the basic level of 40K play to rise dramatically through formulating and understanding the game theory that we utilize in warhammer.

I really like the graphing system you guys are laying out, but I do have one thing to add to it.

The positioning of the 4 points of Durability, Mobility, Assault, and Shooty is important. My first assumption when looking at the graph is that more area is better and that even though there are a lot more factors, the area of that graph can give you a rough idea of the power of an army list.

The problem is that the order in which you put the 4 factors can greatly effect the area. For example if we were to take a fictional army that got a score of 10 in mobility and 10 in assault, but only had a 2 in durability and a 2 in shooty. If you organize the axes in the same way as above you are going to get a total area of 40 ((10*2/2)*4). Whereas if you put assault and shooty across from each other and mobility and durability across from each other you would end up with an area of 74. ((10*10/2)+(2*2/2)+(10*2/2)+(10*2/2))

The next question becomes, what is the best layout?

I think that the way to approach that is to look at which of those 4 factors have the most impact on the effectiveness of each other.
Shooty and Assault by far the least impact on each other in the fact that while they are both useful neither one really helps the other.
Both Durability and Mobility have a big impact on how well or how long you can make use out of the shooty or assaulty elements in your army.
While Mobility can have a pretty strong impact on your overall Durability through denying shots, I think that it is relatively easy to factor that into an armies durability rating.

So based on that, I would recommend putting Assault and Shooty across from each other and Durability and Mobility across from each other.

While it still doesn't paint the whole picture, I think this will give a more accurate first glance at the relationships between the 4 factors and their impact on the strengths and weaknesses of an army.

Raider wrote:
When same types meet, as in assault vs assault or shooty vs shooty, mobility doesnt help. Both armys want to fight at the same range, so theres nothing to gain from a certain distance for anyone. This means that investments in mobility are mostly wasted, so overall the mobile army gains little and suffers from a relative lack of durablity against a force that didnt waste resources.

I don't agree with this.

I think that mobility is always a factor.

In Shooty vs. Shooty it lets the army that is mobile get the best angles, and limit the amount of return fire. Take for example IG vs. Eldar. Eldar can move and use terrain and distance to limit the amount of return fire from a static IG army that can out shoot it.

In Assault vs. Assault army it can be used to get the charge which in almost all cases is very important. It also lets you out maneuver the other army. Take Orks vs. Dark Eldar for an example. The Dark Eldar can fly to the left or the right and hit a part of the Ork army and overwhelm a flank, and the Orks will be slow to react.

(Note: The examples are just illustrative generalizations)

I think that the matrix is too simplistic. For example. "mobility" can mean multiple things - being able to move large distances, or being able to jump shoot jump (or both). It also applies differently when combined with assault units or shooting units. For an assault unit, moving 12 or moving 6 and then jumping 6 would be largely the same (or actually worse), but for a shooting unit jump shoot jump is infinitely better than just move 12. The nature of the platform matters too, since vehicle platforms and non-vehicle platforms have very different types of durability which can also affect the value of the move types - but this is maybe a durability question.. It seems like it would get very complicated, except perhaps that there aren't many units that can JSJ.

There are different types of strong shooting, and different types of durability, etc. Extreme examples:

Dakkafexes don't do much against land raiders

Las/Plas teams don't do much against fortuned, turboboosted bikes

But we all know both units are very good - and both would be fine with their targets switched.

Maybe this just means that individual ratings for durability, shooting, etc need to be lowered, with a top score going either to no army, or only an army that can effectively deal with all targets at all times.

One concept that I like to make use of (and I rarely see anyone else use something similar) is that of time value. There are a wide variety of units out there that have a wide span of uses. However something that not too many people look at too closely is how long that unit will be useful for. You might have a unit that is an absolute hand to hand killer (tooled up daemon prince for example). However they are only going to be of much use if they get into hand to hand. While this may not be much of a problem, it will take some time. During this time, you are not likely to be doing anything useful. So if the prince doesn't get into hand to hand till turn 3, his time value is going to be lower than something of similar power that could get there in turn 2. An extreme example of this would be the reason (or at least one of) that sisters repentia really suck. They simply take too long to get where they can be effective and they are often depleted when they get there. An example of a high time value unit would be dark reapers. These guys have a 48" range on their weapons, so its unlikely that they won't have a target on the first turn. After that, if they have been placed well, they are likely to have a target every turn afterwards. Infantry units tend to have a higher time value than armor value units due to not being affected by escalation rules (time value of units in reserves = 0).

The second part of time value is continued usefulness. Units that have been destroyed are no longer generating any value for you. The dark reapers are likely to last a good portion of the game. Their long range weaponry will keep them out of range for much return fire as well as allow them to be deployed well behind friendly lines. To further boost their survivability, they have 3+ armor saves and are very likely to be deployed in cover. This will, in general, increase the time they are on the board to generate value. Units such as land speeders have a lower time value. While they are likely to be in range for shooting on turn 1 if desired, their low armor value combined with their short range means that they are more likely to suffer casualties than the dark reapers. Predators and Hammerheads that have been shaken or stunned are no longer doing anything useful for your army, so they are not generating any value even if they are still alive. On the flip side, falcons are often tank shocking and delivering short ranged units to the front once they are stunned, so they continue to generate value (although less than if they were still able to shoot). This is yet another reason why flacons are good. The addition of extra armor (or its equivalent) to every transport that can have it is yet another example of time value. A transport generates value by extending the movement of the unit inside (and possibly by blocking line of sight to things behind it). If it isn't doing that, it isn't generating value. The addition of extra armor to allow it to continue generating value when it otherwise would not be able to is worth a lot...certainly more than its current point cost.

So while its good to look at what a unit can do, its also very important to look into how long its likely to be able to continue to do it. If this technique is also combined with the ROI, you can get some very informative information.

I’m going to differ with the general consensus and argue against analyzing armies by the mobility/shooty/assault/durability matrix. It’s an interesting concept, but ultimately I don’t see any use in its application for either new or experienced players.

The biggest problem comes with scoring complex things on single axes.

How do you account for different kinds of mobility? Terminators can deep strike so they’re highly mobile when deploying but are pretty slow afterwards. What about 12” charge ranges (which only increase mobility when I want to assault, if I’m trying to stay at range I’m as mobile as slow as if I didn’t have it at all).

Having single scores for shooting and assault ignores the various targets in the game. An army decked out with masses of plasma guns and lascannon would have a high shooting score, but would be unable to output the shooting necessary to kill a really troop heavy IG army. Ultimately, you don’t just have an overall shooting ability, you have shooting ability against a wide variety of targets, GEQ, MEQ, light, medium and heavy tanks and light and medium skimmers.

For a real game example, my tyranids often have a good amount of high strength shots and a mass of low strength shots, without any decent AP. This allows me to shoot up tanks and poorly armoured infantry, but is mostly ineffective against marines. At the same time my ‘stealers are brilliant at ripping through marine squads, but surprisingly ineffective when faced with horde assault troops like orks. There are plenty of ork armies under the new codex that would show higher shooting scores than my list, but lower assault scores… so by the logic of the model proposed I should assault. Yet experience tells me I can actually kill a lot of orks at range, and would be better off trying to keep the game as a shooting match.

Durability has the same problems with representing a complex thing with a single number. Necrons are really durable… unless the enemy has a lot of St8 guns or power weapon/rending melee attacks. And where would rhino mounted marines fit in… the marines are the bulk of the army and very tough to kill so they should have a strong durability score, but if you shoot up their transports they’re stuck in the middle of nowhere.

Then there’s also a problem with marking a whole army by one scale. You can have an army with static shooting elements holding the flank and centre, while highly mobile assault units race up the other flank. In that situation scoring your army as a whole misses the important difference between the two distinct elements of the army.

At the end of the day, attempting to quantify an army on such a general level will lead you up the garden path. Better to determine everything an army has to be good at (killing various infantry and vehicles targets, holding objectives with tough troops, stealing objectives with mobile units, nullifying enemy units) and determining all the ways you can achieve that, and building a list to meet each need as strongly as possible.

From there you can keep in your mind the abilities of each unit to meet each objective. In each game, based on your opponent’s list, the terrain and the mission type, you can quickly assess what role each unit will need to fulfill for you to win the game.

For example, a unit of ‘stealers is excellent at taking out MEQ troops, whether they’re assault troops or shooting units, is very good at killing static tanks of any sort, and very poor at killing mobile tanks and skimmers. They’re not fast enough to steal objectives, and aren’t tough enough to be all that effective at holding objectives. So each game I identify the best target for some dedicated rending, and send them to it. As such, depending on the terrain and the enemy, I’ll generally assign them the task of killing elite enemy units, and occasionally to killing static enemy tanks.

At the same time, I have units of gaunts… poor at taking out any enemy unit, but they’re fearless and I can have cheap units with lots and lots of wounds, so they’re excellent at tying up quality enemy units for multiple turns. They’re incapable of taking out any vehicle. So generally I identify an enemy unit I want to keep occupied, but will be too hard to get my ‘stealers near, and tarpit the target. However, terrain and the enemy list will often mean I’m better off keeping out of LOS than tarpitting enemy units, at which point the gaunts can switch to objective holders, cheap enough that I don’t mind wasting their limited offensive ability, and with lurking are quite survivable in the right piece of cover.

It isn’t as elegant and it doesn’t come with colourful graphs, but I think it’s a lot more likely to be useful to players.

Has anyone come up with a fitness function to quantify these four traits?

Shootiness = BS * Avg Weapon Strength * Avg weapon Range
Assault = WS * I * Strength
Mobility = Move+Fleet+Assault
Durability = Armor * Invul Sv * T

I am not recommending those as I put almost no thought into them other than to pull the relevant stats. Coming up with Avg strength/range is tricky also. Anyway it is something I have thought about before but never pursued. I would try to find the max in each category and then standardize everything to the max. If you wanted higher resolution you could try to throw in corrections for all the special rules. In the end it would be quite the monstrosity. Anyone looking to do a Math PhD?