Welcome to Librarium Online!
I have many thoughts on 40k and how GW isnt really doing a great job with their points systems for the armies so i always thought to myself "if i am going to pay $40 for a rule book id at least like to know the people designing it for 6 months had half a brain for their numbers" of course i'd end up buying the rule book anyway because of what they did do right (stories and lore). These days i like to occasionally design units and i have great problems coming up with points for them based on logical values as i can't really be bothered thinking too deeply into the inter-relations between other stats and their points.
this thread in particular is going to be an open discussion on the base line of statistics used in 40k and how many points each stat should be worth with the ultimate goal to come up with a generic template (can be reasonably complex) for people who want to design units to get their figures from. Ultimately that means if i wanted to make a 10 stats to everything creature with a (2+ save of course ) monsterous creature i'd be able to come here and find the right sums to get a pretty accurate cost for the base model before moving onto the finer art of points costing special abilities
Unfortunately games workshop loves space marines and have made lots of SM babies so our template for working out stat lines should be a space marine which is WS4,BS4,S4,T4,W1,I4,A1,Ld9, sv3+ as the generic points/stat ratio it's not an easy one to use because they also have the mighty bolter S4, AP5 rg 24" rapid fire and "they shall no no fear" included.
To start this off I am going to hit into the easiest topic for me: the armour save!
simply put from what weapons are available to negate the save and the chance to successfully save a shot, the actual points of the armour save isn't just a simple straight line, when we look at gaunts it's pretty easy to see: things they get to use their saves against = ap - weapons which on the top of my head i can't answer and the chance for the gaunt to survive the weapon is also pathetic
on the flip side looking at the terminator we can list weapons that ignore it's saves: bright lances, melta guns, vindicator shells, plasma weapons. against ap3+ weapons they get around 83% chance of survival compared to the gaunt with a 16% chance of survival on the off occasion it actually gets the save.
on a straight list of % chance to save against shots the table flows: 16%, 33%, 50%, 66%, 83% give or take rounding
the second table is one of pain where I count all the weapons (without multi counting the bolters) in the game and tally their AP to put them into a % figure naturally i can't be bothered doing this right now and will probably edit it later with values.
total AP - :
the idea would be to work out using those two % numbers the % of guns that ignore your save and from there use the data to come up with a good costing for saves.
I'd like to hear what people think about the whole concept i'm trying to develope be it good or bad and maybe even other people's approaches to solving the problem of how many points is a stat line worth??!!
The issue with this is that this is the way that old 40K editions (1st and 2nd) were costed. They were completely divorced of the build of the entire list, and you would have a different points system for each army relative to the units that already exist and the slots that they occupy. One example is:
How much is a Terminator as a troopor scoring unit? There are even then numerous different answers to teh same result (Lysander-wing, Deathwing, Logan-wing etc etc)? How much would a Termie cost in an Iggie army? (a helluva lot more than a marine army!). How much would a Wraithlord cost in a marine army (again a lot more!) or in a nid army (probably less as there are equivalent beasties in the list already).
At the basic level, you take a unit and then another unit from the same army with the same rules but different stat profile to try and cost up e.g. Veteran sergeant vs Tactical marine. Costing a tactical marine vs an Assault marine gives you an indication of what jump-pack infantry get over ground-pounders.
However a near identical marine from the Space Wolves dex costs a different amount, and yet the two are ostensibly the same model.
The new GW system is less transparent, but what it does allow for is tweaking points per unit entry rather than a generic base profile.
I agree whit Kithre as creating custom units is my pet hobby (almoast wrote hobit ). The easiest way to start a model is simply look after a model that is very identical to what you are after.
yeah that's true but it's not exactly balanced finely and i am merely wondering whether it's possible to go down the other route for the basic stat line to get a standard kind of template that would be good to use before you move into the territory of special abilities.
it will take a reasonable amount of math but i think it shouldn't be somthing a unit designer should fear for the stat line is the core component used in the dice rolls and everything else is built over it. sure it's easier to just get a template from another unit of similar power but where's the fun in that?
The problem is a statline comparison isn't even consistent across armies. Look at a Dark Elf spear elf vs a High Elf spear elf. Ignoring special rules (ASF, fight in 3 ranks, hatred etc) the models are identical. Except the points values are materially different - the special rules DO play a part.
A unit without special rules still has the rest of its army to back it up - take for example Dwarven warriros. They don;t have any special rules really, but they are teh cost they are becaue the rest of teh army is chock full of nasty units that can compensate for weaknesses and allow the Dwarven army to field a resilient SCR unit. Older editions of warhamemr didn'ty do this so things were not balanced at all - a High Elf cost the same as a Dark Elf cost the same as a Wood Elf. But army dynamics and costing of elites (+1 WS, +1 BS etc cost almost a fixed PV) meant that the army points values, whilst cosnsistent within a list, were not consistent in comparison to each other's lists.
ANother factor is mounted troops. In older editions they cost double the base trooper + barding/steed cost. This isn#t really a fair reflection of their battlefield prowess.
Thanks for pointing me to your thread, Gintoki.
Unfortunately I am unable to help, despite my love for math-hammer. The problem is the complexity of the issue that would have to be broken down into a sheer intimidating number of factors that would -in turn- each have to be assessed individually - a task that is nigh impossible to complete, and I wouldn't even know where to start.
This, I think, is the reason why such a system is non-existent and why points cost end up being imbalanced, some units being much more effective, points-wise, than others. You can make certain à priori judgements and you can do some playtesting, but you can never foresee every aspect and thus the true value of some units is only revealed a long time after the unit has been released and in play for thousands of games by hundreds of players...
A system for calculating the costs of single unit would have to account for so many things -some of which have been named here by others- at once, that a modular approach seems little appealing. Certain stat combinations are more valuable than others, more valuable to certain unit types, produce greater synergy with certain special rules and/or equipment, are more valuable in certain slots than in others, and benefit specific armies more than others.
An example: increasing the number of attacks in a profile should be more expensive on a model with high strength than on a model with low strength. Giving a unit S5 should be more expensive on a beast than on infantry. Giving a unit the feel no pain special rule should cost more for T5 models than for T3 models, but at the same time should be less expensive on Sv2+ models than on Sv5+ models. Giving a power weapon to an S5 model should be more expensive than giving a power weapon to an S3 model, but this should again be factored by the number of attacks and whether or not the model has a second close combat weapon and so on. The same unit as a troops choice should be more valuable than a unit as an elites choice. Another GEQ-like troops unit should be less expensive in IG armies than in Necron armies who would -as an army- benefit from such an alternative much more. You also have to factor in equipment options like added weaponry or dedicated transports and the like.
I come to the conclusion that such a project -however desirable- is too huge to grasp, because individual modifications can not be factored in independently, but instead influence each other very drastically and thus make it nigh impossible to cover every possible nuance. It would probably not be much more work to actually create every possible statline (characteristics, unit type, special rules, equipment, FOC slot, overall army) and assign it a points value, while this would at least ensure accuracy and balance.
Unfortunately, both these approaches are -in my opinion- not realistically promising. In the end you will have to use gut-hammer instead of math-hammer in this case, and playtest as much as possible, trying to find an optimally balanced cost where the unit does not dominate the game but does not appear useless either. I can calculate the odds of a deep-striking unit scattering into some nearby enemy squad, but I can't do this.
Sorry I am unable to help. Good luck!
This has been hashed out actually, in the back of the WH40k 3rd edition rulebook.
Several things factor into the effectiveness of any given model, weapon, or selection that actually make Warhammer incredibly difficult to balance.
The example that they used in the 40k book is the price of a missile launcher in a Space Marine Tactical Squad, versus the price of a ML in a Marine Dev squad.
In a tactical squad, the missile launcher forces a duality of roles- either fire against a tank, or fire against infantry. If firing against a tank, the other 4-9 members of the squad are probably useless, their weapons unable to pen the armor of the vehicle. If firing against infantry, the missile launcher is only marginally better than a Bolter (assuming you're rapid firing) and can't be fired if the squad has moved.
In a devastator squad however, the entire squad can be equipped with anti-tank or anti-infantry weapons. Now, if the missile launcher chooses to fire on a tank, chances are good that the whole squad has a good chance of knocking the vehicle out. If firing on infantry, the entire squad is probably geared with multi-shot weapons ideal for knocking down the target. Furthermore, the entire squad is armed with heavy weapons, so move/fire is entirely different.
You could extend this to other ML scenarios as well. How much more is a Missile Launcher worth in the hands of a BS4 model than in the hands of a BS3 model? What about mounted on the back of a terminator, who eliminates the move-or-fire restriction? What should you do to include the chance of a nearby model allowing re-rolls?
Those are the problems just for balancing a missile launcher. Other problems include the number of wounds in a squad, the number of outgoing shots or attacks, their ability to withstand inbound damage, ablative wounds, unit strength, and their interaction with other special rules in the unit and army at large.
Warmachine, AT-43, Flames of War, and other games like this have a simpler time of balancing their units because they have fewer special rules and few to no options for equipment or unit size. In the GW game Battlefleet Gothic, players were able to reverse engineer the cost of a cruiser to an almost 100% accurate formula known as the 'Smotherman's'. This was easy because cruisers are fielded as a single-model-unit. Everything about them is stand-alone, and the rules that they follow for joining squadrons is standard across the entire game. However, there are cracks in the game that can be exploited. People don't expect you to make enormous cruiser squadrons, or to link torpedo salvoes. In situations such as that, the value of a particular ship feels like it should triple. So the formula is accurate in certain situations, but you can still crack the game. If you've ever gone up against a pure infantry rush in FoW, you'll understand.
The best way to determine the points value of any given unit is to playtest it within your army and try to guess how much it should be worth. As always, you should always aim to make your unit a little over-priced. Underpricing a house regiment will result in people thinking that it's cheesy. It's better to say "yes, that regiment just rolled your army, but he cost me 544pts" rather than, "I know he rocked your army, but I paid the 25pts that we agreed on!"
whilst i can agree to what everyone so far has been saying there is one factor that is being ignored here: I am specifically asking for a look at the basic stat line and the relationships between the different statistics and not looking to fully balance ideas based on pure mathematical theory. I am instead thinking is it possible to get a good points outline of the basic stats before moving to gut-hammer and failing that is it possible to make good graphs that effectively show the change in results for the basic stats so people working out points costings for units have an idea of what they can tinker with.
this means i really want to know how much statistical power an all 10 stats model has or if an all stats 10 model with a 6+ save could lose to an all 9 stats model that has a 2+ armour save and whether it's possible to make the points system work around that logic before any other factors are put into place.
Now that i know people don't like this approach so this is now going to be a mathematical challenge for people who are interested in working on this line of design and as a first step in the process I think we need to work out which stat combos influence the results.
Combat: Number of attacks ----> Weapon Skill vs enemy Weapon Skill ---> Strength vs Enemy Toughness ----> Armour Save/Invulnerable Save
Ranged: Number of Attacks ----> Ballistic Skill ----> Strength vs enemy toughness ----> armour/invulnerable save
naturally we already know the chances for the individual roll offs because of GW's table the challenge is in working out how much of a difference in one area does it take to balance out another to give effectively the same result: eg does it take 10 attacks at S3 to match the wounding power for 2 attacks at S5?
I guess i'm thinking it's possible to create a p1 x v1/t1 situation for 40k
An all-9-stats model with a Sv2+ would beat an all-10-stats model with a 6+ save into the dust in close combat given only this, but make them both monstrous creatures and the result turns 180°. Comparing statlines won't help if the other dimensions defining a model have such great impact on how the statline reacts to the game world.
It's really a difficult and complicated matter.
Actually, Red Archer, if he is looking for the pure, virginal stat, without any kind of consideration for a unit, special rules, wargear, wargear options, and armybook, there is a way to do it. But it's horribly drawn out and very complicated. I can try to explain it, but I'm no mathematician or even a math major. However, if calculus scares you, avert thine eyes.
Gintoki, I understand the answer that you're looking for. Infact, I thought of doing it, but the math made my head spin and the time involved was simply horrifying. The fact that GW is demolishing online play-aids makes this even more difficult now than it was then.
Here's how I would do this:
Find unit A, declare as constant
Find unit B, declare as group 2: make this unit as similar as possible to the first, with ideally 1 difference
Declare Variable: find that 1 difference
B-A = V, where V is the price of the variable.
You're not done yet. You need to check your accuracy.
Find unit C, declare as constant
Find unit D, declare as group 2: this unit also only includes variable V
Check for price of variable:
D-C = V, is this a true statement?
If yes, then you should ideally check that variable through the entirety of the book, but it is safe to assume that you have found the value of that variable.
If no, continue checking and write down the value of 'V' in each test. At the end of the tests, round the answers V into a single number. This will be the estimated cost of the variable.
That's not so hard, I hear you say.
But wait. When you go through the armybook, the BEST way to solve for any variable would be to determine the exact cost of every variable present. This means special rules, and even unit OPTIONS (devastator marines cost more than tacticals. Why?).
You'll need to also make sure that each of these examples comes from a single armybook, so that you know that the selection of possible army-wide combos is taken into account. Once you have the totals for a Guardsman with a Statline of all-10, you can only make guardsmen units. To make a marine unit, you would need to repeat the system to find a marine with stats all-10. To find the "true value" of any stat, entirely stand alone, you would have to repeat this method through every armybook.
Further more this method means that at times, you might be checking for two variables or more at the same time, because it is rare that any two units will have the same set of variables acting on them.
When checking for 2 variables, you'll have to write down your outcomes like this:
B-A = V
V = x+y where x and y are the two variables.
Then you could end up with statements like
(x+y) = 9
(x+z+w) = 9
This is what makes it so hard. In most cases, you're looking at extremely ugly formulas and trying to solve for a common key. It's advanced calculus really, and while possible, it's just ugly. You could program a calculator to do it, but the program would have to be entirely situational, and would take almost as long to write as it would to do solve the equation by hand.
Now that you know how to do it though, get back to us with your reports. I'll expect them sometime before the apocalypse. That gives you about 2 years according to most sources. I think that might be enough time.