The Issue: Inconsistent Ground State
There has been talk within the balance forum of improved gear resolving certain balance issues. I am now going to go into the problems with Warhammer's character progression model and specifically power creep, as applied to balance issues, and why I will not allow current balancing efforts to continue if there is no solution to the issues I raise.
Failure to satisfactorily resolve the problems presented in this topic (and thus using the old model) will result in the balance forums being locked until endgame gear is completely implemented.
Warhammer's character progression model is an absolute and vertical one. Stats increase, abilities are gained, item procs are added, renown abilities specced for. This results in the base state of the game changing - absolute levels of power increase and this affects a wide number of variables, like DPS/HPS, TTK, move speed, cast time, critical hit rate, etc.
Changing any of these variables significantly is a change in general game state and thus affects the balancing of the game. This results in the game being lifted and twisted off its base design, and the design begins to both underperform and become much more difficult to work with.
Unless two characters with power tier X perform in the same manner, down to the last detail, against one another as two characters with power tier Y, the game has changed and so has the balance. In this way, any kind of customisation is the enemy of balance, because it increases the number of variables which a balancer needs to accomodate for and thus significantly increases the difficulty of the exercise.
If power gain is taken to its extreme of min/maxing, it creates extremes of game state which are even more difficult to accomodate without sacrificing the viability of a class or some abilities. The most obvious examples are:
- Extreme damage (see: Unshakable Focus for an example, though not a power creep example, of what extreme damage can do both to boost and to hold back a class)
- Increased contribution of items to stats, which reduces the worth or impact of other elements of the game (such as weapon DPS and abilities) which then end up contributing relatively less to their stats and can be discarded.
The effects of these factors and others associated with min/maxing and state change in general are wide-ranging, and a full analysis of them goes beyond the time I have here and, truth be told, my skill in explanation. Suffice to say that the possibility of power creep troubles me enough, as a balancer, that I will speak plainly and say that I want it stopped. If it cannot be stopped, I believe our ability to balance all aspects of the game, as well as make more interesting concepts work, will be severely compromised. I hold up Engineer and Magus range as an example.
Now, these points should be accepted straight off, but just in case some people do not understand the above without being presented with examples, I'm going to go through examples of how unchecked and absolute power creep can imbalance a game.
Example 1: Critical Hit Rate
Many of you know I'm an enemy of crit stacking. Crit stacking changes the game in the following ways:
- Provides a significant power increase to tactics and abilities which proc on critical hits.
Not much needs to be said here. The proc rate of a tactic or effect is increasing with game state, so the tactic itself becomes more effective than it was before, while other tactics remain constant in effect.
- Provides a significant increase in effectiveness to tactics and mechanics which increase critical damage.
Annaise16 has mentioned previously the idea that a tactic is balanced around granting 10% defense or 10% attack. However, a critical hit rate tactic's value increases in effectiveness as the player stacks crit, to a theoretical maximum of 33% DPS at 100% crit. Because of the way critical damage tactics outperform other tactics when critical hit rate rises, as well as the burst / RNG applications of a lucky string of crits, lack of a critical damage tactic has been cited as a flaw in those classes which do not have them.
The Bright Wizard and Sorcerer mechanic is also affected by this flaw.
Example 2: Armor Stacking
Armor stacking is often cited in Engineer/Magus discussions, as well as in the context of physical RDPS in general. The ability of classes to stack insane amounts of armour, aside from imbalancing the game by significantly changing mitigation percentages, results in a huge slash to the viability of physical classes that are not given the ability to negate that armor to a significant degree. This situation can be avoided entirely by... not allowing extremes of armor.
Example 3: Constant Value Effects
Many effects in Warhammer have constant effectiveness, and if the effectiveness of scaling skills increases, these effects are left behind. Classic examples of this are absorption shields, stat modification abilities (see: item reliance above) and certain damage morales.
Conversely, some effects which were improved for the sake of later states have now become imbalanced in this state, with the defining examples being Cutting Claw and Force Opportunity, which strip large amounts of armor.
The "proc meta" thread is a classic example of this - proc effects are constant damage and thus tend to be left behind as attack power increases. This topic was partly motivated by a number of players posting that improved gear would render the problem moot.
Example 4: Bottom-End Creep
I think everyone will be familiar with this one. As gear levels increase, a new player in T4 will be pitted against increasingly unfair opposition, with an effective power level multiple tiers above his own. This leaves him with no chance, and the fun of playing is rapidly lost.
Dev teams tend to compensate for this effect through bottom-end creep, in which access to lower tier sets is expediated for the sake of getting players into sets which can put up a respectable fight against the death machine sets of the high tiers, or through adding gear specifically designed to make a player into a walking brick, but with meagre offensive power or effectiveness at their role. This is highly undesirable, yet is not efficiently controllable with an absolute system.
Example 5: Broad Design Change
In order to counteract lack of thought in design and extremes of power, often it is necessary to introduce crutch items and effects which are designed to compensate for the power extremes introduced. These can be considered as analogous to improvised changes applied without due consideration to an originally well-thought out design. Needless to say, adopting this approach further corrupts the original design, as the effect is rather like taking a beautiful picture, spraying it with sewage, letting it dry and then attempting to go back to it and pick off the muck to get the original picture back again. There is simply no need to introduce stupid effects which end up requiring more stupid effects to negate, and imperfectly negate at that.
Example 6: Customization Killing Balance Killing Customization
Customizsation is only worthwhile if well-balanced. If there is a clear winner among a set of choices, the other choices are worthless options which, practically speaking, do not exist. Thus, the game resolves to a state where customization is an illusion and a dominant setup or spec exists - defeating the entire point of allowing customisation in the first place. This, to an extent, will occur in every game, but it should still be controlled - and thus any customization system must always yield to balance considerations lest it be rendered pointless itself.
Solving This Problem: State Stabilization
State stabilization means the following:
- - Above a certain tier of gear, a given set of characters must perform in much the same manner against each other in combat regardless of the tier of gear which they are using, so long as each and every character is using the same tier of gear. This means that two groups in Annihilator should have the same underlying game state, or as close to, as two groups in Sovereign. Stats may of course differ in accordance with the distributions and possibilities available via lateral progression at Sovereign level compared with Annihilator (see: later points), but a player in Annihilator should be able to acquire Sovereign, enter a battle with other Sovereigns and play the game in exactly the same manner he did at Annihilator.
- - No item proc or ability which is a straight increase in power is allowed to exist. Any such effect designed to increase a character's power must have a drawback which maintains the character's overall power level while enabling different strategies and means of play.
People will talk about the intricacies of stat customisation. They will talk about how it's boring when a tier of gear is exactly the same as the previous tier, just more powerful. This is nothing more than a smokescreen.
All people actually care about is that whatever gear they've just acquired makes them better than people who don't have it.
All the increases in stats resolve to something very simple. Effectiveness. You hit harder. You heal for more. You can withstand more damage. The players don't actually care, in combat, how that condition is satisfied - as long as that condition is satisfied, and players in better gear are superior to those who are in inferior gear, then the players are satisfied and the intent is accomplished.
There are a number of ways to implement state stabilization with respect to gear, and the following simple ones are those that occur to me now:
- - Ultra-simple: Gear has a tier level, and the sum of the tier levels of a player's gear is used in an opposed check when interacting with another party. The relative factor, with a suitable cap, is used to modify the output of certain calculations (for example, for damage.)
- A little more complex: Gear awards points for a secondary type of specialization - for critical hit rate, etc. The number of points allocated in total is compared when two players interact, and this is used to scale the total application of those bonuses from this secondary specialization to that interaction. In this way, a character in Sovereign having allocated x% crit from secondary specialization will receive that bonus in full against a character in lower tier gear, but it will not apply at all against a foe in Sovereign himself.
Bear in mind that to satisfy state stabilization, it is not enough to provide counters to a certain specialization or ability that is offered with gear. The effects MUST BE NEGATED AUTOMATICALLY between players of the same rank, or players will violate the state by having different extremes of customisation. This is why absolute power creep is so bad.
The advantages of stabilizing state are:
- Consistent state for balancing. The ground state, and thus the balance, does not undergo extreme changes no matter how many tiers of gear are added.
- Better control over min/max. Min/max is by its very nature game-breaking. If allowed to be taken to extremes, the game will be impossible to balance.
- Fine control over the bottom-end phenomenon. Because the relative difference in gear for the sake of the power bonus calculations can be capped, this prevents the need to hand out mid-tier gear when high-tier gear is prolific. Simply cap the maximum tier difference for the calculation. The result of this with a tier cap of 2 would be:
Sovereign vs Warlord: +1
Sovereign vs Invader: +2
Sovereign vs anything lower than Invader: +2
Thus a player can only have two tier's worth of advantage over another player at any given time, and the intolerable extremes applied to players with lower tiers of gear are eliminated while preserving the advantage of higher tier gear over lower tier gear.
In Closing
I must reiterate that I feel strongly about this, and I view all of the problems raised within this post as being very serious and in need of resolution. The availability of the balance forum thus hinges upon finding a resolution to them.