7th Ed. Standardized Comp Score

[Caneghem]

I have been thinking about the issue of comp lately, and how it is usually very much arbitrary and each tournament tends to have their own rules. I've seen it handled all kinds of ham-fisted ways, where obvious prejudices against certain armies harshly penalize them while other armies are severely underestimated. When you start getting down to the level of micro-managing power dice and set artificial limits, it starts to get really obnoxious and cuts down army diversity.

So I got to thinking about how the problem could be fixed once and for all, and of course my solution involves crunching tons of numbers! Once this project reaches its completion however, all of the current army books would forever be assigned their own set of point values, and you'd never need to do them again (until the book was re-released).

What it boils down to is this: we have to "fix" the books GW has released. I don't mean that we change the actual point values of everything, that would be insane because of all the different point rules GW has in place. Each model firstly needs to be assessed one at a time, and be given comp values based on what they can actually do. This can be easily done by way of a simple computer program and someone dedicated enough to carry out the calcs.

You have to figure out how you are going to weight everything, and this as always is the most controversial part. If done correctly... list design suddenly becomes much more open-ended and fluff-friendly. For example, jungle swarms are pretty widely held as a "fluff choice", because of how they perform on the table compared to their points. Well, once the comp-weights kick in, it should be feasible to use jungle swarms so long as you can perform some tactically useful function with them, as in the long run they will help you win out by helping comp.

So all of the army books have these new comp values attached to each possible choice based on what they can actually do. Now comes the tricky part... developing a user-friendly application that can quickly generate the comp score of any list. If done right, the tournament organizer should be able to plug a submitted list into the program in a couple minutes and come out with an adjusted score.

So now, the TO has the comp weights of each list. The highest score is the hardest list, and it sets the standard for the tournament. This list has a value of "1". If someone turned in a list with a comp score that was half of the hardest list, they'd be at a "0.5"... and so forth. Scores for each battle are determined as follows:

VPs scored * (Opponent's Comp/Your own comp)

If you manage the following VPs against the player with the toughest list with your own list which is a .5, it would look like this...

(You scored 458 VPs, opponent 512)
You: 458 * (1/.5) = 916 final
Opponent: 512 * (.5/1) = 256 final

Now keep in mind, a .5 would be only "half as nasty" as the hardest list in the tournament. I expect that there will still be the charlatans who will intentionally build the weakest army in existence and just dance around and deny points as much as possible. An easy way to nip this in the bud would be to set a point value range for the tournament... so under this comp system you could set the point value of the tournament at 2000-2250 for example. This would prevent someone from turning up with a 500 point harassment force with the intent of taking out a couple targets then trying to grab for table quarters and bank on multiplied comp scores. Also, some lists that are just inherently hard like Daemons can opt to give up a few points to appease the comp gods, maybe taking 2100 points instead of the full 2250. The great thing is, the system would be so flexible it could handle it all.

This type of system would allow for players to use the full range of an army's models and still be "in the running" for generalship... so long as they know how to play the game!

Now... somebody start putting this into practice!

 

[strewart]

Wow... I kinda see what you are getting at, but that kind of system would be very very hard to figure out and judge. I think the value (comp value) of a unit would not only be subjective based on the person, but also dependant on how much of the unit is on the table. It isn't really linear if you see what I mean. 1 fanatic in a game should not be just 7 times less deadly than 7 fanatics, 1 engine is less potent than half of 2 engines.

There is an easier solution. You said that this isn't exactly changing the points values of units, but that is kinda what it boils down to. So really GW should employ a LOT more playtesting to the armies before the books are published, even to the point of releasing online beta versions to people. If not to everyone, then a large number of people around the world. Army books would still sell, and essentially GW is a model company not a book company, most of their money comes from the models not the books anyway.

Or of course leave torunaments completely unrestricted.. That is kinda what they are meant to be even though a couple of books are a lot more powerful than the rest and would dominate. Maybe lists could be judged on a case by case basis.

 

[Caneghem]

Wow... I kinda see what you are getting at, but that kind of system would be very very hard to figure out and judge. I think the value (comp value) of a unit would not only be subjective based on the person, but also dependant on how much of the unit is on the table. It isn't really linear if you see what I mean. 1 fanatic in a game should not be just 7 times less deadly than 7 fanatics, 1 engine is less potent than half of 2 engines.

You'd have to factor everything in, including multiples of the same choice. I did kind of gloss over this section, but it would be a ton of work figuring everything out. It would be flexible enough to go in the other direction as well... the cost of infantry models would go down starting at the 21st one you add to a unit, going lower and lower (26th, and at 30th). The cases you mentioned were examples of "more is always better", so that would need to be factored in as well. EotG would be pretty pricey to begin with, and the fact that they add power dice would probably mean that the second one would be significantly more costly than the first... because you would be greatly expanding your magic phase as well as putting down another great model. My thought on power dice is similar... you'd pay more for each power die you added. Someone with 2 PD will pay less per PD than someone getting to 6... who would pay less per die than someone going to 10... etc. Once you get into the "out of control" ranges, the power dice cost will be fairly exhorbitant... but still legal.

There is an easier solution. You said that this isn't exactly changing the points values of units, but that is kinda what it boils down to. So really GW should employ a LOT more playtesting to the armies before the books are published, even to the point of releasing online beta versions to people. If not to everyone, then a large number of people around the world. Army books would still sell, and essentially GW is a model company not a book company, most of their money comes from the models not the books anyway.

Getting GW to change its ways doesn't strike me as an easier solution... in fact it seems like GW is going in the opposite direction. A recent Standard Bearer was singing the praises of a gaming group who liked to just "fudge the rules" and it pretty well expressed an attitude of "you're lucky we're making rules at all". Desiring a system in which competitions can be fairly held and skill be the deciding factor is decidedly OUTSIDE the realm of GW's thinking... and to an extent there is nothing wrong with that. I'd wager that the bulk of GW's customers are not hardcore tournament types.. and think the rules are just fine how they are!

Or of course leave torunaments completely unrestricted.. That is kinda what they are meant to be even though a couple of books are a lot more powerful than the rest and would dominate. Maybe lists could be judged on a case by case basis.

This is an ok solution, but this becomes a game of turning in the hardest list the judges won't reject. When it comes to game time, things are even more lopsided as some people don't "work the system" and those that had their list rejected a couple times, made a couple cuts, and just made it under the mark... well they have the toughest lists in the tourney and suffer no composition penalties. And again, personal preferences and biases rear their ugly heads... you'd be surprised at what lists get let through!

 

[slannfrog]

Does anybody have the GW Comp Rules/Calc Format LINK for tournaments???

Thanks.

 

[Barotok]

I've been thinking about the feasibility of such a system, not only to standardize comp scores but to judge the strength of individual units on a slightly smaller scale then widespread multi-army equation. These of course are two seperate topics but it would stand to reason that the massive amount of complex equations required to standardize judgement of compostion would be broken down to, at some point, individual units and models. I'm not sure I have the skills to assist in such an undertaking but have been playing with some formulas for basic calculations to judge the strength of a unit, given its size, static CR, frontage and average expected wounds dealt. As you can see, these smaller formulas would be a factor in larger equations that would calculate in to a comp score. I'm not sure of the details though. Thoughts?

 

[Caneghem]

I've been thinking about the feasibility of such a system, not only to standardize comp scores but to judge the strength of individual units on a slightly smaller scale then widespread multi-army equation. These of course are two seperate topics but it would stand to reason that the massive amount of complex equations required to standardize judgement of compostion would be broken down to, at some point, individual units and models. I'm not sure I have the skills to assist in such an undertaking but have been playing with some formulas for basic calculations to judge the strength of a unit, given its size, static CR, frontage and average expected wounds dealt. As you can see, these smaller formulas would be a factor in larger equations that would calculate in to a comp score. I'm not sure of the details though. Thoughts?

This is pretty much what I was thinking, it would revolve as much as possible around calculations.

Each model type would be put through its paces... and I am thinking the best way to do this would be to create a series of benchmark targets. Basically you run the numbers of a matchup of the unit in question striking first against a target, then in striking last against the target. To give a very rough example... let's say a block of 20 dwarf warriors meets a block of 20 saurus spear-warriors. Here's what would happen...

Saurus charge: 10 attacks yield 1.25 kills.. 3.75 dwarf attacks back yield 0.42 kills = result +0.83
Dwarves charge: 5 dwarf attacks kill 0.55 Lizards 18.9 saurus attacks back gives 2.36 kills = result +1.81
Second round of combat: 5 dwarf attacks kill 0.55 Lizards 18.9 saurus attacks back gives 2.36 kills = result +1.81

Avg result for dwarf warrior benchmark = +1.48

Now imagine arranging a series of these benchmarks and having a computer program do all the work... that's pretty much what I'm thinking. Given a broad range of "test combats", this would give a fair assessment of a unit's overall combat prowess in terms of weapon skill, strength, toughness, wounds, initiative, and attacks, as well as all forms of saves in combat.

The trickier things will be movement, leadership, etc. These would also have to be weighted based on unit type. I'm thinking movement would have a multiplying effect on the combat result score... so a unit that performs awesomely in combat while also having great speed would score a good bit higher. This would need to be evaluated at the "combat role" level.

You'd have to have different benchmark sets depending on what sort of unit you are looking at... broken down into "combat roles". You wouldn't evaluate skink cohorts using the infantry benchmarks... as then they'd be insanely negative in terms of comp while still being an excellent choice! Perhaps a fair way to do it would be to run all applicable benchmarks... to cover all normal uses for a unit.

Similarly you'd have shooting benchmarks... with special rules like move and shoot and the like having a multiplying effect on the score much like movement and combat. Move and shoot also extends the threat range greatly. A unit that can shoot 18" and move 6 would have a threat range of 24 inches in front and 22.5 inches to the rear... while the same unit without move and shoot has merely an 18" range in front, a difference of 28.5 inches if you're counting...

Eventually though, you could work out what you deemed to be a "fair" set of benchmarks. Only after the setup is settled upon do you start running the different armies through the paces, and in that way you eliminate your personal bias.

 

[Barotok]

It sounds to me like you'll be setting up a massive formula based on several, possibly 100s of smaller parts. While that is the only way you could hope to standardize comp score there are a few problems. Firstly, the way the forumlas are calculated will create a bias towards certain comps. The example that comes to mind is the skink cohorts once again. While a miserable failure vs. any kind of elite infantry, their mobility gives them the advantage. This gives them multiple uses, IE some players will use them for things and other players will use them for others. This makes judging their value highly difficult because there is no standard by which to judge.

Another example would be if I build a cannon heavy rock-paper-scissors army of say, dwarves, but can't fire cannons accurately due to the fact that I am a terrible distance judge, the potential for damage goes way down, when I am playing the list. Compare that to someone who is an eagle eye with distance judging. This player would maximize the effectiveness of the damage potential of such a list. What I am trying to say really boils down to the fact that some comps are frowned upon because of how easy they are to effectively play while others, which are more difficult don't get a second look, unless a skilled player is using them. This is of course a generalization, but I think you can understand what my point is.

Secondly, measuring a units damage potential against similar types of units (infantry vs. infantry) seems to add a bias all of it's own. With a game that offers great advantage to players who are able to positionally exploit a troop mismatch, direct comparison of similar troops will ultimately tell you which ones are better. For example, my sarus with HW+SH are not going to last against a similarly geared unti of chaos warriors, toe to toe for several rounds. In fact, doing the math, you'll find that the warriors (provided they charge) will win by an average of 3.1 leaving you with a 52.3% chance of making your break test. So I am not sure if there is a fair set of benchmarks that could be applied as a blanket to the game as a whole.

Perhaps instead of measuring individual units and comparing them to others a different system could be utilizied in which units would be measured for their effectiveness in combat, but composition would ultimately be the determinant factor in the score. Unless I am grossly misinformed comp scores relate to how 'cheesy' the army is to play against. I've seen tourneys where the opposing player would give you the comp score for your list and seen tourneys where it's measured in all different ways. The bottom line is that players who use an extremely magic heavy list will generally get a lower comp score as an example. This could be true for shooty armies or any other style rock-paper-scissors list. The reason I refer to them as such is that some armies have the tools to effectively deal with high magic while others simply can't. Some stand up to shooting better while others crumble. Those low comp lists are generally no fun to play against, especially if you don't have the tools in your list to effectively deal with them and make for a crushing one sided game. However, the next player who has the counter to that list crushes our magic heavy army because he does have the tools.

I'd venture that measuring individual units for their effectiveness on the battlefield in terms of mobility, damage out put and combat resolution is a great start for such complex calculations. However, after that a totally unbiased system would favor a balanced list in which shooting, magic, avoidance and any other factor which is 'unfun' to play against when your opponent has it in excess would lower your score not in a linear fashing but an exponential fashion. So the more magic that you stack in your list, the worse the score gets and each respective magical addition adds an increasing penalty.

Here's an example of what I am talking about. Going back to our friend who decides to dump a ton of cannons or artilery into his list. For this example, the higher the comp score, the worse it is. The closer to 0 you are, the more balanced your list is viewed as.

Infantry unit: 100 CP
Cavalry unit: 200 CP
Artilery unit: 150 CP

Our player has 2 infantry units, 1 cavalry unit, and 6 artilery units.

Infantry CP: 200
Cavalry CP: 200
Artilery CP: 900
Total Comp Score: 1300

Not a balanced list at all! Now lets add a penalty for having excess number of units of a certain type. Unit CPs increase 1.5x for every additional unit. Lets do the calcs.

100CP + 150CP = 250CP for infantry
200CP = 200CP for cavalry
150CP + 225CP + 337.5CP + 506.25CP + 759.3CP + 1138.9CP = ~3117CP for artilery
Total Comp Score: 3567

Ugh, thats a high number and you can really see where adding so much artilery to the list increases the comp score to the point where the 6th and final piece is over 1100. Remember, the lower the score, the better for this excersize.

Now lets say that our cannon toting friend has a change of heart and swaps 3 of his artilery units for 2 cavalry units and a third infantry unit, bringing all the categories to an even three each, which in this example is 'balanced'. Lets run the calcs again and compare.

100CP + 150CP + 225CP = 475 for infantry
200CP + 300CP + 450CP = 950 for cavalry
150CP + 225CP + 337.5CP = 712.5 for artilery
Total Comp Score: 2137.5

So cannon head takes 1400 points off of his comp score in favor of a more balanced list, which will ultimately be more fun to play against from every opponents point of view. Will he excell in the shooting phase anymore? No, not any more than the next guy's army but he is now not totally screwed if someone shows up with a list that's designed to counter a 'shooty' list.

So in conclusion, you can take an individual unit's combat prowress and equate it to a number, add those numbers together with a multiplier in there which is biased only against unbalanced lists and you have a comp score which is standardized only by numbers. Obviously, there are numerous extra rules and calculations that need to be added in order to help certian armies balance (you can penalize a TC or VC player for going magic heavy or a dwarf player for having too many guns) out because of what they're natural tendencies are.

Hopefully, you can use some of the ideas here as a framework or skeleton for what you're looking at doing. If you need help, count me in, as I'm very interested in something like this.

 

[Caneghem]

Wow, great to see lots of feedback Barotok. I'll give it a full read tomorrow but I noticed a few things on skimming... EDIT: I wound up giving a full reply... don't worry, I won't add to this Berlin-text-wall!

1) range guessing
Cannons and the like will be evaluated based on the "ideal guess". As a former dwarf player, I've seen folks much better at statistics than myself actually put out graphs of cannon accuracy based on the ideal guess, and they had it in the context of using the different dwarf runes to improve efficiency. Tracking down these numbers for how often different warmachines will hit will make their average damage easier to calculate. I can't even begin to fathom the numbers for a stone thrower with that scatter in there... but again it will start with the concept of the "perfect guess" dead center in units or on top of monsters.

2) skink cohorts
All possible unit roles will be evaluated, so cohorts would be evaluated on their shooting AND combat ability. Units will be given a value based on their overall potential, and it then becomes the responsibility of the player to fully utilize that value. No doubt the little skink units are worth more than their 50 points would indicate...

3) Break tests, Leadership, and movement
These are, as I said, the harder things to quantify as they are really situational. Movement allows for more dictation of combats, and so it will carry a price tag its own that will vary depending on the other things the unit can do. I'm still mulling this part, but straight up evaluating damage potential and factoring in all relevant combat stats seems like a logical first step.

4) Redundancy
Certain things in the game just get better and better the more of it you take. Magic is an example of something that gets exponentially better, since beyond a certain point you will totally overwhelm your opponent's defenses. There needs to be a redundancy multiplier involved that increases the price of power dice without setting an absolute cap. My goal is to create a system that has no hard caps. So you go up against someone with a magic spam list with your balanced list. The reason this is usually no fun is that you can't win by the rules given in the warhammer book. However, with standardized comp, your balanced list might be sitting at around a .5 while he may have the "hardest" list and have the 1. This would mean your VPs are doubled at the end, while his are halved. This makes victory possible, and it also makes victory more challenging for the magic spammer. For the sake of completeness, each army needs a pricing scale for its power dice, as some do well enough with a few while some (like TK) pretty well rely on having a lot to begin with. Magic isn't exactly an easy thing to put a price tag on, since the variety is endless. The fact that magic is so very flexible will increase its price tag significantly. I agree that too much magic takes some of the tactical element from the game, but my goal is to factor in the price well enough so that such a list is still allowed but it will demand near-perfect play from the magic user.

Shooting is quite similar in that the more tends to be the merrier (for the shooting side anyway). Again, the penalty for repeating such units must be evaluated on a case-by-case basis, and you should see similar, but probably not as harsh, penalties for really laying shooting on.

5) "However, after that a totally unbiased system would favor a balanced list in which shooting, magic, avoidance and any other factor which is 'unfun' to play against when your opponent has it in excess would lower your score not in a linear fashing but an exponential fashion."

An unbiased system would not favor any particular list over another, and that is one of the prime advantages of doing a complete evaluation the way I am suggesting. The cost of increasing magic will be exponential in most cases, since it has essentially no diminishing returns (wheras shooting is limited by LOS and board space at least!) Shooting will be similar, but probably not as severe in most cases. For dwarves salivating at the idea of lining up shield toting thunderers and still getting decent comp, don't forget that they perform well in combat too! As to things like avoidance play, I don't want to eliminate certain play styles completely, as they add flavor to the game. But I do think that all units that can use the old "shoot and run" tactic will have increasing costs for redundancy. A balanced list that has some shooting and some troops will probably wind up with a lower comp ratio, in which case your own shooters can go after some of the "avoiders" while your foot troops isolate and kill a few other of the buggers. Your multipier should see you through.

Essentially the plan is to go through each unit and...
1. Apply common sense to what you can expect that unit to do in the game
2. Run numbers and predict expected damage to establish relative efficacy of the unit in each category
3. Evaluate whether adding a copy of that unit should be cheaper, the same price, or increasingly expensive

How this system could revolutionize the game, and incorporate fluff into tournaments
In war, it is very rare for both sides to be equal. GW has offered up an illusion of equality by assigning ballpark point values to things, but all this does is create more unfairness than if players said "I'll take one of these, you take two of those", and just balanced against each other using common sense. Most players realize that the point scale just doesn't handle the realities of the game well enough.

Now imagine a tournament where you show up and NONE of the armies are pronounced as equal, but where victory is still totally dependent upon the skill of the player? That would make for much more fun and memorable games which would ironically be closer to the spirit of what GW intends than current tournaments.

Say your roving band of chaos warriors (balanced list) happens upon an envoy of some of the most powerful wizards and machines the Empire has to offer? Do the Chaos gods not smile upon you more for having attempted such a mismatched fight? Scaled points will make this scenario still enjoyable as the Chaos side would not need to kill as much to gain victory.

Or say for example you assault a fully armed dwarf gunline with your invading force of dark elves... isn't it typical of invading forces to take double the casualties of the defender in most scenarios? Even chipping away at such tough defenses would be a feat worthy of being called a victory... assuming for the sake of fluff that the rest of the invasion force is waiting to come in and exploit the hole in the dwarf line you heroically created.

Currently, fluff must be forgone for the sake of "fairness". If this project were ever completed, it would inject a whole new life into the tournament scene by allowing for themed lists and most importantly VARIETY. You wouldn't so much consider "I can take this and this and still get good comp". Instead you could consider what aspects of the game are your strengths and weaknesses, and build a list that is designed for your play style. It might not involve the most powerful units of the army, but the background points will mean that you can take pretty much what you want.

Wow, that went long... too much caffeine late at night for me!

 

[Barotok]

As far as range guessing is concerned I was using it as an example to clairfy my point despite the fact that it is a legitimate concern. It seems like most players using lists involving heavy artillery seem to be proficient at their use, and you'll certainly find that, at the tournament level, players will be highly accurate with their usage of these weapons. I'd go as far to venture that the only variables present in these situations will be the dice/scatter being as I've seen players perform startlingly complex mathmatical calcuations based on previous movement of a unit, where the artillery piece sits in the deployement zone and the angle at which the shot is fired.

As far as those units, like skink cohorts, who seem to fall between the cracks on specific use perhaps it would be fair to judge each unit for all its potential uses, rendering an overall score for that unit. Non-shooters will score a 0 in the shooting department, monsters score 0 in the skirmishing departement, flyers score 0 in this department etc. If you're going for an overall equal and unbiased rating system, treating each unit the same will most likely be instrumental in keeping it balanced.

On break tests:
Break tests are a function of leadership, which should be the ultimate measurement. Evaulating units based on their leadership is a simple probability calculation and the extra rules regarding break tests, such as battle standards, general's leadership and anything else may be disregarded considering that each army will most likely employ all of these things (IE BSB & General).

On Movement:
You mention that movent allows for dictational controll of combats. As a general rule of thumb this is true. If you're HEs are facing off against my sarus, I will most likely never get a successfull charge off. However, rarely, other troops like perhaps a small unit of skink cohorts, sets up a bait charge that allows for my unit of sarus to get an ideal charge off on a faster enemy. Given this won't happen frequently in tournament play but with careful placement of units can help dictate wether or not an opponent willingly takes a charge at which he is at a disadvantage because he is forced to choose between the lesser of two evils. Also, don't forget about the "Waaaagh" ability that the O&G army has. From a pure numbers standpoint, extra movement once a game isn't a huge advantage unless it's coupled with the strategy of a clever player (IE O&G infantry can get a charge on a faster unit). With a rule like this, I just consider my Sarus slower than Orcs, despite the fact that their movements are equalled on paper.

On Redundancy:
I agree that the penalties for redundancy should be different for different tactics. Capping doesn't seem like a good idea, as you said, at all to me either considering the fact that the rules do not specifically state any kind of cap for power/dispell dice. I'm not sure I am on board for adjusting the value of magic for each individual army list considering that would require a lot of weighing and most likely test playing to get right. Perhaps a simpler approach would be to generalize the value of basic WH magic with specific weightings for armies that have army specific magic (IE TK, VC, SKA, etc).

Also, an ultimate goal of being able to derive modifide VP calculations based on a comp score would require an already balanced comp scoring system that produced consistent and predictable results based on a variety of scenarios. Being as buildling a standardized comp rating system is such a large undertaking it may be best to focus on generating consistent results from the calculations being used before trying to equate that to VP multipliers, with the end results goal to produce a comp score and not modifying VP multipliers.

In short, get a working comp score system that produces solid results before trying to scale VP multipliers seems like the easiest route to go.

Two final comments that I was thinking about at work today:
1) This kind of project is going to deal almost exclusively with averages and therefore will not accurately represent the sometimes one sided and unexpected outcomes on the battlefield.
2) Games Workshop's games are admittedly always going to be out of balance, points wise. This almost sounds like an attempt to fix GW's mistakes in some areas, in regards to points cost, since you're giving eveything a weighted value based on what it is capapble of on the field.

 

[Caneghem]

"I'm not sure I am on board for adjusting the value of magic for each individual army list considering that would require a lot of weighing and most likely test playing to get right."

If only all schools of magic were created equal! Most armies that can take just plain vanilla wizards that choose from the schools of magic.. they'll all be scored pretty much the same. Once you start getting into army lists with crazy magic phases and unique spell sets, you really just need to go case by case.

"Also, an ultimate goal of being able to derive modifide VP calculations based on a comp score would require an already balanced comp scoring system that produced consistent and predictable results based on a variety of scenarios. Being as buildling a standardized comp rating system is such a large undertaking it may be best to focus on generating consistent results from the calculations being used before trying to equate that to VP multipliers, with the end results goal to produce a comp score and not modifying VP multipliers. "

Sometimes it is fun to think about what will be possible when it is all working. Unlike GW, this project would need to involve heavy amounts of testing and debate so that fair values can be established.

You may have missed in my original post on the thread, but what I'm aiming at is basically a point total for each list based on whatever scale seems most useful. I'm thinking more along the lines of going for a larger scale to better reflect subtle differences. Not all 8 point infantry models are created equal, and rather than get into the 7.9, 8.0, 8.1 decimal nonsense, it will probably being on a 79, 80, 81 type of scale. So some lists might total up to 22000 points and others 20400 points and so forth.

The application for tournaments would be to simplify these points as follows: the highest list, say it is 25000 points is given a comp of "1" for the tournament. Everyone else's comp would then be determined as a percentage of that score. So for example 22000/25000 = 0.88. That percentage can be called your "comp score" for the tournament. It also helps to iron things out so there is a smoother penalty/reward scale. Say the system isn't ever totally perfect, and your list probably should get a .86 but you wind up with a .87. It gets to be so close to the same penalty/benefit that you don't get totally shafted. Some comp systems use "tiers", but there is no way to assure equality even within tiers. It is also less significant of an error than you'd get in a 1-10 rating system or some such thing.

"In short, get a working comp score system that produces solid results before trying to scale VP multipliers seems like the easiest route to go."

In total agreement there, I was just salivating at what becomes possible when the point values of units are finally ironed out.

"1) This kind of project is going to deal almost exclusively with averages and therefore will not accurately represent the sometimes one sided and unexpected outcomes on the battlefield."

The game of warhammer will still use dice! There will always be flubbed attack rolls and insane courages. But these add flavor to the game and only the worst kind of criminal would seek to extricate them. I guess the simple fact is... they can and do happen to everyone, so it is already balanced out. Unless you mean something different?

"2) Games Workshop's games are admittedly always going to be out of balance, points wise. This almost sounds like an attempt to fix GW's mistakes in some areas, in regards to points cost, since you're giving eveything a weighted value based on what it is capapble of on the field."

That is precisely what must be done. The theory here is "aim small, miss small". In seeking to be very precise on how each individual unit is evaluated, the inevitable errors in judgement will be much less noticable, and on the whole lists should have fairly accurate values.

BTW: This thread wins the prize for fewest posts with the most text!

 

[Barotok]

Still in thought over this overall concept which perhaps should be renamed to "Rebalancing Warhammer"?

All joking aside, I have been trying to come up with a logical way to determine a unit's DPR(damage/round) in CC, pulling from a wide range of ideas. Some have been testing their combat skills against an enemy of equal stat lines (itself), but this does not generate varied values that seem like will be necessary to differentiate units.

With my understanding of math, there is really only one other option which is to use one troop type to measure all combatants effectiveness against. The next logical question in that line is, what troop type?

Perhaps my sense of logic is skewed but the average man seems like the best measurement benchmark for a basic value such as DPR.

Perhaps someone with greater understanding of math could shed light on what other method(s) would be better used for this calculation.

*Edit*
Should have prefaced this message with a statement indicating that this is but one of the smaller pieces needed to perform accurate comp calculations.

 

[Celticfire]

Borotok, good idea, but you would need to test more than just humans. here is what i can do to help with this problem you are creating.
1) choose the race/unit being tested
make sure that the unit that you test is tested against these following 3 armies, as they will act as a control, giving you a more acurate result

2) choose the types of combatants

Humans- probably the most neutral/basic army in the game
High Elves- at this point in the calculations ignore the ASF, so they are a weaker higher WS army, but still the stat line is pretty normal
Orcs- not the goblins, just the higher T orcs.

Infantry-choose the basic core troop from each army, Elven spearmen, orc boyz, and whatever the human is, im not sure i think it is also a block of spearmen
Cavalry-if possible choose the heavyer cavalry from each section ummm, boar boyz, dragon princes, knights of the circle?
Charoits-boar chariot, whatever is the more popular Elven chariot, and the Humans may not have one, which is fine
you get the general idea, test your unit against every type of troop from each of these armies, you should not have to go to the extremes, such as testing against monsters, flying and some of the more specialized things out there. all armies have infantry most have cavalry; these will be your two biggest spots that need to be tested

3) mathhammer it. ignore special rules at first, just do the stat lines of each type of troop.
assume
frontal charge
15-20 man units
handwepons and shield
full command
from there add in any special rules that might pertain,change some variables such as two hand wepons, size,flank/rear chrges, add in a basic charecter, but be carefull that that he is not too specialized as that coulsd skew the results.
as to a magical unit that you can test everything against, there is none, each troop type is different and is played diffrently. to both of you in your endevors, i wish you luck,
EDIT: thanks

 

[Barotok]

Forgive me for getting off subject for a moment, but I read somewhere on the forum already that a continously scrolling text window while posting can be fixed by changing browsers. I'd link you the thread, but can't remember where it is.

 

[strewart]

Yeah it appears to be a problem with internet explorer not the board, and there isn't anything we can do to fix it. I have been unable to emulate the problem myself in Firefox. Here is the old thread discussing it or a similar scrolling problem:

http://www.lustria-online.com/threads/post-size-limit.1440/

 

[Barotok]

I run IE and have never had that problem despite the average length of most of my posts.

Celticfire, you are outlining the logistical steps to summarize combat values in real conditions which is good. However, to standardize scoring, individual models need to be added together for a total value in a way so that different units may interact with one another by simple addition. Trying to weigh factors like, additional characters within the unit will simply bog down calculations forcing multiple combinations of units and characters due to special rules. Perhaps I misunderstood your post but it seems like you are wanting to calculate multiple factors in the same calculation which ultimately won't allow for the many combinations of units that people seem to create when making lists.

I also was thinking about special characters. Because of the extreme rules that frequently come along with these beasts, it makes sense to start calculations with out these factors, adding them in later. Obviously these characters come into play, but should add significant weight to an army, considering their abilities.

 

[Jive Professor]

I think the real problem is that these sorts of things never reallly work. To combat the subjectiveness of most comp systems, we create an even more complicated, and still subjective comp system?

Lists just need to be taken on a case by case basis. Judging units by themselves in a vacuum doesn't really solve much, as units always work in conjunction with the rest of the army.

 

[Barotok]

Does GW have a subjective system by which they use to determine point values for their models? How would creating one based on the math behind each unit be more complex than what GW is using currently? When I read through a new army book, I often wonder how they come up with their values. I'm sure they do use the math, but I just tend to wonder how much.

You're right about the fact that judging units in a vacuum doesn't solve much, but it tells you about their individual performance as compared to similar units in other lists. While certain things are difficult to calculate, I'd say that a good percentage of the calculations will start with models and the units they are part of. This creates smaller building blocks towards a bigger picture.

 

[Jive Professor]

Does GW have a subjective system by which they use to determine point values for their models? How would creating one based on the math behind each unit be more complex than what GW is using currently? When I read through a new army book, I often wonder how they come up with their values. I'm sure they do use the math, but I just tend to wonder how much.

You're right about the fact that judging units in a vacuum doesn't solve much, but it tells you about their individual performance as compared to similar units in other lists. While certain things are difficult to calculate, I'd say that a good percentage of the calculations will start with models and the units they are part of. This creates smaller building blocks towards a bigger picture.

I am not denying that GW is using what is essentially an entirely arbitrarily and subjective system to come up with points values. I am saying adding ANOTHER complex layer of mathematical mumbo jumbo will not solve anything, just create new problems.

Broad hard caps and specific lists need to be in place for ease of use, not complex logorithims. No more than 10 PD, no more than two of the same Rare, no more than 4 war machines, no special characters, etc.

 

[Barotok]

I'm still not totally convinced that it will create more problems. When coupled with the hard caps you're talking about, which I have thought about several times, but failed to mention, I think that it's possible to normalize most lists down to raw numbers for at least a comparison.

I will agree that as you increase one aspect of the list, the worse it gets for composition. Too many warmachines, magic, shooting and special characters are great examples of what can make the game grossly imbalanced, and potentially play to game ending results for lists with strengths that prey on others. I'd be interested in seeing if after removing all of those extra numbers, if there would be anything left to normalize. What I mean is, if limits were placed on certian aspects of armies, would the numbers not really tell us much?

 

[Celticfire]

Celticfire, you are outlining the logistical steps to summarize combat values in real conditions which is good. However, to standardize scoring, individual models need to be added together for a total value in a way so that different units may interact with one another by simple addition. Trying to weigh factors like, additional characters within the unit will simply bog down calculations forcing multiple combinations of units and characters due to special rules. Perhaps I misunderstood your post but it seems like you are wanting to calculate multiple factors in the same calculation which ultimately won't allow for the many combinations of units that people seem to create when making lists.

Sorry you missunderstood me, i said that if it was needed to test charecters, and other variables, that that it could be done, but that it should be done after you a\have completed the basic comp value for the unit. otherwise, as you sais it would become too confusing.

I also was thinking about special characters. Because of the extreme rules that frequently come along with these beasts, it makes sense to start calculations with out these factors, adding them in later. Obviously these characters come into play, but should add significant weight to an army, considering their abilities.

i have never been to a tourny that has ever allowed special characters. i assume you are making this list for tourny use. i may be mistaken, and they are frequently played in other ares, but from my experiance, they are too imbalenced to attempt to try to find a comp for.

EDIT: Sorry about that