Ok last question:don't you have a thread that summarize what's the current status of the work.

Like:

Basics for rolls: Thread A. Status: Validated
Character Creation: Thread B: Status work in progress.

and so on...

I could have sworn we had at least a partial summary. I may be remembering a few topic-specific summaries and confabulating them into a single master summary.

Something to work on.


The core dierolling mechanics are currently unchanged from base SR3. We investigated a bunch of ways to eliminate the perennial 6=7 problem, but the simple solution (reroll 6, add 5) further compresses the already-none-too-wide range of achievable TNs and the solutions that had nicer probability curves were decidedly baroque (reroll 5 and 6, add 4 was the leader of that pack).

There was a push to reduce the places where Open Tests are used, and possibly eliminate them altogether, but it's not a high priority for me so I don't know whether it will happen.

The last thing in the pipeline is that there's a lot of interest in solving SR3's Impossible Tasks problem (exemplified by Invisibility, where the casting mage can get more successes than your average or even non-exceptional above-average person has Intelligence), with proposals including ways of trading off TN penalties for extra dice or allowing exceptional single-die rolls to generate more than one success, but there's not a lot of clear inspiration on that topic so it's been back-burnered for the moment.

~J

So I'm working on some example mages, and it's reminded me of a proposal that I want to run up the flagpole as long as I have people's attention (well, to the extent that I do). As it stands, Drain is not really a tactical resource—taking any amount of it is essentially unacceptable due to the TN mods produced, with a few boxes of wiggle room in case of things like High Pain Tolerance. The Trauma Damper is totally broken, but there are a whole bunch of spells that are practically unusable without it (the alternative being a big stack of expensive foci).

My thinking was to add an additional Damage Track for the Awakened, the Drain Track. Drain would go onto this track (Physical Drain would still go to the Physical track), no TN mods would be applied based on the Drain Track, and the Drain Track would overflow into Stun. This would go hand-in-hand with an increase in the difficulty of totally resisting Drain, possibly the Drain-at-full-Force rule. The Drain Track would recover like Stun Damage; I'm not sure if Stun would have to recover before Drain, or if they'd recover at the same time (possibly with a penalty, like one of them taking double the base time to recover when the other isn't empty).

There's already a proposal out there to set TN mods based on the sum of total boxes across modifier-giving tracks (so a Rigger with Light Physical (1 box) and Light Stun (1 box) operating a vehicle with Light Damage (1 box) is facing +2 TN, the equivalent of three boxes in any one of those tracks, rather than +3 which would equal 6 boxes in any one), so there's a bit of a risk of overcomplicating damage tracks, but I do feel like right now any spell that can't be reliably cast without Drain is nearly worthless, barring a few exceptions that can mostly be cast during downtime.

Thoughts?

~J

i like the latter better, because it makes things easier for anybody, instead of just for magicians.
giving them a complete new track for any frain they might get(which is easy to resist in most cases), grates on my magic hating nerves like not much else <.<

Well, the point would be to make Drain no longer so easy to resist in most cases.

Right now, as you note, most mages will typically only take Drain on a really bad roll—TN mods from damage are so awful that if a mage doesn't have an excellent chance of taking no Drain casting a spell at a given Force, they generally won't cast that spell at that Force. If that spell needs a high Force to be effective, they probably won't take that spell to begin with. This is why the Drain-Power-is-full-Force rule is so ineffective; there are a few spells that are sufficiently useful to make players change how they allocate Spell Pool, but in general it just moves more spells into the "useless because of hard-to-resist Drain" category.

The idea of the Drain Track is that all of a sudden taking Drain isn't the end of the world, so we can make it harder not to take any. You can cast spells that you might not be able to fully soak—but you're limited in how much Drain you can fail to soak. On the one hand, this does strengthen mages in the sense that a bunch of spells suddenly become practical, like the FooBall Combat spells or possibly the Elemental Manipulations, but on the other I feel it could weaken them by removing the ability to cast F6 M-damage Stunbolts all day with negligible chance of Drain (with only 7 dice on Drain resistance, you've got ~98.24% chance to resist Drain—you need to cast it 39 times before your chance of having taken any Drain rises above 50%. 7 dice is pretty easily achievable without Pool—a Dwarf mage can do it with raw Willpower alone—and it gets more reliable quickly as you add dice).

Plus, it would let us nerf the Trauma Damper.

There is some risk of there being something really powerful you can slip into those 10 boxes, and it also powers up mages in short bursts—in something like Food Fight a mage could open up and freely fill the Drain Track, knowing they'd probably be able to rest afterwards—but my initial evaluation is that this would reduce the power of mages while increasing their versatility somewhat.

Is your evaluation different?

~J

If they can cast a certain spell twice before passing out instead of once before passing out, that effectively DOUBLES their already considerable firepower quite a bit . .
If you can't resist that spell yet? tough, invest karma. or get magic loss after spending very much money on the trauma dampener. and then invest the karma to get back what you lost. your choice.
The drain track trainwreck would make it so mages do not need to ever spend karma to soak that drain, or spend money and then karma to get rid of the drain, but let them cast it, laugh it off and then cast it again a bit later.

A seperate Drain track would open up the possibility of having its TN modifiers at L,S,M and D levels only affect magic related tests instead of (nearly) all tests as in the case of stun and physical. But you'd really need to increase the likelyhood of taking drain in the first place in order to make this viable

Side note:Since you're toying with the idea of a third damage track, take a note for another one: The fatigue rules on athletics and their "virtual" effects on the stun track cause some oddities and complexity as well. So a seperate track would do no harm in terms of complexity and could remove the oddities.

So this is clearly another one of those cases where in the time since I last looked at a proposal I went and forgot most of the messy bits to the analysis. I'm splitting the Drain-track discussion off into another thread.

Also a progress update: I got somewhat busy, but also have been working on a crude character generator app that should speed things up substantially (and as a bonus maybe be convertible into a proper character generator afterwards).

Doing chargen by hand isn't too awful, though I'm going to need to come up with some good quick-reference cost charts for skills that include specializations since they're encouraged so much more under this system. One thing that's been mysteriously tricky, though, is remembering the marginal cost for improving an attribute; it may just be more than a decade of having "new value times two" drilled into me, but adjusting for racial bonuses/penalties keeps taking several seconds more effort than I expect. Not too awful; the only two cases in which it's really nasty are either making a bunch of characters or that final phase of chargen for inveterate optimizers where you start looking at tweaking numbers to see if there's a character you like better somewhere near what you already have.

The first case is what I'm doing a lot of, but most people probably won't encounter it. For those who still do, and those who run into the second case, I think the existence of a character generator is a sufficient response—it's not like it's fundamentally impractical to make a character by hand, after all.

~J

Ok, having brushed up on my linear programming I've given another go at the resource formula.

To start out with, I needed to identify desirable properties of a resource formula above and beyond matching the specific karma-nuyen exchanges listed above. Aside from some common-sense bookkeeping (you always get positive nuyen for karma, etc.), I identified three properties to start from:

Property 1: the marginal nuyen return is nondecreasing. You always get at least as much from spending one more karma as you did from last karma point you spent.

Property 2: the resource function is a step function with equal step lengths. That is, from 1-x karma you get n¥ per point, from x+1 to 2x you get m¥, etc.

Property 3: resource blocks come in at least vaguely tidy quantities. At no point does spending 1 karma get you 1,637 nuyen or something.

There are a number of other properties we might like to have, but I started there. That produced the following code, in Python using Sage:


(Apologies for the mess, I've spent the past few years in very un-Python-like languages (Haskell and ML-derivatives) and it shows.)

As it turns out, the "nice" resource levels are actually much nastier to work with, while the "naughty" levels have produced answers for a wider range of minimum increments and step lengths.

I had to maximize something, so I maximized the return on the highest resource level. Finding a better choice for this is likely to help further exploration.

This gave me the following, using the "naughty" values:

For 100¥ increments, 5-karma steps:

• 1-5: 300¥
• 6-10: 600¥
• 11-15: 700¥
• 16-20: 3.5l¥
• 21-25: 4k¥
• 26-30: 4.9k¥
• 31-35: 5k¥
• 36-40: 5.1k¥
• 41-45: 5.3k¥

And so on. This exposes some of the deficiencies in my constraints—the ratios between one marginal rate and the next are all over the place. I mean, they are in BeCKS as well, but BeCKS ranges from 1:3 change (500¥/pt to 1.5k¥/pt) up to 3:4 (15k¥ to 20k¥). Even in this snippet we've got 1:5 all the way up to 50:51, and that unevenness continues through the entire range.

Ideally I'd like to minimize the distance between the maximum and minimum ratios, but that's not a linear constraint; I'm going to have to look more closely to see if I can express something sufficiently like this as a linear constraint.

Also, the solution makes use of 100¥-increment values all the way up; I think a sliding constraint on the increment shouldn't be too hard, and if the result still has solutions that would make them much nicer.

So then, for 10-karma steps, 100-nuyen increments:

• 1-10: 800¥
• 11-20: 1k¥
• 21-30: 4.9k¥
• 31-40: 6k¥
• 41-50: 6.8k¥
• 51-60: 11.3k¥
• 61-70: 14.5k¥
• 71-80: 14.7k¥
• 81-90: 15k¥
• 91-100: 24.5k¥

Still pretty funky, but hints of something reasonable may be peeking through.

(For reference, I'd thought I'd managed to find a step+increment combination that gave a solution for the "nice" values, but I can't find it anymore—it's choking on everything I try to give it. Either I have a bug somewhere or "nice" is a fantastic misnomer.)

~J

What are those values, they are not following property 1 at all?

By the way, I don't see why it's important to fix that.

They're the nuyen you receive for spending an additional point of karma, based on the total amount of karma spent.

So by that chart, 1 karma gets you 800¥, 5 karma gets you 4k¥, 10 gets you 8k¥, but 11 gets you 9k¥ instead of 8,800¥. Since each return increment is larger than the previous one (800 < 1,000 < 4,900 < 6,000 < …), they do follow property 1; if you swap the first two rows (so 0-10 is 1,000¥ and 11-20 is 800¥), then it wouldn't obey property 1.

~J

Ah I see. I ve took your property one as:

•1-10: 800¥ to •11-20: 1k¥ => augmentation of 200 nuyen
•21-30: 4.9k¥ to •31-40: 6k¥ => augmentation of 1,1k per karma

and so on in order to make your function having an exponential curve.


But I still do not see what you want to achieve with that. It just looks counter intuitive. What's your goal exactely?

Well, I'm following the example of both BeCKS and point-build—going from 5BP to 10BP gets you 70k¥ in point-build, while going from 25BP to 30BP gets you 350k¥.

The goal is to have the return be supralinear. The multiples of increase don't need to be consistent or increasing (they certainly aren't in point-build; if we discard the 10x increase going from -5 to 0, they range from a high of 4.5 from 20k¥ to 90k¥ to a low of just over 1.5 from 650k¥ to 1M¥); I'd like to constrain them more, but need to either find ways to express something like that constraint linearly or start brushing up on nonlinear optimization.

I did have a go at maximizing the summed pairwise differences between the levels; at a 100¥ increment and a 10-karma step that gave me the following:

• 1-10: 100¥
• 11-20: 2k¥
• 21-30: 4.9k¥
• 31-40: 5k¥
• 41-50: 7.5k¥
• 51-60: 11.3k¥
• 61-70: 14.5k¥
• 71-80: 14.7k¥
• 81-90: 15k¥
• 91-100: 24.5k¥

Which is still far from ideal (several tiers still too close together; 100¥/karma at the bottom tier is just insulting) but shows some promise. There's a lot that can be gotten out of tweaking the optimization objective, but I'm deferring serious exploration there until I make it stop producing hundred-nuyen increments past the first few tiers.

~J

Adding sliding increments wipes out all possible solutions a lot faster than I thought. It looks like I'm going to need a better handle on the shape of the problem if I'm to have any hope of producing a clean solution.

Edit: actually, not necessarily. I just realized that the way I implemented sliding increments had incorrect interactions with other constraints resulting in stricter constraints than desirable (some possible solution space was being discarded). Will have to fix that and try again.

~J

So I've been mostly splitting my time between character-generator building and picking away at the edges of the resource question, but I had a sudden flash of insight while in the shower this morning: cheap, absorbable specializations might make Trolls a kind of mundane melee-Adept, especially if unlimited at chargen. I wanted to build one right quick as a proof-of-concept:

Assuming unrestricted specializations:

Troll (35kp)

Attributes:
190kp net

BOD: 6 (0)
QCK: 5 (36)
STR: 10 (40)
CHR: 4 (30)
INT: 5 (EA) (44)
WIL: 6 (40)

Reaction: 5, Combat Pool: 8

Skills:
139kp net

Polearms 5 (specialization 10) (40)
Unarmed Combat 5 (specialization 10) (40)
Throwing Weapons 4 (specialization 8) (26)
Stealth 5 (21)
Electronics 2 (4)
Etiquette 3 (8)

Contacts:
1kp net

2xL1 free
1xL1

Resources: 200k¥
50kp net

Edges/Flaws:
10kp net

Exceptional Attribute: Intelligence (10)

Not too shabby, I think. If we cap starting specializations at 7 that's 30 karma worth of spending shifted to post-chargen, but the individual levels come quickly.

For reference, if the base Attribute can be gotten to 12 (through 'ware, triple-cost spending, or Adept powers), a 6 (12) specialized skill costs 57 karma. That's actually a better result than I thought; I didn't think it was going to be so cheap as to be problematic, but there's a nice kind of symmetry that that last six points of specialization cost nearly as much as a whole new skill to 6.

~J

i don't really get what you mean with cheap absorbable specialisations . .
but 10 dice + combat pool in any skill on chargen is pretty OP in SR3.
And the Base Attribute can be gotten to much higher with Trolls.
STR6+4=10 already. Change MetaVariant and you are at 11, exceptional and you are at 12.
It's pretty trivial really. Trolls really WERE Powerhouses under SR3.

Well, the cheap part is base SR3: as long as you don't go above the linked attribute, a level of specialization is a third the price of the base skill. The difference absorbable specializations makes is that you don't lose karma if you improve a base skill you took a specialization on, so you can be aggressive about taking specializations, even on skills you know you want to raise later.


Eh. Not so bad, I'd argue, due to the costs involved and the restricted range of skills that…

Oh. Wow. I just realized that Heavy Weapons is Strength-linked. I mean, I'd still argue that the range is restricted, but it's a lot more open than I thought when I started my reply.

But anyway, pool use is capped by base skill, which weakens the power of high Specializations; it's powerful, but also not cheap, and Adepts have had the ability to field similar starting die totals on a wider range of skills without it being too bad. On the other hand, I guess there is some room to worry about the interaction with Muscle Toner. I'll look into it.


Not so much, I'd argue. Strength is the least valuable attribute, and non-absorbable specializations are really hard to justify for the kind of primary skill you'd want a high level in; you had a few potent niches (archer-artillery, polearm-devastator, throwing-adept, wallhacker), but especially given the costs (both in BP and in penalties to extremely valuable stats like Quickness and Intelligence) they tended to be very narrow.

With the proposed changes, I think they're set to be a lot more powerful, though there are still a bunch of drawbacks.

(That said, I'm still inclined towards some kind of cap on specializations at chargen—though Dwarf Mages with high Spellcasting worry me more than melee or Heavy Weapons Trolls.)

~J

Oh ho ho ho, no!
STR, under SR3, is very much a master Stat for any combat oriented Character.
STR based weapons are the end all be all of SR3 basically, because of the way how the Power and Damage Levels of the SR3 Damage-System work O.o
You can get a Troll up to 18 STR if you want to really hardmax. A certain Bow can, if you can get it, deal 22M Damage over 1.5 Kilometers.
A Pole-Arm with Dikote will Deal STR+4=22D Damage with a 3 reach combined. Granted, these are extremes, but this is doable in Char-Gen.
If you go for Ghoul too, it gets only worse. Depending on what melee rules you use, you can get a close combat attack up to STR+(STR/2)+Weapon-Modifier.
for a Troll, that could be 18+9+4=31S Damage. Add in 2 Net Hits and you are at 31D Damage.
And it Adds Recoil. At STR18 3 Points.
Out of combat, stuff like lifting/carrying Gear makes STR important too.

And yes, an Adept can get 6+6+6 = 18 dice Pool at char gen for several skills too.
Which is one of the most broken builds you can actually get under SR3 i think . .

I would have expected that you'd mention the two adept dwarfs with 26 to 36 melee dice without the use of combat pool (IMG:style_emoticons/default/grinbig.gif)

I was going by pure skill, improved skill and combat pool here, no bonus dice at all from anything else.
because these can be gotten by other people too at least in some parts i guess. that and i never was any good with magics <.<

Both dwarfs would get to 18 (21) dice without combat pool, if you were to remove all other sources.
And they both are my personal reference for testing the "abuse" potential of Jon's system, once it's finished.

More later, but where does the Dwarfness come in? You can hit 18 before pool with an Adept by Skill 6 + Improved Ability 6 + Ambidexterity 6 on a dual-wieldable weapon, but that still only gets you 6 points of usable pool.

~J

They come into (my) play once you try to get those "insane" pool sizes of 26 to 36 dice without combat pool on a starting character. The mentioned dwarfs are extremely min-/maxed (more ore less) one trick ponies that were built as proof of concept and in order to test certain aspects of chargen. Since BeCKS (and possibly your SeCKs) somehwat try to create more well rounded and consistant characters while putting up a higher bill for extreme specialist, they are perfect candidates for my personal testing of your final product. If the two can still be built in the same way, not much of a gain or a loss. If they can't be rebuilt, you achieved part of your goal. If they can be recreated and even expanded upon, then part of your mission failed.



I'd argue that improved ability explicitly raises the connected skill level and thus the correct number of combat pool dice for the main hand (since the dual-wield rules explicitly limit it to main weapon) is actually 12 (but starting characters not reaching that number under normal conditions). But for the mentioned dwarfs the restrictions of pool dice for the skill tests is rather unimportant because of the number of dice they roll without combat pool, which leaves them with the option of completely using it for dodging and damage resistance where combat pool dice aren't retricted in any form.

How do you get 26 dices? Huh What?

Seems there are people even better at Min/Maxing than me . . *jealous*