I had an unexpected surprise when Greg Porter himself responded to my post about scaling in GURPS and EABA. His comments and the back-and-forth he had with David Pulver are not to be missed.

What? You missed them? Hie thee to the comments section of that thread. Read. I’ll wait. Good? OK, we continue.

But his original reply deserves more than to be buried in a comment section, so I pulled it out and replied point-by-point. There are places where we agree, and places where we don’t. My posts are indented (because I’m responding). His are in purple/bold, in the main text.

**The concept [logarithmic scaling] is not new, even if not fully chart-ified in the past. Hero System is a “doubling every +3” system for things like Strength, and I think Bill Willingham’s Underground (Mayfair Games) also had a progression-based chart.**

Definitely not new, but also definitely one of the best, I think, ways to approach a universal scaling for a generic system. I had wondered if setting the “number of divisions per x10” equal to the sides on the dice while rolling 3d6 would work out as generic (I was hoping that 3d10 would work out well with a 10-step per x10 scaling), but it didn’t. At least, it didn’t with the elegance of the GURPS 3d6 and 6-steps-per-x10.

**There’s good/bad, optimum/non-optimum to all of these, and it often depends on what you are looking for and what the system is trying to do. DC Hero’s “doubling per +1” is indeed brutal, but probably necessary if you are going to put Batman and Superman in the same scene and not have to fill Superman’s sheet with 12-digit Strength values.**

Superman is always tough, because his powers fluctuate with the writers’ needs. Being able to push planets around is going to challenge pretty much any quantifiable system. Even description-based games such as Fate will probably have some issue with that, though “as strong as the scene needs him to be” is much more tractable with a game like Fate than GURPS, with its bias towards the quantifiable.

**GURPS works pretty good at the human level, but is human-oriented enough that going to superheroic levels is often a problem.**

I think the log scaling that I am trying to wrap my head around (and that +Sean Punch looked at in the most recent Pyramid) fixes that somewhat – actually, I think it has the potential to fix it entirely. EABA with it’s inherently log-based chart also scales from fleas to kaiju. I haven’t internalized the system yet – still reading through it – but I have gathered that already.

**And EABA has its quirks, too. It doubles every +3 for Strength and every +2 for everything else (and as designer I can tell you that took a -lot- of tweaking). And you don’t have to memorize 20 levels. If you know that each +2 is x2 and each +1 is x1.4 (or for most purposes, x1.5), then if you know a particular value it is pretty easy to scale up or down. For instance, 25 meters is a ranged difficulty of 12. What’s the difficulty at 50 meters (x2 distance), 100 meters (x4 distance) or 12 meters (x1/2 distance)? You didn’t have to memorize 20 levels or even look at the table to figure it out. The same could be said for -any- of the other games with regular progressions.**

Where I think that this runs into issues is what if you approach it as “the range is 2900m?” In GURPS, for example, if I know that 20m is +6, then 30m is +7 and x100 for that is 12 more, making it +19. For EABA, each +20 is x1000, so that makes it harder. Though if I happen to know what 3m is, then I can just put that and add 20 to it. I’m sure it gets easier with practice, but “oh, it doubles each time” isn’t easy at the table, but adding or taking away zeros is, since we’re used to base10 placeholders.

GURPS’ scale is very nearly the same as yours, of course. You have +20 per 1000x, repeating, while GURPS is +6 per 10x.

It’s a small point, but I’ll stick to it: a smaller number of values until a progression repeats exactly makes it easier to play without resorting the The Chart, A Chart, or Any Chart. This is true enough for the fan base for which I write that it pretty much gets pounded into my style. And you’re talking to the guy who put transcentental equations into GURPS for the purposes of bow design, and whips out any convienient math function at the drop of a hat – since I love systems with underlying mathematical rigor because they scale well and can be tested for extremes easily.

All of that is trumped by the players and GM being able to instantly get to the “roll the dice” part. If you as designer didn’t feel that the x1000 per +20 wasn’t important, you wouldn’t have put it in – so people (and the game rules) are making differentiations based on that progression, and a lot of things seem to happen at breakpoints. You have to know that 25 is one of the points on the chart.

Actually, looking, it’s not on the chart. 23 is on the chart, which is sensible (it’s 2^4.5 power, rounded to no digits). But you need to know that it’s 23, and whether you round up, down, or nearest (EABA is not unique here) and that 23 is a unique level, but 230 is not – that’s 250. 500 is on the chart exactly, but 50 isn’t. So I’ll stick to my point here and say that people are going to want to memorize the progression, and that 6 or 7 levels per x10 is a more natural fit than 20 levels per x1000. It’s not a game-making point, but I think it’s a legitimate one.

**I will disagree and say you can’t actually set x1.0 to zero and have things work out if you want to do “table math” -and- have the values be useful in game terms in relation to things like Strength rolls and the particular game’s dice conventions.**

This may well be true. It really depends on how often one will do manipulation with the log of the actual value and have it matter, as opposed to just working with modifiers. If is all you ever need to know is the ratio of two lifting forces (200 lbs force is fighting 150 lbs force, that’s always 1 step away on The Chart or the GURPS progression. That will be a +/-1 bonus or penalty in either system.

If you’re going to be asking for the actual force (or distance, time, or whatever) values frequently, then setting the zero to 1.0 is going to be more important. As a game designer, one will have to pick. It should be relatively straight-forward to do one or the other.

**I -think- that values are usually set to minimize the number range and math for the normal expectation of play. So, a GURPS default target size of +0 means no math operation is needed for the default target size. Which happens to be a unit of measure that is not 1.0. Setting default human size to “1” (for a +0 modifier) would means the game’s measurement scale has to go from being in yards to being in “person heights”.**

As David noted, the basic scale for GURPS probably should have been 1 yard, which is the unit of hexes used (so attacking at 1-yard adjacency – melee combat – is at no penalty naturally), and a human torso is roughly one yard tall. That would probably have avoided some gymnastics. But you’ll always run into this, since (for example) the default amount of lifting power for Joe Average is not 1 lb., or even 10 lbs. It’s 20 lbs (for the basic increment of lift for ST 10), and from there, it’s ratios – where a lot scaling would work rather perfectly!

One of the reasons I’m so interested in the log scale for GURPS and ST is that so many of the quantities that GURPS relates to the ST score have ratios as the most important point. Multiples of Basic Lift define encumbrance, and HP are derived from ST, and thresholds for wounding are defined as things like HP/3, HP/2, -2xHP, and -5xHP. All of those are easily picked from a logarithmic list.

Thanks again for your considered reply. Having you, +David Pulver, and +Sean Punch discuss this stuff on my blog is ridiculous fun for me!

I think the issue of scaling in GURPS is an old old won and in fact prompted the genesis of Fudge (which led to Fate). The idea of scale in Fudge is that the base stats are relative within the SAME scale.

There is a separate scale factor that is added in if two combat of different scale interact. Each +1 Scale is 1.5 better than the previous level.