Admittedly, I haven't yet given this question one iota of thought towards an answer; however, that's to be expected, given that the question occurred to me two minutes ago, long enough to get onto this site and start writing this sentence. I'm posting first and thinking later, after I click the Post button, while others may (or may not) come up with their own thoughts and post them as responses.

I've been doing a lot of research for my system, during which I came across this article on Tactics by Brian Gleichman quite a while ago. I'm sure the regulars have seen it. (Newbies to the Forge, read it! And it's mother, this set of articles!)

I've taken the article on tactics to heart, and am attempting to incorporate it's lessons into my designs for a combat system. I'm aiming for a gamists'-dream-system, but one in which they have to actually fight for their victory at the gaming table and will hopefully find it impossible to win with just an optimal character build (though obviously such a build will give them an advantage in any case.) This article suggests that a really juicy tactical system must, by definition, include dissimilar assets.

Here's where it occurred to me that a character point-build system is likely to conflict with this goal. I think it's obvious that when a player builds his character in power, they are going to choose options that give them more power, as well as make choices to reduce their weaknesses. In other words, the dissimilarity in their resources dissolves over time. At the highest heights of power, tactics disappear and give way to simple twitchy video-game combat.

If I want characters to grow in power, but I'm also keen on using a point-build system, is this dissolution of dissimilarity inevitable? If I come up with a solution, will it always be a hack fix or can there exist a solution that fundamentally solves the problem?

I'd be interested in hearing if anyone else has come across this or a similar problem, and what approaches you took to deal with it.

Daniel

It's only a problem if the only things that are of use tactically are internal aspects of the character. And that's... pretty much never true, right?

Though not in this arena of design, I recently in Actual Play (it happens) had to ask my players if they minded if one character was seemingly disproportionately powerful in the social sphere (A Peer in Victorian England). They were fine with it.

Also: if your designs intend to make it so that optimal builds aren't all that is needed for victory, isn't that just shifting the optimal definition build around? For instance if you need offense, and defense, and manuever, and intelligence, and trickery,and etc ..oh and you'll need to lead 100 farmers and support their culture in order to hold what you have won, well then having the ability to deal with farmers just became part of the optimal design.

Thanks for the article links!

I am hoping that someday there will be a wiki with links to every Forge and non-forge article that seems topical.

Those are some neat articles. I disagree with his evaluation of RPS though -- it's very useful as a way of guaranteeing and requiring dissimilar assets since every unit type has a weakness and a strength. (The Fire Emblem video games are a good example of how to leverage RPS relationships to increase tactical play.)

Anyway, my impression is that you want a game where it's important to build an optimal character, but it's also important to have good moment-to-moment tactics. i.e. you want to reward skill both in deck-building and play.

Is the game you're working on meant to encourage team play among the players, or to be more player-vs-player? Or maybe strike a balance between them, like in Agon?

In a game that encourages team play, players are more likely to focus on enhancing strengths than on shoring up weaknesses because other team members can compensate for those. e.g. The close-up melee guy will likely concentrate on that and leave long-range damage to the sniper. As situations become more dangerous, some shoring-up will occur to make sure the character isn't taken out of play. e.g. The sniper will invest resources in defense and minimal close-up combat skills so she can survive a few rounds of melee combat in an emergency.

One thing to note is that you can easily combine classes and point-buy systems -- they're not mutually exclusive. Most MMORPGs do this. So does D&D 4th ed, if you think of the power slots as points. (IIRC, attributes are point-buy as well.)

If you want classless point-buy, there are still things you can do to encourage niches and variety between characters. One is to make chains of prerequisites, so that Power X requires Powers Y which requires Power Z, and once you have Power Y, Power X is more useful than unrelated Power A.

Another way is to adjust the costs of specialization so it's cheaper and more effective than being a generalist. In most point-buy systems, higher skill ranks cost more than lower skill ranks. But if the difference in cost is much greater than the effectiveness gained, people will tend to spread their points around. e.g. in Fable 2, going from level 4 to 5 of the Inferno skill costs 50 times as much as going from level 1 to 2 of anything else, so there's a strong incentive to buy a level or two of everything.

Basically, if you want to encourage players to play dissimilar characters, make sure there's incentive for variety between PCs, and that there's incentive for specialization. Variety of challenges is a good way to encourage PC variety, since no one PC will be able to handle all challenges by herself.

Depending on if its team play or competitive play, you'll get different results.  If you're building a character, you aren't going to want to be scissors, where any rock automatically beats you. Instead, you'll want to be a paper resistant rock.  If you're doing team play, you'll want a rock, a paper, and a scissors.

My game is team play and (more or less) point buy.  With every group I've shown the game to, people have specialized instead of being (weaker) but rounded characters.

Holy jeepers. This question/topic is right up my game's alley -- both in theory and in practice. Nevermind that I'm in the middle of an edit/revision.
I'll jump right into some of the elements from Gleichman's essay.

Gleichman seems to refer to classes (along with Selene Tan) as a solution to your problem, Daniel. D&D might be a poor example because, in some versions, players are able to mix and mash classes -- essentially finding an optimum build.
Here's an example from my game: characters choose 1 superpower, out of 11 different superpowers. They're stuck with that power, and there's a huge, steel-plated Wall-o-China between powers. No one gets two different superpowers. (Note: my game's theme is not capes&spandex heroes, so single powers are okay.)

So, then I must be building for team play, then? David C is right that team-play will lead to deeper specialization. However, I'm in the middle of a gamist/right-to-dream design goals. How will a solo player (teleporters, super speeders, flyers often go on their own) fare without being able to rely on some other dissimilar assets?

On top of the dissimilar classes of character (wizard, thief, etc), there can also be dissimilar 'classes' of action (melee, projectile, etc.) In this, a player is able to become more-or-less generalized within his class.
Maybe a wizard class has many more projectile actions, but he can also point-buy heavily into his few physical actions (making them awesome but limited.)

In my game, I've set up 7 different action classes.
Each superpower, out of 11, has access to three or four action-classes.
(Teleporters have physical, projectile, and psychokinetic actions, for example.)

While at first blush, you'd say, "By generalizing within each class, you've allowed players (with enough points) to reduce the difference between the powers."
I'd call you a genius, then. You noticed that the number of character points floating around has a bold impact on dissimilarity of these assets (your characters' classes.)
Rather than reducing the number of points available on the board, I raised the cost of actions (then allowed each action to be customized within itself.) Now, due to the economy of points, players have to make a lot of front-end decisions before each combat.

As mentioned, I could go the other direction and limit the number of points available on the board. However, notice Gleichman's distaste of limiting a player's options:
"As a general rule, increasing the number and types of resources you need to manage increases the tactical play of the game." "As a system weakens character niche, it reduces tactical play."
While character points aren't exactly a Gleichman resource (ie: consumed), they are valued in a point-build game. With more of them on the board, players have more decisions on-hand.
Now, players can improve small things (+1 damage for a specific action costs 8 character points) or big things (+50 maximum health costs 100 charcter points, or 1 rule-breaking super-action costs 1,500 character points).
Certain powers allow players to mitigate the effects of certain action classes. Teleporters can easily evade physical attacks; pyrokinetics have a defense against ice attacks; telepaths can avoid mental attacks but no one else can.

But to avoid videogame-twitchy combat, designers have to make sure that the dissimilar assets (each character's actions) are meaningful in their use.
Obviously, some action classes mean different things for different objects or items on the board.
However, characters just slapping each other with these actions they built ahead of time creates a sense of meaningless use of fortune. (IE: I built my powers to teleport behind and hit you.. and it has a 85% chance of success.. and just when I want to use it, the dice fail me.)
To avoid that, each action allows a reaction from the attacked player.
(I teleport behind you to hit you, you throw a fireball at me, I drop something heavy on you, you burn the ground beneath me).
The instigator keeps up the actions (which can be used only one-at-a-time -- no duplicating actions)
Now each of those character actions -- played in sequence -- create a sense of maneuver.
Even if the dice bite me once, they can't harm the entire chain.
Players' decisions before a fight (and during the fight) then determine success -- and speeding up a Pace of Decision.

My conclusion-solution to your question isn't exactly easy. I ate up Gleichman's essay like it was my daily bread for about three weeks while I was writing out the designs for the above. What inspired me was his implicit endorsement of multiple, interacting decisions.
My design for combat seems like a lot of wheels spinning inside of other wheels, which turn other wheels at right angles. But enough weight is given to each decision so that the players understand (and between turns can modify) their characters' powers as it relates to everyone on the board  -- friend or foe.

Recognizing a natural tendency in players to try to generalize, I flipped the scale over.
Instead of allowing players to start as a generalized character and coaxing them into specialization, I forced them into specialization corners and allowed them to generalize.

An interesting note: For one play-test, I forced everyone to be a teleporter. The characters were different enough that their play decisions forced one player to equip himself with military gear (non-superpower equipment), another to specialize a handful of actions with no equipment, another to generalize his evasion and use one trusty attack item (a gun).

Include prerequisites. Make things harder to rank up once you already have "enough" of them, so that people are more balanced.
If you want to steer towards a particular "build," Fighter, Wizard, etc. then make it so that there are plenty of prerequisites for progression in a given area, so you're more likely to choose a role and stick with it because "you've already gotten this far," but also make an increase in something proportionally expensive to how much of it you already have.
You might have two different systems, with one being somewhat like Feats, and another being somewhat like Skills, and force them to cooperate with each other. For the Feats, you simply select based on the given prerequisites - these will be spells and the like. For the Skills, take "ranks" in them that get more expensive over time, say, at a rate of an increase of 1 for every 3 you already have. The Feats will have a certain levels of certain Skills as prerequisites.

Hopefully this is a bit more intuitive than "D&D but ranking your skills gets even more expensive!" It will require tweaking and balance to be sure, but I think it's an achievable way of doing it.

Thank you for the input, gentlemen, ladies.

HeTeleports wrote:
My conclusion-solution to your question isn't exactly easy. I ate up Gleichman's essay like it was my daily bread for about three weeks while I was writing out the designs for the above. What inspired me was his implicit endorsement of multiple, interacting decisions.
My design for combat seems like a lot of wheels spinning inside of other wheels, which turn other wheels at right angles. But enough weight is given to each decision so that the players understand (and between turns can modify) their characters' powers as it relates to everyone on the board  -- friend or foe.

Recognizing a natural tendency in players to try to generalize, I flipped the scale over.
Instead of allowing players to start as a generalized character and coaxing them into specialization, I forced them into specialization corners and allowed them to generalize.

An interesting note: For one play-test, I forced everyone to be a teleporter. The characters were different enough that their play decisions forced one player to equip himself with military gear (non-superpower equipment), another to specialize a handful of actions with no equipment, another to generalize his evasion and use one trusty attack item (a gun).

ShallowThoughts wrote:
That idea of forcing early specialization, with generalization only coming later, is a great idea.

That's a good point people have made about whether the game is cooperative, independent (the players work alongside each other but not necessarily in concert) or competitive. I'm shooting for the middle option, independent for a few reasons, mostly because I don't want extreme dependence.

Generating prerequisites and dependence is one way to approach this, but I'm actually trying to avoid this. It's a similar thing as simply making classes, except that the "nicheness" is emergent between the individual skills and powers.

Ideally, I want a system to allow not just fighters and wizards, but a special configuration of fighting and wizardly abilities.

This conversation makes me think that what I've been looking for is nicheness without naming the niche ahead of time: Properties of a niche is that it is specialized to survive under particular conditions. To what degree specialization should occur, and what conditions it is adapted for, depend on context.
This makes me think that what I might do is, instead of hardcoding nicheness into the system, I should instead code a "nicheness lever" under control of the PCs.
They could specialize to whatever degree they want, but the more they adapt to that niche, the harder it should become for them to change and re-adapt to new conditions.
However, because the niche is user-defined, if they keep to one consistent vision for the character, they shouldn't need to change at all.

(Hmm, now I wonder .. is this really what I want?)