So, this is it. SGM 1.0 (beta), a complete matrix rules rewrite.

It features:
- logic connected tests
- AR slower than VR
- unified mechanics (hopefully)
- very few loopholes (also hopefully)
- a shitload of examples, example thresholds and baseline
- agents in check (worse than hackers, due to logic mechanics)


It is still in the writing and its definately not complete at the moment, but I think it is in a state to show somebody.

Disclaimer: This rule set is not the holy grail of matrix rules. I wrote it mostly because I could (and was asked). If it is better than the standard rules, everybody can judge for himself.

Feedback very much apprechiated.

On page 14 you say:
The threshold is 4 times the rating of the lowest
involved encryption program. The interval is 1
minute for live data and 1 hour for static data (as
more powerfull algorithms can be used there).

Shouldn't that be the highest involved encryption program? I mean, otherwise I could take an encrypted file, encrypt it with a rating 1 program, then break it trivially.


edit: Oh, and BTW... THANK YOU SANTA!!!

I think he means 2way comm. If one side has 3 and the other has 5, you go off the 3.
3*4=12TN

It was meant that way: If 3 people communicate with encryption 2, 4 and 6 ratings, 2 would be used for the decryption test.

Any way to say that in less than 2 sentences? Maybe add an example . . .

Oooh, I see what you're saying.
Check, I agree with that.
But you see how I got to my interpretation, right? I guess it's the difference between stored data and live traffic.

I think it's enough to say that encrypted communications effectively function at the rating of the lowest encryption program involved.

I didn't read it in details, but it looks nice. Your sample systems remind me of those I designed for my matrix security handbook .

You have taken a totally different approach from me. I'm not fond of closed rulesets and I there are some basic rules I don't like (the security tally for example) but I'll try to give you some feedback anyway.

For now, there's just one thing that really bothers me: from what I've read, a rating 5 agent is as good as a logic 5 + hacking 5 hacker. Am I missing anything ?

4+4 hacker (a hacker gets +2 dice for VR) without edge specialisation and any codeslinger advantages.

Furthermore ive balanced in an implant that adds +1 IP and another implant that gives -1 threshold (basically 1 hit that is not capped by programme) per test. Thats the one that also blakkie is propagating.

The maximum for an agent is 12 dice. The maximum for a hacker is somewhere arround 26 dice.

That's going to suck for TM's if they dont get an equal echo to have -1TN...
Though that's my first thought without reading over the file fully.

TM balancing is a different thing alltogether.
I think Ill implement my changes from SHP, which is an extreme boost frfor TMS from RAW.

As it is TMs are so underpowered at the moment, I would never recommend playing one.

Just a note, since you are talking about your SHP. I noted when you explained 'SECKS', you dont give out a 'baseline' Karma number like BeCKS does. Based off the calculations from the sample characters though, would it be safe to assume that 580k would be an acceptable substitution to 400BP, on average.

Thanks Serbitar, I'll check it out when I have more time.

Don't forget to update your sig. I still see 0.9 there.

I give a number, should be 600 karma if I remember correctly.

Regarding Active Alerts:
You may wish to re-evaluate having an Active Alert result in a decrease in System Rating. A +4 bonus to the firewall rating against identified intruders is nice, but is it worth lowering the maximum rating of the programs running on that node? Especially considering that node security often involves agents acting as IC?

If you're going to have Active Alerts impact a matrix attribute, I would think Response would be the better fit conceptually. Firewall is checking things more thoroughly, so you get slowdowns more easily.

Personally though, I wouldn't bother with having an impact on any matrix attribute. While previous editions had alerts that caused impacts on Host performance because they triggered effects that had Host-wide implications, such as boosts to IC ratings, SR4's Alerts are specifically targeted against individual detected intruders and only apply when the firewall rating is used against the intruder. In short, it's merely the firewall doing what it's designed to: making life hard for unauthorized users.

the problem is the following: Technically, you can just put all and every access ID on the "intruder" list and just leave active alert on forever. This would mean free (and extremely high) firewall rating. Thats why I wanted it to have some severe drawebacks. Response was not severe enough for me (especially for clusters and devices).


Maybe I can find something more suitable.

Thats an idea I could live with. Though you still have the question "If the firewall can do this +4 thing, and it doesnt have any drawbacks, why isnt it doing this all of the time against everybody?"

I agree with Runnerpaul.

Here are some more comments :

1) After thinking about it and taking a deep look at the probability tables, rolling logic+hacking and capping with the program level seems nice. It takes logic into account, remove the default problem, and blends with the basic game mechanics.

2)

Response to your [4], Blade. You have to remember that most matrix nodes, read servers, are way more powerful that a personal commlink. When you see the rating of the node 1-6, you are seeing the max amount of processing power it can give to each individual user/icon on that node. If they acted like personal commlinks, how would a company be able to operate with hundreds of personnel accesing the central database or have thousands of customers using their online store at the same time?

The way I deal with it, I consider that each connected user has, at his disposition, a full node of its rating.
For example, a corp office server with a rating of 3, will allow its 500 user to use it at the same time and each one of them will be able to use a rating 3 Edit progra while a rating 3 analyze program (used for system monitoring) and a rating 3 security agent will ensure the security of the user.

If you consider it as Serbitar does (or at least as I understood it), nothing prevent corps nodes from being full of CI and forcing the hacker to resist 300 analyze tests.

300 analyze test would be a -600 dice modificator in my system.

With my system, the chance to detect somebody actually decreases if you use more than 2-3 agents.

Didn't see this...
and 300 attack tests ?

Well, most of the time, the run is over when you are detected in a cluster. If it wanted it could shut down and throw you out and reboot in 3-6 seconds. So the question is rethorical anyways.
The best answer would be some fluff reasons about having too much IC is bad for overall performance and such . . .

But a question for you: What stops all the IC from your 300 seperate nodes that make up your cluster to go to the one node where a hacker is deteckted and attack him?

I really see no difference between your and my system . . .

Edit: I have a better answer. Attacks behave just like matrix perception. Every 2nd third and so on attack in a pass get a -2 dice pool modifier, for the same reason like matrix perception. Sounds good.
Thanks for the good argument.

You can send the IC to that nodes, but you'll be lowering the defenses in the other nodes of the clusters...

This way, you can send another hacker (or an agent) to distract security.

Compare it to physical security : to secure an appartment, you'll only need 1 or 2 guards. To secure a big building, you'll need a lot more. If you send the 300 guards of the big building in a room where you noticed something, the rest of the building won't be covered.

If you compare to real world security guards, they would send a few for support and sound the alarm... You wouldnt have 300 guards/police trying to cram into a single room to mob one guy...

@blade:
We are talking about attack IC. That is not used 99% of the time. The patrolling does the patrol IC.
And I think about 10 IC would be enough, leaving only 3% "unprotected".

Another question: Does one have to log on into each node in your cluster, or how do you navigate in the nodes? If one has to log on every time, thats nightmare before christmas for every hacker. If one doesnt, the IC can just go everywhere without even using an action (same node) and it is again equivalent to my system.

Do personae in the cluster see each other when they are in different nodes?

I consider clusters more or less the same as you do : if set up this way you can see the content of the other nodes and the other nodes can see you (but you can also set it up so that everyone feels like he is in his own node). You can't exactly travel in the node. I mean, you can go anywhere you want in the node, but you'd still have the same computing power with the same ressources linked to you (you'd still have you same Edit program).

So the IC could go everywhere and still be powered by their "node"? Whats the difference between your and my interpretation?

The difference is that there's a cap to the number of IC. If you send more IC somewhere, you'll end up lowering the defenses elsewhere (just like in a building) allowing for a second hacker to make his way.

Which means that most of the time the hacker won't face too many opposing IC.

You can even consider (at least, that's how I do) that attack commands can only affect "icons" in the same "node" as the IC. In that case, even if the node cluster can suspend every program in that particular node to allow for more CI without degrading response, the cluster node won't be able to send an army of CI in the single node.

My only criticism (and this is as much of the rules themselves as your guide to them) is that it seems like there are a lot of rolls that a hacker's got to make, just to (for example) edit the feed on a camera. I'd rather just set a difficulty for that and do it in a roll; otherwise you're back to the old problem of "decker-time".

Actually, that's one of the trouble with a closed rule set.
Hacking something is carrying out some kind of procedure (there aren't many different ways to act) and might transform hacking into a succession of rolls.

That's why you must not forget the almighty roleplay.

If you deal with combat this way :
- I shoot goon 1.
* Rolling ensues *
- Goon 1 is wounded. Goon 1 shoots you
* Rolling ensues *
- I shoot goon 1 once again.
* Rolling ensues *

this won't be very fun. It's the same with the matrix. It's not :
- I hack an account.
* Rolling ensues *
- You pass. There's a IC
- I spoof the IC
* Rolling ensues *

It's very interesting, any chance on brining back security sheaves/tallies?

Last night, I thought about a rule to quickly resolve hacking (the way you quickly resolve looking for information on the street by rolling Charisma+Etiquette(street))

The hacker rolls Logic+Hacking, the TR depends on the overall rating of the node :
Rating 1 => TR 1
Rating 2-3 => TR 2
Rating 4 => TR 3
Rating 5 => TR 4
Rating 6+ => TR5+

Modifiers :
VR HotSim : +2
(Hacker's program average rating - target node's firewall rating) / 2
Negative modifiers if the node has some additional security (CI, encryption, data bombs) : -1 to -4
AR : distraction modifiers
Wound modifiers
Probing the target (longer hacking time) : can trade dices for hits.

The numbers might need some tweaking, but it could be useful when you don't want to spend too much time to resolve hacking.

If there was a way to fit all the rolls of a matrix run in a single roll, we wouldn't have that much trouble.

Fast hacking rules that exactly fit regular hacking rules can't exist. You'd have to insert in a single roll :
The success probability of various logic+hacking rolls from the hacker, the success probability of firewall+analyze rolls from the node, the success probability of pilot+anaylze rolls from the CIs... The treshold of these test depending on other factors, and some test happening only if one succeded... That could be mathematically done... but that'd be absolutely awful.

So either you accept the simplification or you don't have any fast hacking rules.

As I like being able to resolve some hacking of lesser importance quicker than usual matrix actions, I feel a need for such a rule.

Since most of the hacking is resolved through logic+hacking rolls, it should be a logic+hacking roll. The most important part of the difficulty of a matrix run depends on the rating of the target node. That will be the TR. I chose these numbers, because I felt it went along with the difficulty table (hacking a basic node is simple, hacking an average node is average, hacking a secure node is hard, hacking especially secure node is extreme). Other factors that applies to hacking are taken into account through modifiers : program ratings, VR, additionnal security...

But if you feel like calculating the success probability of each action, the different outcomes, the success probability of each one, and so on, and mix all that in one single roll that takes into account each and every part of the standard rules, I'd be glad to see it.

Why not just make fast combat then too? One attack roll and one defense roll, if you dont beat the attack hits, you die. Could result in alot of dead runners though...

The matrix is pretty simplified as it is... If you really need to make it faster, just have your hacker player roll an assload of rolls for what he wants to do and write them down on paper and hand them to you all at once, then you can play with the rolls and just tell him what happens overall. You can interact with the other players while the hacker does all that.

So why limit data search to one roll and not play the entire search, rolling for each node you search, rolling etiquette each time you post in some BBS ?

Why only roll Charisma + Etiquette when looking for information on the streets. Why not roll it for each and every people you ask in each and every bar you go ?

I once had a shaman PC who wrestled each morning with a physad PC, it wasn't really interesting to play the entire combat so we just resolved that quickly with "fast combat" as you describe... So that's not that stupid.

EDIT : I may not have been clear enough, I don't mean that the matrix rules should be dumbed down to fast rules, I just say that it'd be nice to have a quick way to resolve hacking actions that aren't useful, or interesting or fun, or challenging. For example, if the hacker wants to hack every low-rating commlink in an area to search for something or if he just wants to have fun by hacking someone's fridge.

I understand what you mean, but if you are just trying to hack into some idiot's rating 2 commlink or rating 1 fridge, and you cant do it in 1 roll, I think you have a serious problem as a hacker...
Now if you are just doing it for RP purposes, like showing off to a friend or something, I wouldn't even make you roll to hack in if you are skilled enough. Like a Hacking skill 3+ shouldn't have any trouble getting into a fragging coffee maker to change the flavor of your friends drink to strained peas...

But if you are rank 1-2, yeah I would make you roll it out because you aren't that skilled. Hell, this could be the practice you need to upgrade your skill without doing a training roll.

The same with a good hacker trying to hack an average device...

BTW: A 5/5 Hacker in hot VR (12 dice) hast still a 10% chance to lose (not draw or win) in a test against a 3/3 host (6 dice).

negligible rolls start with having 10 more dice (or double the amount, whichever is less) than the opposition (less than 3-4% chance to lose a test).

*bump*

Any more progress Serbitar?

Yes, working on it constantly, but not enough at the moment to publish.
I am still covering up the holes to make it kind of "stand alone" so you can just use it as a matrix rules plugin with SR4 without having to read any of the matrix RAW.

OK, SGM 1.1 up now, with various new stuff. No change history yet, though.

At the moment I opted to leave the system reduction on active alert, to make active alert a choice and not an obligatory thing when discovered.

Furthermore Ive updated SHP to 1.10 (including size modifiers to be used with sensor targeting among other things).

Cool. I love these little goodies of yours.