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.

You shouldn't want to.

If we're talking about a simulation, even a casual simulation, you should want different units to be good in different contexts. The value of individual units should change based on those contexts. Terrain is the most obvious example. A busy board with lots of LOS blockers should devalue a tank like the Predator, which has long range weapons requiring LOS, and begins to miss mkre when on the move.

In order to achieve "complete mathematical balance" tactical and strategic context needs to be tossed. But if you do that, you don't have a wargame.

Probably not. And even trying to do it would almost certainly require scaling/sliding points for different game-sizes and multiplicator effects.

Mortarion at 400something points is a very different beast in a 500 or 1000 point game than he is in a 2000 point or, just for fun, a 5000 Apoc game. This probably no longer the case, but I remember reading some opinion piece or so during 5th or 6th edition that made a pretty good case on how quite a few issues could at least partly be explained by the "normal game" of 40K in the UK being 1500 points usually (and thus presumably what GW writers would play in most cases), where as the "normal game" in the US was often 1850 or 2000 points, which simply played out differently for the effectivenss of many units, the importance of the 1st turn, and a bunch of other things.

Re-rolls or other bonuses change the relative effectiveness of units and would probably need to be priced relative to the units that benefit from it in a given game, not a fixed point cost for a character that brings it.

And of course, differences in missions also bring out different things. All those (surely well-intentioned) ITC houserules and the habit of mixing Maelstrom and Eternal War missions popular in the US also really changes the dynamic of the game, usually placing (even) greater emphasis on board control and large blobs. It's surprising how good Primaris armies actually are/can be in UK tournaments, while also being utter trash in the US scene.

Etc....

Sunny Side Up wrote:
Probably not. And even trying to do it would almost certainly require scaling/sliding points for different game-sizes and multiplicator effects.

Mortarion at 400something points is a very different beast in a 500 or 1000 point game than he is in a 2000 point or, just for fun, a 5000 Apoc game. This probably no longer the case, but I remember reading some opinion piece or so during 5th or 6th edition that made a pretty good case on how quite a few issues could at least partly be explained by the "normal game" of 40K in the UK being 1500 points usually (and thus presumably what GW writers would play in most cases), where as the "normal game" in the US was often 1850 or 2000 points, which simply played out differently for the effectivenss of many units, the importance of the 1st turn, and a bunch of other things.

Yeah obviously anything you flat out can't take more than 1 period is going to have notable scalability issue. Even something as simple as tactical marines would have different impact based on point size if they were max 1 unit ever never mind big thing like primarch!

Similarly psykers increase or decrease in value based on points. Yeah the darn smite stays about constant(though varies in worthyness based on opposing army) but all the others...one -1LD spell in 500 pts is lot more impactfull than same in 5000 pts game.

Looks unfeasible.
I guess that balancing 40k is an NP-hard problem.
In the first place, it requires a statistical model with states and probabilities. This model would be rather large. I guess the studies/questions asked would be posed on a higher level so that statistical reasoning and inference would be necessary. Have a look into the hidden Markov model for speech recognition which is the simplest nontrivial Bayesian networks.

Probably, it is theoretically doable, but the complexity is too high for it to be practical. Certainly it would require much more than a simple spreadsheet
But most importantly, there is the problem of the shifting meta. To be balanced, our system of balancing would need to take the meta in accord. The meta is different from place to place, but this theoretically could be dealt with. However, our very act of balancing will change the meta on which its balance is based, thus creating imbalance once more. Therefore, balancing in such a way is a self-defeating process, and true balance ultimately can never be achieved outside of a vacuum.

 Iron_Captain wrote:
Probably, it is theoretically doable, but the complexity is too high for it to be practical. Certainly it would require much more than a simple spreadsheet

Well sure. Formula just needs to account(not even comprehensive list) for one unit:

game size
scenario
exact terrain(shapes, sizes, density, any special rules etc)
exact opposition army
rest of YOUR army
stats of units keeping in mind value of stats depends on other stats, weapons etc

Of course when this formula then alters one unit it has habit of changing your army which in turn results in formula altering value of other units. GG

Formula sounds like hard work, how do you facture things into a formula that aren't quantifiable? Special Rules, Strategems, Chapter traits,Relics... etc.

I'm sure that technically you could mathematically balance 40k, but it would take way more time and effort than the classic methods.

I helped test and work on a game a friend was making that was heavily math based in the balancing department. He'd constantly track how many turns the game would run, when the first shots would fly, how much damage was done each round by groups of weapons, so on and so forth. The result was an oddly balanced point system, but it required an excel spreadsheet programmed with the formulas. Weapons would cost differently based on what arcs you put them in, the speed of the ship, and other secondary subsystems. It was interesting, but generally not feasible for mass consumption.

Better to just eyeball balance with 40k at first, then refine through playtesting and basic mathhammering.

 Peregrine wrote:



Automatically Appended Next Post:

Breng77 wrote:
No you can not because there are too many variables involved. How do you value a weapon that ignores LOS? That depends on terrain which is not standardized. How do you value a unit with the conscript stat line? The value is different in an army like AM with good shooting, than it is for orks. It is also different depending on buffs available. Balance can only really be approximated with math, then would need adjusting through testing.

These variables are not an obstacle to math. You handle it the same way you handle it with playtesting, by evaluating a unit/list/whatever against a diverse metagame. As a very rough approximation you'd have something like point cost = (best-case value)x(probability of seeing best-case scenario) + (worst-case value)x(probability of seeing worst-case scenario), where each situation contributes a share of the point cost approximately equal to how frequently it occurs. And if you can figure out the relative frequency of each situation well enough to have meaningful playtesting then you can incorporate this into your mathematical model.

Sorry but they absolutely are an obstacle in math, the issue is that given the lack of limitations or standardization in the game the possible metagame against which you need to account for is simply too large. For instance what is the best case value? How do you determine the probability of facing that best case scenario. For instance a best case scenario for devestators might be having a ruin in their deployment zone that allows them to see the whole battlefield, then depending on their weapon choice would determine their best case opponent. How do you determine the probability of that happening, when there is no standardized terrain set up, or even existing recommendation about the level of terrain that should be used in the game. Further in a game that allows things like a force comprised entirely of T8+ 3+ save units, and one comprised entirely of T3 5+ save units there is simply too much to account for to truly have balance. In a game where skew lists are possible it is literally impossible to balance the game because it is not designed in a way that can be balanced. The assumption that you can determine the relative frequency to have meaningful playtesting is laughable. You cannot as the game is currently designed have meaningful playtesting because there is not enough time in the world to test out close to every possible match-up. The only way you can have anything resembling meaningful playtesting is that some options are garbage, and some are obviously good, so testers don't even need to touch some, and always use others. As I pointed out a tactical squad alone has hundreds of possible load outs, you really think anyone tests all of the combinations? And then does so along side every combination of other units in the book, or in the imperium faction?

For math to even begin to be usable GW would need to
1.) Define standard terrain layouts, or at least amount of table coverage, % of LOS blockage, % cover to an approximate level as for it to not be a large factor in unit effectiveness
2.) Define a points cost at which the game is supposed to be balanced (units are not going to be balanced at both 1000 points and 3000 points)
3.) Put more limitations on list building to curb skew lists- if you don't do this hard counters/bad match-ups to lists will basically always exist, because as a list designer you have no way of knowing exactly what you are going to face (especially if GW actually ever gets closer to cross faction balance, bad balance actually makes preparing for a meta easier as I only really need to account for the top several list builds)
4.)Reduce wargear options for units. Unless the idea is that one load out will be the "optimal load out" and everything else fluff or red herring having tons of wargear options makes balancing units much more difficult. A unit with few or no options will always be easiest to balance because you know exactly how it will perform against a variety of targets. Intercessors are easier to balance than tactical marines because you can make assumptions about how effective they will be against say a leman russ, where as a tactical squad with Melta gun, combi-melta, and multi-melta will perform quite a bit differently than one with flamer, combi-flamer and heavy bolter. And before you say that they have different targets in game, if a list of all russes exists both have to be balanced to face it.
5.) Either remove allies, or publish a statement saying factions will be balanced based on their entire ally set. Either marines are balanced standing on their own, or balanced assuming you are taking guard, you cannot have both be functional.

I still think the nay-sayers are being unimaginative.

One poster postulated that a unit had a certain value in a particular context, and another value in a different one. Good point. Would be impossible, for a given game board and mission, for a system to build two balanced forces?

The requirement isn’t that a system can be created that assigns a number of point to each thing that holds balance in every context, the requirement is that the game can be balanced.

I think I’ve found a mechanism to solve the context problem; context is just another input into the model. All this stuff takes is a little creativity.

With regard to the chess comments, I never said that chess was balanced. And of course rooks are better in the end game; we can prove this to be statistically true based on the results of thousands of games... just like I suggested we could for 40k (although with more games required to account for complexity).

Sorcererbob wrote:
I still think the nay-sayers are being unimaginative.

One poster postulated that a unit had a certain value in a particular context, and another value in a different one. Good point. Would be impossible, for a given game board and mission, for a system to build two balanced forces?

The requirement isn’t that a system can be created that assigns a number of point to each thing that holds balance in every context, the requirement is that the game can be balanced.

I think I’ve found a mechanism to solve the context problem; context is just another input into the model. All this stuff takes is a little creativity.

With regard to the chess comments, I never said that chess was balanced. And of course rooks are better in the end game; we can prove this to be statistically true based on the results of thousands of games... just like I suggested we could for 40k (although with more games required to account for complexity).

Well if we throw out the requirement to have balance across all options in all situations then the game is balanced now. Given any mission and game board I can build 2 forces that are balanced to each other. What I cannot do is do so with variety, or without knowing the table and mission, or having the ability to design both forces ensure any semblance of balance. Sorry simply put you are wrong because there are too many contexts that impact balance and too much player choice involved. I think you are not giving enough credit to how much more complex 40k is than chess as far as balance is concerned. I think you would need near infinite games to achieve balance without instituting as I said previously a bunch of rules and restrictions to shrink the number of variables. You can never build a game where a unit is balanced at 500 points and 5000 points. It is akin to saying chess would have the same balance if you changed the board size, or which pieces were available to each side.

Sorcererbob wrote:
I still think the nay-sayers are being unimaginative.

One poster postulated that a unit had a certain value in a particular context, and another value in a different one. Good point. Would be impossible, for a given game board and mission, for a system to build two balanced forces?

You realize there's infinite number of contexes? Every player has basically different terrain collection so it's impossible for GW to know how that affects. Then what about enemy army? That's something GW obviously can't factor in. Point value of your unit changes if opposing chaos army brings instead of 10 chaos marines 30 cultists. Or howabout 5 marines and 15 cultists...Then add in predator. Your units value changes again. Land raider instead? Again changes.

Good luck creating any sort of balanced formula. Only way to get balanced games would be fixed scenarios with fixed terrain(including shape) with fixed forces and looooooots of gametesting.

Stop. I think people are trying to dive “too deep” here and making to too complicated.

First off, a lot of people are saying you have to factor in terrain, scenario, mission, deployment, setup etc etc. You don’t need to factor in any of that in order to balance units against each other. Using the previously mentioned Revier Grappling hook as an example – it is an option you can spend points on currently. You spend these points prior to knowing the table setup (in theory) making them a strategic choice. They have nothing to do with the “balance” of the unit in terms of base costing.

Terrain and mission etc are all part of the strategic side of the game, the part completely in the hands of the players. This should not have an impact on the points costs of units. A predator is still a predator regardless of whether it has a clear los across the table or whether there are los blocking items in front of it. It doesn’t affect the “mathematical efficiency” of the unit, but rather the “strategic worth” of the unit in that given situation. It’s offensive and defensive power in the situations where it arises are what should drive balance calculations – not the “oh, for some reason player 1 put his tank behind a wall and now can’t shoot at anything without moving”.

Balance can be driven by setting a benchmark for various sets of things –
1. Basic stat line
2. Weapons
3. Abilities

For example, a Tactical Marine is the basic stat line for every infantry model in the Marine army. Once assigned an “initial” points value you can then move onto the next stage and determine weapon costs based on their stats and abilities etc etc. Obviously a lot of trial and error would need to be used to begin with. But, once you have your basic costing, you can then move onto another variant of Marine, a Terminator, for example. Using the standard Marine base cost, you can then add, or more points based on a perceived cost of additional stats. Maybe +1 wound is 10 points, +1 armour save is 5 points, and gaining an invuln save of x = 20 points. Of course these are just made up numbers right now, but it is the kind of basic logic you’d need to follow. All weapons would be pointed based off their stats alone, ignoring the “platform” they are on, like they are currently (i.e a lascannon has the same price regardless of whether it is a predator, dev unit, dreadnought or flyer using it).

Abilities will be more tricky to nail down. Things like re-roll 1’s is easy enough – it provides a x% increase in a models offensive output. This can then be based off assumptions that it is used in conjunction with 2 5 man units, buffing the effectiveness of 11 models in total. However, abilities like “character”, “fly”, “psyker” etc would all need to be played with to find a starting trial value.

It would be a massive amount of time and effort in order to point the first army, but then you'd have a model to quickly transpose to all the other armies. The issue then becomes tweaking the costs of each value to ensure everything is reasonable (i.e a Guard Infantry model isnt like 15 points).

Just to point something out about how many options there are an AM infantry squad has 1152 possible load outs. (2 vox options x 6 HW options x 6 Special weapon options x 4 sarge gun options x 4 sarge melee options), That is just a single unit. That means taking just 2 of those units gives you 1,327,104 possible different combinations.

So a single battalion with 2 company commanders(25 possible load outs) and 3 infantry squads has 955,514,880,000 possible combinations. So you need nearly 1 trillion games to even field all of the possible load outs, that is using 2 different units, and not taking opponents into account.

This is why I say that if you want to come close to balancing everything units need to have more limited options. Primaris Intercessors for example have 72 possible load outs (mostly squad size) (3 gun options x 6 squad size options x 2 sarge options x 2 grenade launcher options)

I had help with a doctorate in statistics and game theory who helped me write the formula we used for Azyr for AOS before the general's handbook killed the fan systems.

I will say with advanced mathematics that you can get pretty close to balancing a game, but not entirely.

Azyr used weighted averages. So the points cost of something was how well it performed against every possible combination and then averaged that number out.

The part where it failed were on the abstract synergies. Say dude gives +1 attack to every unit within 9" of him. You can't really account for this accurately because the actual mathematic increase will depend entirely on what units are within 9" of him.

If its a unit of scrubs with low grade attacks, the mathematic increase in performance will be X, but if you have a unit of high quality attacking models next to him, the performance will be X + Y. Coupled in that then you have to account for how many high grade units are in range at any given time.

Its impossible to do properly.

However you can get very close and then you have to have a fudge factor that covers the abstract abilities. Of course, no human beiing alive today could create a system that makes everyone happy. Someone will always gripe.

Even if a model was really worth 100 points, all it takes is for someone to play poorly, or get outplayed and that model gets tanked riight away and they are going to shoot your entire system down as being horrible and unbalanced because that 100 point model didn't return them 100 points liike they thought it should.

Kdash wrote:
Stop. I think people are trying to dive “too deep” here and making to too complicated.

First off, a lot of people are saying you have to factor in terrain, scenario, mission, deployment, setup etc etc. You don’t need to factor in any of that in order to balance units against each other. Using the previously mentioned Revier Grappling hook as an example – it is an option you can spend points on currently. You spend these points prior to knowing the table setup (in theory) making them a strategic choice. They have nothing to do with the “balance” of the unit in terms of base costing.

Terrain and mission etc are all part of the strategic side of the game, the part completely in the hands of the players. This should not have an impact on the points costs of units. A predator is still a predator regardless of whether it has a clear los across the table or whether there are los blocking items in front of it. It doesn’t affect the “mathematical efficiency” of the unit, but rather the “strategic worth” of the unit in that given situation. It’s offensive and defensive power in the situations where it arises are what should drive balance calculations – not the “oh, for some reason player 1 put his tank behind a wall and now can’t shoot at anything without moving”.

Balance can be driven by setting a benchmark for various sets of things –
1. Basic stat line
2. Weapons
3. Abilities

For example, a Tactical Marine is the basic stat line for every infantry model in the Marine army. Once assigned an “initial” points value you can then move onto the next stage and determine weapon costs based on their stats and abilities etc etc. Obviously a lot of trial and error would need to be used to begin with. But, once you have your basic costing, you can then move onto another variant of Marine, a Terminator, for example. Using the standard Marine base cost, you can then add, or more points based on a perceived cost of additional stats. Maybe +1 wound is 10 points, +1 armour save is 5 points, and gaining an invuln save of x = 20 points. Of course these are just made up numbers right now, but it is the kind of basic logic you’d need to follow. All weapons would be pointed based off their stats alone, ignoring the “platform” they are on, like they are currently (i.e a lascannon has the same price regardless of whether it is a predator, dev unit, dreadnought or flyer using it).

Abilities will be more tricky to nail down. Things like re-roll 1’s is easy enough – it provides a x% increase in a models offensive output. This can then be based off assumptions that it is used in conjunction with 2 5 man units, buffing the effectiveness of 11 models in total. However, abilities like “character”, “fly”, “psyker” etc would all need to be played with to find a starting trial value.

It would be a massive amount of time and effort in order to point the first army, but then you'd have a model to quickly transpose to all the other armies. The issue then becomes tweaking the costs of each value to ensure everything is reasonable (i.e a Guard Infantry model isnt like 15 points).

Sorry this is not true, the value of a lascannon is much higher if it can see every point within 48" than if it has a hard time drawing LOS more than 24" in most places, immobile artillery has a greater value in missions that don't involve claiming objectives or needing to move. Something can be great statwise, if it is bad at winning the game it is a bad unit. You mention abilities, the ability to ignore LOS for shooting has a different value on planet bowling ball than in city fight.


Automatically Appended Next Post:

 auticus wrote:
I had help with a doctorate in statistics and game theory who helped me write the formula we used for Azyr for AOS before the general's handbook killed the fan systems.

I will say with advanced mathematics that you can get pretty close to balancing a game, but not entirely.

Azyr used weighted averages. So the points cost of something was how well it performed against every possible combination and then averaged that number out.

The part where it failed were on the abstract synergies. Say dude gives +1 attack to every unit within 9" of him. You can't really account for this accurately because the actual mathematic increase will depend entirely on what units are within 9" of him.

If its a unit of scrubs with low grade attacks, the mathematic increase in performance will be X, but if you have a unit of high quality attacking models next to him, the performance will be X + Y. Coupled in that then you have to account for how many high grade units are in range at any given time.

Its impossible to do properly.

However you can get very close and then you have to have a fudge factor that covers the abstract abilities. Of course, no human beiing alive today could create a system that makes everyone happy. Someone will always gripe.

Even if a model was really worth 100 points, all it takes is for someone to play poorly, or get outplayed and that model gets tanked riight away and they are going to shoot your entire system down as being horrible and unbalanced because that 100 point model didn't return them 100 points liike they thought it should.

Correct me if I'm wrong but AOS seems to have fewer options for most units as well. Units are bought in blocks, and don't typically have much in the way of weapon options. In that case (except the mentioned synergy issue) it is much more attainable to work things out mathematically because there are fewer variables. IT seems like terrain would also have less of an impact because shooting is less of an issue.

Breng77 wrote:
Just to point something out about how many options there are an AM infantry squad has 1152 possible load outs. (2 vox options x 6 HW options x 6 Special weapon options x 4 sarge gun options x 4 sarge melee options), That is just a single unit. That means taking just 2 of those units gives you 1,327,104 possible different combinations.

So a single battalion with 2 company commanders(25 possible load outs) and 3 infantry squads has 955,514,880,000 possible combinations. So you need nearly 1 trillion games to even field all of the possible load outs, that is using 2 different units, and not taking opponents into account.

This is why I say that if you want to come close to balancing everything units need to have more limited options. Primaris Intercessors for example have 72 possible load outs (mostly squad size) (3 gun options x 6 squad size options x 2 sarge options x 2 grenade launcher options)

Why are you trying to balance all the "combinations"? Why aren't you instead trying to balance the individual weapons, individual options and the cost of a basic model without any weapons included?? If a meltagun is balanced with a plasmagun in terms of points then the 1000's of combinations won't matter as everything is balanced vs everything else.

I think you could get very close.

It might be impossible to balance entirely - there are limits to what you can do with a d6 system and discrete low point values.

But you could get it to a level where there were no objectively bad in all circumstance units and no objectively auto take in all circumstance units.

You could avoid the common issue of having two units in a roster that do the same thing, but one is clearly better in all situations than the other so one or both is never seen on tables.

Breng77 wrote:

Kdash wrote:
Stop. I think people are trying to dive “too deep” here and making to too complicated.

First off, a lot of people are saying you have to factor in terrain, scenario, mission, deployment, setup etc etc. You don’t need to factor in any of that in order to balance units against each other. Using the previously mentioned Revier Grappling hook as an example – it is an option you can spend points on currently. You spend these points prior to knowing the table setup (in theory) making them a strategic choice. They have nothing to do with the “balance” of the unit in terms of base costing.

Terrain and mission etc are all part of the strategic side of the game, the part completely in the hands of the players. This should not have an impact on the points costs of units. A predator is still a predator regardless of whether it has a clear los across the table or whether there are los blocking items in front of it. It doesn’t affect the “mathematical efficiency” of the unit, but rather the “strategic worth” of the unit in that given situation. It’s offensive and defensive power in the situations where it arises are what should drive balance calculations – not the “oh, for some reason player 1 put his tank behind a wall and now can’t shoot at anything without moving”.

Balance can be driven by setting a benchmark for various sets of things –
1. Basic stat line
2. Weapons
3. Abilities

For example, a Tactical Marine is the basic stat line for every infantry model in the Marine army. Once assigned an “initial” points value you can then move onto the next stage and determine weapon costs based on their stats and abilities etc etc. Obviously a lot of trial and error would need to be used to begin with. But, once you have your basic costing, you can then move onto another variant of Marine, a Terminator, for example. Using the standard Marine base cost, you can then add, or more points based on a perceived cost of additional stats. Maybe +1 wound is 10 points, +1 armour save is 5 points, and gaining an invuln save of x = 20 points. Of course these are just made up numbers right now, but it is the kind of basic logic you’d need to follow. All weapons would be pointed based off their stats alone, ignoring the “platform” they are on, like they are currently (i.e a lascannon has the same price regardless of whether it is a predator, dev unit, dreadnought or flyer using it).

Abilities will be more tricky to nail down. Things like re-roll 1’s is easy enough – it provides a x% increase in a models offensive output. This can then be based off assumptions that it is used in conjunction with 2 5 man units, buffing the effectiveness of 11 models in total. However, abilities like “character”, “fly”, “psyker” etc would all need to be played with to find a starting trial value.

It would be a massive amount of time and effort in order to point the first army, but then you'd have a model to quickly transpose to all the other armies. The issue then becomes tweaking the costs of each value to ensure everything is reasonable (i.e a Guard Infantry model isnt like 15 points).

Sorry this is not true, the value of a lascannon is much higher if it can see every point within 48" than if it has a hard time drawing LOS more than 24" in most places, immobile artillery has a greater value in missions that don't involve claiming objectives or needing to move. Something can be great statwise, if it is bad at winning the game it is a bad unit. You mention abilities, the ability to ignore LOS for shooting has a different value on planet bowling ball than in city fight.

Of course the “in-game” value of a lascannon is more if it can see everything, than if it can’t – but that ISNT a case of weapon and model balance, that is a case of strategy and game setup. You can’t attempt to “price in” a ruin vs a mountain vs nothing.

If a lascannon shoots a target it will do x amount of damage. If a lascannon doesn’t shoot because of a wall, it’s “average” damage per shot output doesn’t change. Saying that you need to “price in” players good, or bad positioning is stupid. That is an aspect of strategy and generalmanship, not model balance.

If a weapon has the “ability” to ignore LoS then obviously you would then point that weapon accordingly for that ability – not change the points of a weapon that cannot ignore LoS…
Winning a game is all about correct placement, correct target prioritisation and correct decision making in relation to the game at hand.

If a unit has a great stat line but works out badly in one given game setup, is it a case that the “unit is bad” or “the unit was USED badly” in the given game?

For example, a Shadowsword is FANTASTIC at deleting a tank a turn with it’s main weapon. But, if it gets killed first turn because you deploy wrong, or the other player got 1st turn and had 15 lascannons so it didn’t get to shoot at all, does that then make the Shadowsword a “bad unit” and therefore needs a points decrease? Or, is it just that the Shadowsword was played incorrectly? Or was it that the opponents list just happened to counter the Shadowsword?

Those instances aren’t a “balance” issue, they are a list building, deployment or genuinely unlucky issue. (the next game your opponent might have nothing to deal with the Shadowsword, so it completely destroyed their army… Suddenly it’s a good unit?)

Correct me if I'm wrong but AOS seems to have fewer options for most units as well. Units are bought in blocks, and don't typically have much in the way of weapon options. In that case (except the mentioned synergy issue) it is much more attainable to work things out mathematically because there are fewer variables. IT seems like terrain would also have less of an impact because shooting is less of an issue.

Shooting dominated the meta for most of AOS lifetime. The big things to min/max in AOS are mortal wounds (same as 40k) and high volume of ranged attacks. This is followed by high rending and high damage melee attacks.

Some of the units have just as many options as 40k units.

Same concept appliies. You have to get the unit baseline and then figure out how much a weapon increases that baseline and come to the weighted average against every type of defensive stat in the game.

Software modeling helps that a lot.

 Blacksails wrote:
I'm sure that technically you could mathematically balance 40k, but it would take way more time and effort than the classic methods.

I helped test and work on a game a friend was making that was heavily math based in the balancing department. He'd constantly track how many turns the game would run, when the first shots would fly, how much damage was done each round by groups of weapons, so on and so forth. The result was an oddly balanced point system, but it required an excel spreadsheet programmed with the formulas. Weapons would cost differently based on what arcs you put them in, the speed of the ship, and other secondary subsystems. It was interesting, but generally not feasible for mass consumption.

Better to just eyeball balance with 40k at first, then refine through playtesting and basic mathhammering.

Pretty sure there were no auras in that game.

Now I 've run the same model on 40k and will say that the point values in 40k are closer to the median than AOS is, but 40k is roughly 13% unbalanced in a skew kind of way. (that is to say 13% of the game falls outside of a median and either underperforms or overperforms)

Can you guess what most players see in all their games

For example: conscripts when they were released with their busted rules had a 27% skew. Thats huge (what the median is will always be debated but if I'm balancing a system I try to keep the skew between 5 and 8%). The last time there was a skew like that that I saw was 5th edition draigo and paladin buddies (roughly a 21% skew). So 8th edition has released probably the most busted unit of all time in conscripts before their nerf. (I haven't run all the data for most of the editions, but my eldar starcannon spam tournament 3rd ed army was 17% over performing for their points cost)

 koooaei wrote:

 Blacksails wrote:
I'm sure that technically you could mathematically balance 40k, but it would take way more time and effort than the classic methods.

I helped test and work on a game a friend was making that was heavily math based in the balancing department. He'd constantly track how many turns the game would run, when the first shots would fly, how much damage was done each round by groups of weapons, so on and so forth. The result was an oddly balanced point system, but it required an excel spreadsheet programmed with the formulas. Weapons would cost differently based on what arcs you put them in, the speed of the ship, and other secondary subsystems. It was interesting, but generally not feasible for mass consumption.

Better to just eyeball balance with 40k at first, then refine through playtesting and basic mathhammering.

Pretty sure there were no auras in that game.

Kind of? They weren't auras with straight 're-roll dice!' or 'add +1 to everything', but you needed sensors to successfully target ships and you could design some ships to be 'command' ships which could do the bulk of the targeting for other ships.

Another problem with mathematically-based point costs is that eventually you end up with a game of potatoes. No matter what you take -everything's gona be equally effective. You could take an army of potatoes and they'd be as good as a space ship cause the game's perfectly balanced.

 koooaei wrote:
Another problem with mathematically-based point costs is that eventually you end up with a game of potatoes. No matter what you take -everything's gona be equally effective. You could take an army of potatoes and they'd be as good as a space ship cause the game's perfectly balanced.

If you are seeking a balanced game then that's kind of the point. 2000 points is the same as your 2000 points. But yes that was something I heard a lot of with Azyr. That it was boring because everything was too balanced and the same. I find people don't really want balance. They want structure and then they want to be able to effectively min/max within that structure and get their 2000 points to operate like its 3000 points.

Kdash wrote:

Breng77 wrote:

Kdash wrote:
Stop. I think people are trying to dive “too deep” here and making to too complicated.

First off, a lot of people are saying you have to factor in terrain, scenario, mission, deployment, setup etc etc. You don’t need to factor in any of that in order to balance units against each other. Using the previously mentioned Revier Grappling hook as an example – it is an option you can spend points on currently. You spend these points prior to knowing the table setup (in theory) making them a strategic choice. They have nothing to do with the “balance” of the unit in terms of base costing.

Terrain and mission etc are all part of the strategic side of the game, the part completely in the hands of the players. This should not have an impact on the points costs of units. A predator is still a predator regardless of whether it has a clear los across the table or whether there are los blocking items in front of it. It doesn’t affect the “mathematical efficiency” of the unit, but rather the “strategic worth” of the unit in that given situation. It’s offensive and defensive power in the situations where it arises are what should drive balance calculations – not the “oh, for some reason player 1 put his tank behind a wall and now can’t shoot at anything without moving”.

Balance can be driven by setting a benchmark for various sets of things –
1. Basic stat line
2. Weapons
3. Abilities

For example, a Tactical Marine is the basic stat line for every infantry model in the Marine army. Once assigned an “initial” points value you can then move onto the next stage and determine weapon costs based on their stats and abilities etc etc. Obviously a lot of trial and error would need to be used to begin with. But, once you have your basic costing, you can then move onto another variant of Marine, a Terminator, for example. Using the standard Marine base cost, you can then add, or more points based on a perceived cost of additional stats. Maybe +1 wound is 10 points, +1 armour save is 5 points, and gaining an invuln save of x = 20 points. Of course these are just made up numbers right now, but it is the kind of basic logic you’d need to follow. All weapons would be pointed based off their stats alone, ignoring the “platform” they are on, like they are currently (i.e a lascannon has the same price regardless of whether it is a predator, dev unit, dreadnought or flyer using it).

Abilities will be more tricky to nail down. Things like re-roll 1’s is easy enough – it provides a x% increase in a models offensive output. This can then be based off assumptions that it is used in conjunction with 2 5 man units, buffing the effectiveness of 11 models in total. However, abilities like “character”, “fly”, “psyker” etc would all need to be played with to find a starting trial value.

It would be a massive amount of time and effort in order to point the first army, but then you'd have a model to quickly transpose to all the other armies. The issue then becomes tweaking the costs of each value to ensure everything is reasonable (i.e a Guard Infantry model isnt like 15 points).

Sorry this is not true, the value of a lascannon is much higher if it can see every point within 48" than if it has a hard time drawing LOS more than 24" in most places, immobile artillery has a greater value in missions that don't involve claiming objectives or needing to move. Something can be great statwise, if it is bad at winning the game it is a bad unit. You mention abilities, the ability to ignore LOS for shooting has a different value on planet bowling ball than in city fight.

Of course the “in-game” value of a lascannon is more if it can see everything, than if it can’t – but that ISNT a case of weapon and model balance, that is a case of strategy and game setup. You can’t attempt to “price in” a ruin vs a mountain vs nothing.

If a lascannon shoots a target it will do x amount of damage. If a lascannon doesn’t shoot because of a wall, it’s “average” damage per shot output doesn’t change. Saying that you need to “price in” players good, or bad positioning is stupid. That is an aspect of strategy and generalmanship, not model balance.

If a weapon has the “ability” to ignore LoS then obviously you would then point that weapon accordingly for that ability – not change the points of a weapon that cannot ignore LoS…
Winning a game is all about correct placement, correct target prioritisation and correct decision making in relation to the game at hand.

If a unit has a great stat line but works out badly in one given game setup, is it a case that the “unit is bad” or “the unit was USED badly” in the given game?

What if it is more than one game setup? And you keep saying player decision. Unless the player is setting up the terrain, it may not be their decision but instead the terrain that dictates the effectiveness for the weapon, which has nothing to do with generalship or strategy. If you are playing on a table where you can never use the maximum range of the weapon no matter what decision you make that matters, just as a table with more LOS blocking terrain favors units that ignore LOS versus one with little to no LOS blocking terrain. So how should that be costed if we don't know what the terrain will be? Do we cost it assuming it will be able to hide in a good position for its range, or do we assume that it will be a situational ability?

Same with the lascannon, if the table set up or mission dicates that it won't be able to fire at peak efficiency for say 3 turns of the game, why is it still worth the same as it would be in a game where it shoots 5 times?

Your method is the following - cost everything as if we are playing in its optimal condition, then if those conditions are uncommon those units now suck, and those that frequently have optimal conditions are great. Which is what we have right now.

 auticus wrote:

 koooaei wrote:
Another problem with mathematically-based point costs is that eventually you end up with a game of potatoes. No matter what you take -everything's gona be equally effective. You could take an army of potatoes and they'd be as good as a space ship cause the game's perfectly balanced.

If you are seeking a balanced game then that's kind of the point. 2000 points is the same as your 2000 points.

I agree, that is the point of balance.

However, due to different weapon options, units would still be different in terms of their efficiency vs different units. For example, a Space Marine with a bolter might be as “balanced” as a Devestator with a lascannon, but the lascannon will do better vs tanks, while the bolter will do better vs chaff. That is then where strategy and the table come into play, as it’d force you to attempt to correctly use the units, or suffer an impact on overall efficiency.

Kdash wrote:

Breng77 wrote:
Just to point something out about how many options there are an AM infantry squad has 1152 possible load outs. (2 vox options x 6 HW options x 6 Special weapon options x 4 sarge gun options x 4 sarge melee options), That is just a single unit. That means taking just 2 of those units gives you 1,327,104 possible different combinations.

So a single battalion with 2 company commanders(25 possible load outs) and 3 infantry squads has 955,514,880,000 possible combinations. So you need nearly 1 trillion games to even field all of the possible load outs, that is using 2 different units, and not taking opponents into account.

This is why I say that if you want to come close to balancing everything units need to have more limited options. Primaris Intercessors for example have 72 possible load outs (mostly squad size) (3 gun options x 6 squad size options x 2 sarge options x 2 grenade launcher options)

Why are you trying to balance all the "combinations"? Why aren't you instead trying to balance the individual weapons, individual options and the cost of a basic model without any weapons included?? If a meltagun is balanced with a plasmagun in terms of points then the 1000's of combinations won't matter as everything is balanced vs everything else.

Sure they do, because those things cannot be balanced in all situations unless they are identical. If they do different things, regardless of points costs then the combination of Melta + combi-melta + lascannon, has a different value than melta + combi-flamer + autocannon. Either all of these options are valid and balanced or, there is an optimal load out. Doing it by model ignores the impact of the unit as a whole. A single special weapon in a squad may have different value points wise to multiple special weapons.


Lets put it this way, doing it your way is going to end up as good enough, because you are never going to balance every possible option available.

Automatically Appended Next Post:

Kdash wrote:

 auticus wrote:

 koooaei wrote:
Another problem with mathematically-based point costs is that eventually you end up with a game of potatoes. No matter what you take -everything's gona be equally effective. You could take an army of potatoes and they'd be as good as a space ship cause the game's perfectly balanced.

If you are seeking a balanced game then that's kind of the point. 2000 points is the same as your 2000 points.

I agree, that is the point of balance.

However, due to different weapon options, units would still be different in terms of their efficiency vs different units. For example, a Space Marine with a bolter might be as “balanced” as a Devestator with a lascannon, but the lascannon will do better vs tanks, while the bolter will do better vs chaff. That is then where strategy and the table come into play, as it’d force you to attempt to correctly use the units, or suffer an impact on overall efficiency.

Which only works if there are limits on list building to avoid skew lists.

Not while maintaining unit differentiation and customizable armies. You're still going to run into the issue of Player A bringing only light machineguns while Player B brings only mainline tanks, or "All Snipers: The Army" having differing performance on Planet Bowling Ball vs. The Maze of Arbitrarily Tall Hedges.

In the context of a single product - e.g. the Death Guard half of Dark Imperium vs. the Space Marine half it should be achievable.

In a game where winning the single die roll for first turn can give a +40% win rate, good enough balance is probably fine. It's psychologically useful for players to be able to blame poor performance on luck or game design.

Kdash wrote:

 auticus wrote:

 koooaei wrote:
Another problem with mathematically-based point costs is that eventually you end up with a game of potatoes. No matter what you take -everything's gona be equally effective. You could take an army of potatoes and they'd be as good as a space ship cause the game's perfectly balanced.

If you are seeking a balanced game then that's kind of the point. 2000 points is the same as your 2000 points.

I agree, that is the point of balance.

However, due to different weapon options, units would still be different in terms of their efficiency vs different units. For example, a Space Marine with a bolter might be as “balanced” as a Devestator with a lascannon, but the lascannon will do better vs tanks, while the bolter will do better vs chaff. That is then where strategy and the table come into play, as it’d force you to attempt to correctly use the units, or suffer an impact on overall efficiency.

And here is where part of the "omg fix this mess - the game is not balanced!1" is coming from. Tacticals may be well balanced mathematically but if you're only facing tanks, it doesn't seem so to you.