Game design musing: ST schemes

Log ST. You know, ST schemes under which every +X points of ST equate to some multiple of lifting power. This might be x2 lifting power per +5 ST, per HERO System, in which case ST 15 lifts twice as much as ST 10, and ST 105 lifts twice as much as ST 100.

Log ST was the choice for the original superhero game that became HERO, and is often suggested as a great fix for GURPS‘ (sometimes alleged) troubles with four-color action. But could it work in GURPS? Here are some recent emails and forum postings on the topic follow. (My text is in black.)


An email

You state that Quad-ST is likely the best solution. However, after careful consideration, I came to the conclusion that Log-ST is actually the only way it really makes sense.

I should clarify that – I don’t necessarily think Quad ST is best for everything and anything, but do think it’s the best solution that could actually be implemented in 4e, w/ SJG’s and players’ general approval. In other words, I would see Log ST as too harsh a change to be practical, regardless of its merits. To whit (wit?):

In fact, looking over the attributes (minus ST) one can make an argument that they are Log-based already. Your thoughts on ESCARGO seem to suggest this as well (for skills and attributes, minus ST).

Agreed – for attribute > 10. For attribute < 10, we’d have to accept that attributes switch to a different scheme, unless we accept that an attribute of 0 is not “zero”, and that attributes can be negative.

That’s one of the troubles w/ Log ST, at least in terms of player acceptance. Say we use the “every 10 ST doubles lift” scheme. Then half the lifting power of ST 10 has to be ST 0, and half of that must be ST -10… I know that HERO does fine with these negative attributes, and there’s nothing wrong with them in theory, but I don’t see SJG going for it.

We can hand-wave the “zero issue” for non-ST atts, as in “they’re low enough to be zero anyway; it’s not relevant to characters.” But ST that goes lower, down to 1 and yet lower, then lower even more, is both normal and needed. We need to accept ST 0 and negative ST as normal – or, declare that the rules are different for ST < 10. That’s an option, but it tosses out the neatness of Log ST…

One of the best things about Log-ST is that the percentage of increase is constant at all levels of ST – meaning that the costing scheme (using the one from ESCARGO) would actually make sense. So, in your estimation, what advantages does Quad-ST have over Log-ST?

As above, main one would be conservation of a foundation of GURPS: an attribute of 0 means “zero”.

There’s another: conservation of a ST-to-damage ratio that’s (arguably) meaningful.

Again, let’s say +10 ST = double lift, and let’s also say damage = ST/10 dice. ST 20, with twice the lift of ST 10, does twice the damage. But ST 110, with twice the lift of ST 100, does only 10% more damage. Similarly, ST 110 is – what, 1024 times the lift of ST 10? But a mere eleven punches from a man with average ST would inflict as much damage as a punch from the monster with the strength of a thousand men. I always thought this was wonky in HERO. Mixing “exponential” ST with “additive” damage gives funny results – and I can especially see it messing up “uber-systems” like Vehicles, which make big use of clean math.

Even tougher: If damage has a linear relationship to the ST score, then how do you solve the low-ST issue, where half normal ST would inflict 0 dam, and lesser ST would technically require negative damage?

Exceptions and patches can be used to solve these, but again, it’s a shame to have to “break” a system that was chosen for its (very real!) technical merits. For what it’s worth, Quad ST doesn’t face these particular troubles.

Or perhaps damage under Log ST should be reworked to solve this issue, so it doesn’t scale linearly with ST. After all, it’s not quite linear in GURPS to begin with – but then again, it almost is. The required reworking of damage under Log ST would make for a big change in GURPS – not a bad thing in theory, but maybe impractical in terms of acceptance.

In short: Log ST does have great merits (I especially like how it allows ST back into the Contest mechanism), but it’s quite a change.

Final thought: if one were to adopt Log ST, is there any reason to stick with ST 10 for humans? As long as zero and neg ST are unavoidable, why not set the “fulcrum” right on the zero point – i.e., positive ST means greater than the human norm, negative ST means less? Kind of like Strong/Weak Will in 3e… (It fits my claim that ST isn’t really much of an “attribute” at all…) A quirky idea, but I like it.


A forum posting

However, it has been pointed out by TBone that the one major problem with LogST is the fact that ST 0 is not zero strength at all but is simply a very low number. And the ST scale would have to continue to extend into the negative numbers. This is a bad thing. Thus in the end, I have to agree with him and go with Quad ST as the most logical choice for ST in 4e.

Hmm, now that I’ve worked to convince you of the flaws in Log ST, can I also say that I think it could be made to work in the game?

I haven’t worked through this in detail, but here are my untested thoughts on how to shoehorn Log ST into GURPS. (Use +5 ST = x2 lift, for example purposes):

a) Yes, you have to allow for negative ST, unlike the other attributes; ST 0 is just another normal point on the way from ST 1 to ST -1 and lower.

But we already have traits that can be negative, zero, or positive – think of Strong/Weak Will. So this doesn’t have to be seen as evil…

(Hmm, why not ditch HERO‘s lead, in which a normal human is Log ST 10, and make normal human ST 0 instead? This isn’t necessary to implement Log ST, but it’s worth considering…)

b) Ditch any sort of HERO-like “+5 ST = +1d damage” scheme. The ease of doing so is an attraction of Log ST, true, but the scheme just doesn’t match GURPS. Funny things happen when you use additive arithmetic on multiple occurences of log-based damage. And what the heck do we do with negative ST? Use negative damage??

Instead, let’s some sort of “+10 ST = x2 damage; -10 ST = x1/2 damage” scheme. It may require a damage table like we have now, but it meshes with GURPS (especially Vehicles) and works at any ST level.

I think the above lets Log ST work; b) in particular is key. And we’d get the great benefits that you’ve talked about in the past. Reasonable cost for supers. Contests of ST that work just great. Maybe even workable skills based on ST. And ST would mesh better with other attributes, which are arguably log in nature.

(Ironically, by admitting that ST doesn’t really fit in with the other attributes and taking the liberty of futzing around with it massively, we’re able to get it to act more like an attribute…)

But I just don’t see SJG going for a change that big. I guess that’s my only objection with it as a serious proposal for GURPS 4e.

I think we both agree that in a brand new RPG, properly-implemented Log ST would be the bee’s pajamas.


A forum posting

Note that in a four-color supers game, the +5 ST = +1d damage rule would work just fine. It might be worth noting that as a customization, either in the Basic Set or in a new GURPS Supers. Of course, you’d have to refigure a few other things as well; all the weapons would need to have different damage values, and DR would become a function of log thickness, I believe.

For the record, the problems with a HERO-like ST score that mixes linear damage with log lifting ability are as follows. Using HERO as an example (human ST = ST 10, 2d damage; +5 ST = +1d damage, x2 lift):

a) Funny math. In HERO, ST 60 offers about 1000 times the lifting power of ST 10 — yet a punch from that monster does no more damage than six punches from a normal schmo.

Not that that’s necessarily a terrible thing, especially when DR is involved. But there’s more that’s funny. At ST 5, doubling your lift (+5 ST) also doubles damage. At ST 10, doubling your lift adds only 50% damage. At ST 50, doubling your lift adds only 10% damage. Why?

b) The low-end. One-quarter human ST (ST 0) does 0 damage. Combat at small scales won’t work. I don’t know what patch HERO uses to handle these, but a patch is needed to allow some damage at ST 0 and lower.

In short, the incompatible mix of linear damage and log lift is lousy for real-life simulation, and would wreak merry havoc with systems like Vehicles.

However. Your post is saying, “For four-color supers, to heck with real-life simulation; HERO-like ST works fine.”

I’ll agree. Punching power follows no logical rules in comics, and nobody cares about the workings of low-end ST in the genre, either. The “problems” can be brushed aside.

More and more, upon hearing good comments on this list, I agree that the key to enabling comic-book supers in GURPS is to give up any pretense of real-world physics and rules based on these. Instead, go for bogus physics that play – well, like the alternate-reality physics in four-color comics.

I haven’t thought through HERO-like systems for GURPS supers in detail, and certainly haven’t played it. But it’s worth looking at, I think, as an optional system. (Many things would have to be changed in “four-color GURPS“, as you point out; the logical place for the entire system would be in the Supers book.)

Gamers would still have the option of playing supers under normal GURPS rules, for all the gritty extreme effects that this entails (people-pulping punches, etc.). It actually makes for a fascinating look at “What if supers existed under real laws of physics?” Ooh, messy…

I wonder whether SJG would go for it. Although GURPS‘ goal is “one set of rules”, SJG has always realized that no one set of magic rules works everywhere; a new game background may call for a whole new system of magic.

Is it such a big leap to allow that the world of four-color comics calls for its own unique twist on physics and power?

2 Comments

  • Demi B

    There exists a ST/Damage/DR scheme that I think works well: BESM 3rd Edition. Mind you, it is designed as a simplistic cinematic system, and doesn’t have the fine grain of GURPS much less Real Life, but the mechanism is elegant and scales well.

    The mechanism is Size Templates. Each level of positive or negative Size Ranks gives a specific set of modifiers, each with itemized point values. Amongst these are: damage delivered, and armor rating (which can be negative, meaning extra damage received). Of course, the values are inverses and additive, and only apply to size differences – pixy on pixy fights use no modifiers since they’re the same size, and pixy on giant fights quickly end in a little red smear since that’s 4 or more Size Ranks difference.

    The system partially decouples ST from damage (super-simplified “Body” stat contributes to combat, and combat rolls determine damage), but understands that smaller targets are more fragile, and larger ones more robust. So finds a way to scale damage values to the target in an extensible logarithmic fashion.

    When I first read the book, this particular system reminded me strongly of your ST/size musings from years ago, and I wondered if David Pulver had read your works.

    • tbone

      I’ve read a little bit about the BESM system, but haven’t played or read it, so I’m not qualified to comment.

      I’ll do so anyway: )

      What you describe sounds interesting. If I understand correctly, both the Pixie and Giant would use “normal” stats and dice and rules in combat, but any results would be shifted up or down 4 levels, as appropriate.

      I wonder why the use of a Size Template, though, instead of just appropriate mods to ST, Damage, and DR. Many characters might modify those stats without unusual size, and might mix them in odd ways: a Super with massive ST alone, or massive DR alone, etc. I assume the system handles those cases just as well. If so, is Size Template just a convenience for combining such mods in an odd-sized creature?

      I’ll have to look for a chance to play and fully grasp it!

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.