Is there a typo or did they purposely make Activesofts only available to purchase up to Rating 4 and Skillwires up to Rating 5? (If that's the case, do they expect people to fill up there Skillwires with a Rank 4 Skill, another Rank 4 Skill, and a Rank 2 Skill?)

If that is the case, can you decide to not use ALL of an Activesoft that you have? (You ran out of room to fit another Rank 4 Activesoft so can you cram it in as a Rank 2 Activesoft?)

1) It could be a typo. It's happened before, but you got the RAW right. r5 Activesofts would be really powerful and I think the cap is fine

2) I'd expect that they could scale. Regular programs scale down if you system can't handle the full rating so I'd let it fly at my table. There is nothing to say that it does so I guess that would be a house rule technically.

Oh I absolutely agree rank 5 would be powerful... But that's also an extra 3,000 nuyen (making each skill 15,000 nuyen)... That's a bit pricey even for that much power. I'm just curious as why they would make Skillwires able to go to Rank 5 yet the only Software that can be used with it (Activesofts) can only go up to Rank 4.

It'd be nice if we could get an official from CGL to comment on this (Unless that was a flaw left over from FanPro that was never addressed)

The problem is cracking software is easy. So if you have 3 people on the team with Skillwires then they only pay 5K each and the hacker just cracks it for them. I think the reason skill wires go higher is so that more Skillsofts can be used at once. Rating 5 is 1 level 4 and 2 level 3's.

Probably because you can theoretically program your own activesofts. You can get a lot of game-breaking equipment by coding/building it after game starts.

The reason that skillwires can have a higher rating than the activesofts that require them is because you can load a maximum total activesoft ratings of 2 times the skillwire's rating.

So, while your activesofts will never be higher than 4, you can load more of them by having skillwires rating 5--up to 10 total ratings. As Jonny Reload said in his original post, that could be 3 skills at ratings 4, 4, and 2, or any other combination that sums to 10.

I think the OP thinks it's silly to for the capacity of max ranked skillwires to be 2.5 max ranked activesofts. I don't think the answer you want exists ... other than that the dev's don't think it's silly.

On a semi-related note. What do you think it takes to create an activesoft. The software table doesn't list a time, and software skill doesn't say that you can specifically. Activesoft's gear description says "Recording and programming physical skills". Is the recording device just a sim-rig, and then people use the software skill to turn data into an executable? I'm inclined to think that it requires more resources in some way because I feel like it makes them fare to common. I don't much about software buisness models but if a rating 4 activesoft goes for 12,000. All the programs I know that retail for 12k a pop are pretty colossal projects.

Or the people making them just like outrageous markups.

It requires many thousands of recordings (from many different people) to smooth out the motions and get every possible reaction into the soft, and then man-years of effort to optimise the decision trees that select between the platonic ideals of the movements. Automated tools would help, but a manual touch is necessary to make a quality end product.

At least, that's how I'd run it. As a CompSci student I know how much effort it takes to create any significant software ready for real world use.

And we don't have expert systems, VR, or the processing power of the 6th world. Some of the things they are doing now with genetic programming are downright amazing, and if you add in the processing power of the 6th world they become even better. VR allows "intuitive" coding, in that the coder can grab bits and pieces of code and combine them, with the expert systems and processing power smoothing out the problems. And a rating 6 agent is damn near an AI, a few custom ones designed for software design allow for a lot of other interesting things.

So basically, how coding works and how difficult it is for us today has no real bearing on how hard it is in the 6th world in 2070.

Sure its difficult, but a signal individual can code an entire rating 6 Agent in 6 months without much optimization (Logic 5(, Software 5(7), PuSHeD (1), Facility). That means you can buy 5 hits per interval and with a rush job the time is halved, so 1.5 months per interval. And coding an agent (based on price) isn't that much less difficult than an Activesoft (2.5K per point vs. 3K per point). With a group it gets even easier.

I thought you could get R6 skillwires. Which makes 3 x R4 skillsofts perfectly valid. That and R6 active softs would be totally OP.

how hard is it to code a piece of level 5 active soft?
could one with a logic of 4 and a skillwire-set and bought skill of 4 just like that write up a level 5 skill? O.o

Personally, I find the coding rules a little ridiculous, particularly when it comes to skillsofts and agents. They're their to provide an interesting pasttime for PCs that mechanical effects to balance it against other runner hobbies like drone customisation. The problem with it is that at the very least there should be costs, "materials" and inconvenience required. At the very least, if you want to make a rating 5 activesoft, you'll need to find someone with at least that rank, possibly higher, and convince them that they want to contribute to the project the tens or hundreds of hours required to record their motions and experience. The programmer then needs to codify each of those actions into a response and a system that allows the skillwire system to pick the appropriate response to the opportunity. Given this I'd probably ask that anyone programming it should probably have an appropriate knowledge skill at a good rating, access to at least one participant willing to put in large quantities of time to record the muscle memories. Without these things, autosoft coding just ain't happening. Even with these your single contributor autosoft may not be fully effective with differing body type, for instance if you give the little, low-strength guy a Unarmed Combat soft recorded solely from a troll, power-based wrestler they won't get the full gains the skill offers.

Making softs can't be easy going because their existense makes the game world different on a basic level, removing the concept of unskilled labour, but it is not an assumption of the setting that everyone uses them, because it remains more cost efficient in monetary terms at all levels for firms and individuals to invest in training. More than the cost effects, it is the huge social effects of softs that we need to contain, because cheap 'wires and 'softs move the average dicepool for an NPC in a proffessional skill from 4-6 to 8 at a cost of 10K Nuyen plus 15k per skill (This is assuming no modifiers for pulk purchasing and licensing and no breaking of copy protection or that the firm's own software). This sounds like a small change, but it means that you expect NPCs in day-to-day situations to complete extended tests twice as fast. Yes, this is a sizeable investment, but they are literally getting twice as much work done and there are a number of areas where the present discounted value of that to a firm over years of a unit of labour's output that this becomes worthwhile and the other advantages such as being able to shift workers between linked areas and have them maintain their ability with no additional training time which increases the homogeneity of labour and reduces the costs of worker turnover and absense. If a firm does have its own relevant rating 5 skillsoft, then for less than six months' pay to a low lifestyle individual, meaing that for large firms who can reduce the fixed cost of developing the soft over a large number of employees or if the firm can appropiate income through making the soft available externally, then even under RAW it can make sense to have them pay to make workers more productive, since over the course of less than a single year they can make the money back through increased productivity.

Obviously with wires costing essence they aren't for everyone, but in terms of other downsides, more than any other mod they control your body, able to interpret and overule your mind and reflexes. If they get hacked, the effects are probably comparable to pscyotropic IC, only without the need for you to ever use VR.

Sorry, stream of consciousness.

I don't think PCs can program activesofts just by rolling their software. You would need some source for the sim recordings that form the basis of the softs. You can't make that by yourself just typing away on a keyboard, it requires lots of time and effort and potentially special equipment. Though it might be possible to yoink that data from someone, and then you could program the activesoft yourself.

if you can code something like that you're usually a hacker anyway, so just break into some software forge and grab what's up

This is why I see rating 4 being the highest level you can get an activesoft for.

An active soft can provide muscle memory and the like, but it can't provide all of those other things, let's take driving. The active soft let's you dive really well but it can't make you look for openings in traffic, note the small oil patch on the road, read that the car in front of you is preparing to change lanes, etc. All those little things that add up and separate the elite from the rest.

I have an interest in AI; expert systems exist already. They are, literally, codified expert knowledge - it analyses the significant details as an expert in the area would and produces the result that an expert would, just a hell of a lot faster than an expert because human cognition on a single task is sloooow.

As for "intuitive" coding practices; programmers today use IDEs and typed code even when there would be a market for visual programming metaphors (indeed, Garry's Mod has an addon that implements one) were it more efficient - this implies that they are not more efficient. The grind is not going to change all that much just because of VR. We already make use of premade blocks of code, we in the industry call them "libraries" and there are so many of them it would take years to get your head around them all. The Torronto gaming Jam provides a list of suggested libraries that is as long as your arm - and these are a small selection of the multimedia and gaming-related libraries available.

I can rationalise the "rating" of a "program" (because like hell you use the same program to edit camera feeds and text) being the size/variety of the suite of tools you're using to perform an action and software skill being used to tie in new additions instead of actually coding - which would require knowledge skills in programming, software engineering, software design and an appropriate subject knowledge. Otherwise the software industry just doesn't work under RAW - half the time you'll be scrabbling to find bespoke jobs for your top end coders to blast through at lightspeed because you've written all the programs worth writing.

This way your power user can upgrade his programs, but still pays material costs (paying to acquire the components) and takes time (it takes a long while to get your settings right and learn how to use your new tools properly), and programmers don't spend half their life sitting on their hands.

i doubt programmers ever sit on their hands... they find and fix bugs and security holes, they update new versions to include the most relevant new data... an activesoft in a quickly-changing field will need a lot of updates to stay current. others will be more basic, requiring only small patches to include small bits of information that have added to the field.

in a world of hard caps, having a rating 4 is still quite good. its better than a professional level of skill. if your unskilled, then you just got 5 extra dice to use.

If we do see a rating 5 activesoft, i would expect it to behave as skill 4 with a specialization. not unbalancing, but quite useful.

Bugs are not found at a linear rate, they are a randomly distributed occurrence. And 20% of bugs account for 80% of reports over the lifetime of the product, whereas the remaining 80% will only be responsible for 20% of all bug reports. A thought occurs; the 20-80 rule is recursive. This means that 4% of bugs account for 64% of the reports, with a further 32% being responsible for 32% of reports (there were two groups of 16:16, so I merged them), and a whopping 64% being responsible for a meagre 4% of reports. You can apply this repeatedly as an approximation.

We can assume that bugs will be noticed in an order pretty much in line with their reportage percentage (which is close to their expression rate), so the four percent that contrbiute the most reports will be fixed rapidly as they are most likely to be reported first, they'll probably get caught in pre-release testing. Next the 32% that contribute an equal percentage of the reports will be get around to being fixed; but the bug expression rate is only about 36% of the initial rate, and it takes about 16 times as long to fix all of these bugs - the lifecycle of each bug is about 2 times the length of the bugs in the first 4%. Now the remaining 64% of the bugs will take 256 times as long as the first 4% because their expression rate is so much lower than the first 4%, each each individual bug takes about 16 times as long as in the first 4% to express, get reported and get fixed (the fixing time is assumed to be pretty much negligible for every bug).

That's an awful lot of hand-sitting if you ask me. Especially as the programmers, by certain interpretations of RAW, have written every useful program possible (they're available at chargen, afterall). Security holes have the same lifecycles as bugs.

Let's not even get into how a maintenance programmer needs a completely different skillset and mindset to a programmer in the process of writing the damn code, just as a janitor uses a different skillset to the construction labourers. These skillsets are not all that transferrable, so a programmer used to writing new code will be less effective as a maintenance programmer. It's inefficient to tell your application writers to maintain code for the rest of their lives.

(Let me say now that I understand that the 20-80 rule is an approximation, but it fits for a number of phenomena in the real world - it's a useful rule of thumb.)


As for integrating new information; under the "6 months for Agent 6" model this will take a short period of time. Certainly functionality will follow the same pattern as bugs and security holes; you'll take 16 times as long to come up with each feature making up the tail end 64% even though the implementation time is assumed to be pretty negligible in comparison. A lot of hand-sitting results because SOTA advancement in any single area slows down over time.

Now, one could use a smaller team of programmer and shuffle them between projects - this has conflicts with the fluff because it's assumed that corps have large code farms that absorb most of the programming talent. There is no reason for large project teams if a lone person can write rating 6 programs in a matter of months (instead of centuries) and so the demand for programmers will be smaller than for employees in an area which have slightly more serious investment requirements. You'd find that most universities will not teach Computer Science because there would be few enough jobs in the business that most people do not want to learn how to program (main function of CompSci courses, no matter what they say) and so there'd be little in the way of profit oppurtunity for the universities to provide the course.

The only way to break from the hand-sitting is to require that most of the software development business does not fall under the software skill and therefore does not take ridiculously small amounts of effort for lone programmers to produce new products. One has to assume that code farms are truly necessary to write anything of appreciable worth. This also takes away the capacity for PC hackers to claim that they can expand their capabilities for free, which is a bit imbalancing.

My understanding is that Activesofts are SIMSENSE programs, not just regular computer programs. In order to create the Simsense portion of the activesoft, you need an actor, not a programmer. That's why they are the most costly programs, even more than pilots. That's why they are the most difficult to make, why the rating doesn't go above 4. If you think the cliche "those than can do, those that can't teach" has even a little bit of truth, you might imagine that the highly rated performers are too busy to create activesofts.

Examples: Wayne Gretzky and Michael Jordan are listed in the skill ratings table at level 7. I imagine that this was when they were in their prime. Do you think these athletes were still rating 7 when they retired? Do you think they needed the work?

Another thought occurs, would it really be so unbalancing to just make a super skillwire system that can use more at the same time, just op the rating of the regular system to 6 or 8 or whatever of make a new one?
I've always liked the idea of having certain skill sets that you could load (like a B&E Set, or a Pilot Set) and just switching them around, but it's quite hard because of the 10 skill levels limit on skilwires...

I think the brokeness of skillwires is that the other downsides aren't really a problem. Expert system gets around all the edge issue to a large degree and having to spend an action to switch softs is not really a major problem as in most situations you can affort the time, so the only controlling factor is the limit in total ranks and maximum rank. You could allow players to "unbalance" the rating points to increase capacity, but the aim of rating 5 wires is already along similar lines as they only improve capacity. On the otherhand I can't see anything preventing you from having multiple skillwires.

Keep in mind that the Expect System only allows a person to spend Edge to reroll a failed test. It doesn't allow any other Edge use.

In my experience, the most common use of edge is to pull a terrible roll up towards its expected value. Particularly when rating 4 softs mean you probably have a reasonable poolin the 6-10 range before mods.

Note that the Expert System's wording specifies failed test, not failed dice. If you get a terrible but otherwise minorly successful roll, you are stuck with it.

Yeah, but all tests are either threshold or opposed. There's not much reason to roll Edge if you succeed. This doesn't even get into Extended tests, either.

Where what you say comes into play is if you botch. You succeed with a nasty caveat. You can't spend Edge to negate that extra annoyance, by your interpretation. However, you can still spend Edge on a critical botch, since that by definition is a failure.

Glitching is specifically what I was referring to.

Might I ask what you meant by the comment I bolded above? I know that you and I have been over this exact subject before, so you know I am not making things up, or spouting house rules. As such, I am at a loss in understanding your inclusion of that particular qualifier.

If it will help, I gladly quote the text in question ...

Actually, I can't recall us debating on this particular topic before. This isn't the first time you've associated me with an argument I haven't participated in; you did the same thing in an argument over flechette rounds. Just because you lost an argument doesn't mean it was me!

Back on topic: the problem is that there may not be a clear and solid demarcation between a failed "test" and failed "dice".

For example, let's say we have an Extended Test. The character makes his first roll, and does not make the threshold. Has he failed? If he has, he can then spend Edge. If not, by your reading, he can't; but when is the test failed to the point where he can spend Edge? Where is the dividing line?

Not to mention that your interpretation seems awfully wonky, if you can spend Edge on a critical botch, but not a plain everyday one.

Alternately, just because you want to deny (or have forgotten) it doesn't mean it didn't happen.


What crap! Can you give me any other example where canon has confused the word 'test' with the word 'dice'? Can you show me in the errata where this supposed error was fixed? I think you're reaching.


I'm not quite sure what canon has to say concerning Edge use and Extended tests. Can you give me a quote or page number that deals with this subject?



Wonky or not, my interpretation is canon! I even provided the quote to back it up. You however, have not provided any canon evidence to the contrary.

The only reasons that I can see why you would continue to call it 'my interpretation' are either to spew your usual crap all over the SR4 system, as is your usual modus operandi, or it's personal. I'm kind of hoping it's the former.

I have to agree with Fortune on this, you can spend edge on a critical botch but not a regular one. As for extended tests, how does edge interact with those in the first place (I'm honestly asking because I've never got a straight answer on that point).

I shouldn't have to, I expect that you'd know at least the basic Edge rules by now. However, if you want to get technical, there is absolutely no mention of "failed tests" on p67 of the BBB. Instead, all it says is that one option is reroll all the dice that did not score a success. It doesn't say anything that makes a difference if the test succeeded or failed.

As a piece of friendly advice, you really should know the technicalities of a rule before you start acting like an expert on it.

First of all, *that* was personal. An Ad Hominem in the most classic sense. You need more self esteem, if you're so sensitive to counterarguments that you'll fly off the handle like that, and resort to insults and illogic.

Second, I will continue to call it "your interpretation" until you can provide proof-- actual proof, not just a single twisted quote-- that demonstrates that there is a clear and convincing demarcation between failed dice and a failed test, *and* that Edge treats the two differently.

You aren't Toturi, by any stretch of the imagination. You are not the book-ninja of Canon-Fu. You just made another mistake, that's all.

Dunno about the Extended test; that's just another weak spot in the rules IMO.

But as for your first point, here's the rule from p67:

That's all fine and dandy, but I never asked about Edge. I asked for an example of where canon confuses the word 'dice' with the word 'test'. Your attempt to side-track the conversation with insults failed this time.


That's correct, which is why the description of the Skillwire Expert System explains how it works in explicit detail. Note that it doesn't even reference the Edge rules, nor give the page number in the BBB, as any Edge use other than what is described is specifically barred.


It's quite funny how you stoop to the same tired tactics whenever an argument goes against you in any way.

As for your response to Emperor Tippy ...


He was specifically refering to Edge use with Skillwires and an Expert System, Skillsofts being the topic of this thread.

Skillwire Expert System: allows to use Edge to re-roll a failed test.

Failed test: One that did not reach success.
Re-roll a test: Just that.

Unless we get a declaration of intent, this seems to be a completely different use of edge than those uses offered by the main book. It´s re-rolling the failed dice, only that you have to re-roll the successes, too. Quite sensible to keep skillsofts from replacing real skill...

You get to re-roll critical glitches (failure for sure). You get to re-roll an opposed test if you loose.

You do not get to re-roll: glitches (flawed success), tests where you simply want more successes (say combat).


Things that I am not clear about:
- If you re-roll an opposed test (as opposed to failed dice), does the opponent have to / get to re-roll his dice, too? Him rolling is part of the opposed test.

- How is re-rolling an extended test done? I´d personally limit the expert system to "re-roll a single time in face of a critical glitch, or re-roll the last test". This is a consequence of us handling extended tests as arrays of normal tests, according to the definition of extended tests. Only critical glitches and too little successes on the last test would bring about failure and enable the re-roll option.

Not sensible: developing a totally new use for Edge that in some ways is *better* than the normal ones. The way I've seen it done, if you botch or critically botch, you can't spend Edge on a reroll; you can only negate it. Sensible would be to use the existing rules, instead of going for a wonky, highly-legalistic, and confusing new angle that has nothing to do with the current rules.

By restricting Edge usage to just rerolling failed dice, skillsofts are still not as cool as real skills.

If you're saying that pulling an ill-defined, totally new, non-working, completely-out-of-the-arse contradiction to well-determined rules is not sensible, then I agree with you.


Technically speaking, there isn't a "last test", you either meet the threshold or give up. The test cap is an optional rule that isn't even encouraged to be brought into play all the time. But if you want to be pedantic about the whole thing, the extended test is considered all one test. So, you should get to reroll all the dice you used.

I think that the "failed test" incorporates by reference the edge rules. It doesn't create new Edge rules. Just because they didn't include the page number doesn't mean they weren't trying to incorporate the original rules. I think the sillier assumption is that they were trying to create a whole new way of rolling Edge with two words, "failed test." Developers do not hide elephants in mouse holes. They don't change the landscape of Edge rolling with two words.

This is an ambiguity, to be sure. The real question is, should you allow two words, without any explicit declaration, to create a new subset of Edge rules? Or should you assume that the two words are referring, albeit inartfully, to the existing rules? They both require you to leap a logical gap, but I think the latter is clearly the shorter gap to leap and therefore the correct one.

1. The opponent wouldn't reroll their dice. Only the expert system user would reroll.
2. The BBB describes an extended test as a "series of tests", which means that multiple tests are involved. Which would mean that you could use edge on any test in the series in the normal fashion.



I don't understand how rerolling a test is better than rerolling failures. While you might think that using the normal reroll rules is sensible and such, the developers must have felt differently. It's a cold hard fact of RPGs that no matter what you think is sensible, it doesn't mean that the developers will share your feelings. Arguing that something is more sensible only works if you're discussing house rules, and not for discussing what is said in a rule book.



Obviously, you don't know game developers well enough. Not only would they hide elephants in mouse holes, but they'd paste fake antlers on them and tell you that they were their pet jackalope "Charlie".

The problem is more than two words. There is no longer such a thing as a 'failed' die anymore. If a die is a 5 or a 6 it is a 'hit' now. So, if we were wanting to use the normal edge rules they should have said: "re-roll all dice that did not achieve a hit". That, my friends, is an elephant in a mouse hole if I ever saw one.

"Reroll all non-successes" is pretty much how the Edge rule is worded. My argument is that from a gameplay perspective, it's more sensible, easier, and smoother to use the existing rules rather than half-assedly describe a totally new one just for skillwires. You're right that developer intent might be something else, though.

Yes, playing with limited extended tests seems to be an optional rule, and is one we use. Anyway the extended test is defined as a set of tests, so any other rules refer to a single test in the row.

You are right that re-rolling a whole test instead of negating a glitch makes that use of edge way more efficient. And you are right that it is a new mechanic. Enhanced game mechanics as effect of augmentation does however not sound wrong in my ears.

Of course you can reroll or get additional dice. You just can't be sure that the (critical) glitch will go away.