2018-03-26

Point-Buy Systems Considered Harmful

A few weeks ago in "Sarge's Advice", there was a pair of questions (#8-9) on possibly pricing out individual attributes for a unit, such as movement, hit dice, armor, and morale. I feel like that bears on one of the main themes of this blog -- but when I went looking for a link to a post on exactly that point, I couldn't find any. Here I'll try to rectify that; my main thesis is this:

Game systems that price out individual attributes for figures, on an à la carte basis, are always inherently broken.

So: don't do that. (You'll see in my response Monday that I resisted approaching that idea.) Here are the top three examples that come to my mind:
  1. The Monstermark System for OD&D (in White Dwarf #1-3)
  2. "Creating New Character Classes" in 2E AD&D (DMG, Ch. 3)
  3. "Calculating Magic Item Gold Piece Values" in 3E D&D (DMG, Ch. 8)
You can probably identify others; and not just in D&D, of course. All of these systems were inherently broken and, to my knowledge, no one uses them anymore (they simply didn't stand the test of time). The 3E system probably stands out as the most poisonous of the lot, because at least the other examples had the grace to mark themselves as "Optional", but the 3E system did not. (OMG the rivers of blood that flowed over that topic on ENWorld back in the day, as every player thought themselves immensely clever for seeing that they could make a 2,000 gp infinite-healing magic item.) Now that I think of it, back in the day I also built a robot character in a Champions campaign this way that also completely broke that game.

The basic problem with any of these systems is that there's probably some synergy between two or more attributes that the designer didn't/couldn't observe, that allows constructing an overly-powerful figure when combined. And also: Some number of attributes that overlap redundantly, so that paying points for one becomes a waste as it doesn't have any benefit in the face of the other attribute. Or: The attributes are situational in nature, such that for the same price on one type of figure the ability is useless, while on another type of figure it is overwhelming. Here's what I wrote regarding the Monstermark system (the one I most recently dealt with in detail):
A giant rat given magic-to-hit defense is effectively unbeatable by the PCs it normally fights; but a very old red dragon, given the same ability, would have little effect against its high-level opponents (surely wielding magic weapons already). If ghouls have possibly paralyzing attacks, then it makes a huge difference if they have one attack for 1d6 damage, versus three attacks for 1d2 damage (even with nearly the same expected damage). Centipedes and carrion crawlers, with a base damage of zero, even with poison or paralysis, would generate a product that is still zero by this multiplicative system. And so on and so forth.
Granted that we shouldn't engage in point-buy attribute systems (because they are always broken), we therefore have this corollary:

The acid test is gameplay.

That is: once you've designed a unit, the only way to judge if it's properly balanced cost in the game is to play a lot of games and assess its real, on-the-table, utility and power. This is hard. This is why, throughout this blog, you'll see me emphasizing simulating our game designs in software so that we can test the various units a few tens of millions of times before adjudicating a final price. Of course, this was not a thing that anyone could even imagine before copious computing power was at our fingertips all the time. Historically, I assume that most game design was done by creating pieces for a game, and then pricing them as a near-afterthought (e.g., Mike Mornard has made this point many times); in my game designs the structure of the pieces is only about 20% of the effort, and the other 80% is adjudicating the right price (via simulations).

Examples of systems that we've simulated in software for pricing purposes:
Blogs where we've discussed this overall issue in passing (search for "acid test");

Do you have other examples of broken à la carte point-buy game systems? Or any you think aren't broken? Discuss below.

21 comments:

  1. Vampire: The Masquerade and storyteller in general... The system is so broken that I don't even know where to start...

    ReplyDelete
  2. Well on one hand, I remember the feeling of refreshing clarity when I switched the B.A.S.E. campaign from the chaos of Heroes Unlimited to the order of Mutants & Masterminds. (Although eventually my modus operandi was using the random tables in HU to come up with the ideas and using M&M to convert them into something playable.)

    I think the big problem with point buy RPG systems is that it introduces the concept of "builds" into character creation, and instead of being interesting fictive entities, it turns characters into bundles of optimized stats with no deeper meaning or personality. Eventually this leads to some classes being the "best" ones that everybody should play if they want to "win", and lets other classes die on the vine as they're not optimized enough to "compete".

    Bah! Nerts to that! Tell me a story and let me tell you one back. The numbers are the skeleton, not the skin! Exoskeletons are for ants!

    ReplyDelete
  3. I'd go you one further and eliminate attributes (mechanically treat is if they were all 10). Do they really add anything to the game?

    ReplyDelete
    Replies
    1. You can do that. All my henchmen are attribute 10 unless we say otherwise. There is very little benefit from high stats in the older sets and they work fine.

      In my Treasure Hunters Prolix, we roll stats but we only record if they’re high, normal or low.

      Delete
  4. These systems are inherently broken in the absence of a referee who says, “yes that’s terribly clever, but you can’t have it because I say so.” Champions especially so.

    ReplyDelete
    Replies
    1. Having played oceans of Champions, I agree with you. I would never call Champions an OSR-type game, but it is refreshingly old-school in that it says "the system is founded on having a GM with a clear vision as to what they want for their game and clear dialogue between the GM and players as to what's possible, because otherwise any moron can break this game in 5 seconds", instead of just saying "the rules are all". The goal is to be able to model everything, and since the universe isn't balanced, neither innately are the rules of a system modelling universal outcomes. They even put examples of how to easy it is to make a character that can conquer the world with a starting point total in the main rulebook in 4th edition. I think playing it with Pathfinder/3rd ed fans, who are used to being rewarded for and being able to play optimal power builds created by obscure synergies, would be interesting.

      However, for old-school D&D my back innately goes up when I see point-buy anything. I think that's my experience with all sorts of broken spell point systems talking.

      Delete
    2. Great example with the spell-point systems.

      Delete
  5. I have a different point of view. I think Savage Word does a good job. Yes, you can min max characters, but most will turn out about even. I like Victorious RPG from Troll Lord Games. Yes, since it is set up to quickly create characters there is room to min max. However, depending on your gaming style IronMan can fight better than Sherlock Holmes, but SH can find more clues. I happen to like good ability scores and I have created methods of generating the ability scores I think are worth having. Nothing I have written is to say that an opposing view point is wrong, but I have my preference.

    ReplyDelete
  6. Not a role-playing game, but the sci-fi tank game system Ogre/GEV evolved a points system with copious amounts of playtesting that is fairly balanced. There is a formula to calculate unit costs (both giant cybertanks and conventional armor) based on its movement type, weapons strength and range, etc.

    ReplyDelete
    Replies
    1. Where is that? My 1987 "Deluxe Edition" doesn't have any of that.

      Delete
    2. There was an article by Henry Cobb reprinted in the second edition of The Ogre Book that detailed the cost formula. There's also a version available here and a web version for Ogres available here and one for conventional armor here.

      Delete
  7. Power and combat effectiveness in OD&D can be so tippy and random - based on a million factors you can't anticipate - that, in my view, any value assignment system that goes beyond the very simple is a waste of time.

    For monsters, I just use hit dice plus 1 or 2 possible additional pips, based on armor class, special powers and so on. Would say, Monstermark be more "accurate" than that? Perhaps a little, but I can't imagine that would be enough to justify the extra work. And even if it's slightly MORE accurate, it's still itself going to have tons of problems, as you suggest.

    But here's the other thing. Suppose your system - simple or more complex - underestimates the value of Medusae by, say, 50% - whatever that would even really mean. So what? On many dungeon stocking systems (including that of OD&D and my own), there are all sorts of OTHER random factors that may have a greater effect on things, form the NUMBER of monsters encountered to which LEVEL of monsters the table tells you to pick, etc. In other words, since you will always potentially have imbalances, adding one more potential one isn't that big a deal. Does that make sense?

    ReplyDelete
  8. For many years GURPS was our go-to system and it still sees a fair amount of use.

    The thing is that it's easy to munchkin GURPS, especially when you're playing games with special powers or high point values. We all knew it, players and GMs alike, and we agreed not to do that. GMs always had veto power and we rarely had to use it.

    So I would say that a point buy system isn't broken, it's a different mindset. It gives the players more flexibility to customize their characters but it requires more buy-in from them to not game the system.

    ReplyDelete
  9. I used the class design system from AD&D, to great effect, but I also agree with your conclusion. The difference lies in the application. As the DM, I'm responsible for deciding what rules to use. My preferred method is to apply as much math as I can get away with (always keeping an eye to the results). I would never let my players in on that design process, however, which is where most point-buy systems break down.

    ReplyDelete
  10. I'm reminded of a Gurps game back in college in which a friend played a character affectionately known as "super kitty". In that system you could generate extra ability points by taking hindrances. He made his character an ordinary house cast, which incurred all manner of hindrance points (small, can't talk, no hands, etc.) but then made up for these with amazing abilities that effectively neutralized those (magical armor, can fly, telepathy). The result was a totally broken character that was amusing for a time.

    ReplyDelete
  11. Crabaugh: "Customized Classes", Dragon #109. I consider the XP chart of a D&D class to be equivalent to a unit's point-buy.

    Also all of Player's Option 2E. You could multiclass with Cleric, spend points to get access to the spheres that give healing and raise dead, then blow all the rest of your Cleric points that should go to sphere access on amazing class abilities elsewhere. Also everyone just ditched taking name-level followers and got 5 free points to spend on something that made the character stronger.

    ReplyDelete