I was running a game last night and the hacker hacked into a laptop, setting off a restricted alert, which in turn launch an IC whose only intent was to trace the hacker and report its findings to a spider. The IC was only rating 4 with a rating 3 track program, so 7 dice. The hacker, had a 5 stealth, so for the Tracking extended test, that acts as a negative modifier, so now the agent only has 2 dice to throw.

Based on RAW, extended tests would lose 1 die each roll attempt, so the agent gets 2 dice in IP 1, another die in IP 2 and then is done in IP 3, as it has no dice. The threshold is 10 to complete a trace.

Is this correct? I don't have a problem with it if it is, but it certainly seems that some of these lower rating nodes/agents are going to find it very improbably to trace a stealthed hacker. In turn, that makes me feel a hacker is risking a lot less than I previously anticipated.

Plus, with the extended test dice pool dropping by one each roll, if a hacker can perceive its being traced, he can basically focus on redirecting, thus pumping up that threshold and win in a battle of attrition...at least in regards to being traced.

You have to realize that as far as RAW is concerned, a stealth 5 program is about as powerful as it can be. So it makes sense that a basic tracking program will not be able to trace it. Often with traces it comes down to either being impossible to trace or being guaranteed to trace.

I thought the rule about losing dice every extended test was optional though, unless it got changed in 4A.

Unless you are a Technomancer, which is why we are basically Matrix Ninjas.

But yeah, if I am detected in a Node so far my primary worry has been getting hit by Black ICE while the system telecomms their best Security Hackers from offsite to run a Trace on me. White ICE really dont provoke that much of a resonse from me other than maybe having themselves shut down for a little while. Its just not worth doing much else against them, chances of getting a good Trace on me are low enough already (And thats without using any Proxies or other tricks)

IIRC, that's correct. I suspect the developers were sick of "extended test" being read as "keep rolling until you win," so they imposed the harsh -1 per roll penalty on extended tests. However, in this case, it's to the player's advantage, at least when the player is the one being traced. Let's see... assuming no redirect trace actions, and a spider with Computers 6 (specialty tracing hackers) and a rating 6 Trace program against a hacker with Stealth 6... That's 8+7+6+5+4+3+2+1 or 36 dice total. That's, on average, enough hits to complete the trace, but certainly not enough to buy those hits, and even a mildly successful action or two spent redirecting the trace would make the threshold essentially unreachable, as would using a proxy server (+4 threshold to be tracked, although you suffer a -1 penalty to Response for routing it through another server).

In short, if you've got a State of the Art Stealth program (or, god forbid, you're a technomancer) and you run through a proxy server, there's a good chance that even a tracing specialist won't be able to find you.

Can we please stop this "-1 dice per roll in an extended test is the writ of the rules" bullshit? The rules do not work that way.



Your GM may limit the number of rolls, and if they choose to do so then the suggested way to do this is to subtract a die for every roll in the test. The implication of the statement that you can (instead of must, will, or should) is that you are not going to under normal circumstances.

Fair enough, but I am the GM at my table and I have ruled it as all extended test have limited rolls this way.

Personally, I like the limit, and I just wanted to make sure I understood how I was ruling to work.

The problem is that it makes a lot of extended rolls impossible or close to it. Assume a reasonably competent professional Talismonger who is good enough to be in business, but isn't spectacular: Stat 3, Skill 3. They want to refine some radicals. Unless I miss my guess, that's a threshold 10 (I think? Memory is spotty) roll. Supposing the talismonger exhausts their rolls, they'll on average roll:
1: 6 dice, 2 successes.
2: 5 dice, 1.5 successes.
3: 4 dice, 1.2 or so success.
4: 3 dice, 1 success.
5: 2 dice, .6 succeses.
6: 1 die, .3 sucesses.

For a total of 6.6 successes. Which isn't enough to do basic radical refining.

I much prefer to say "You cannot roll more times than your skill x 1.5 (rounded up)" with no die dropping. So the enchanter would roll 6 dice 5 times and get, on average, 10 successes.

One subtle nit: "should" and "can" carry the same legal weight and are not on the same level as "must" or "shall". So when I see "should" in a rule, that means I can choose to use that case or not

8 dice - buys 2 successes
7 dice - buys 1
6 dice - buys 1
5 dice - buys 1
4 dice - buys 1

That's six successes there, after 5 tests, if you buy them. Which means your spider needs 4 more out of 6 dice, not bloody likely. However, if you use edge and roll it out...you might have a shot.

Edit: Half asleep and got some rules backwards