Which means, even in the best case of a 10/1 having good odds of "cancelling each other out," it still means that they create a situation where the dice pool (and as such the stats of your character) don't fucking matter.

Say the difficulty 10 situation is a called shot to the head, with a pistol, in low light conditions (pretty common stuff, right?). You're trying to take out some leather-and-denim clad vampire in a back alley behind a nightclub somewhere, and you want the extra damage of a headshot.

Shouldn't it matter whether you've got a Dex of 5 (the absolute best a human being can ever have) and a Firearms of 5 (better than most DELTA operatives) or are a bumbling clumsy oaf with a Dex of 1 and an absolute 0 in Firearms and have never handled a gun before in your life?

Well, sorry, too bad. It doesn't.

Ok, I have to post because some of the things being posted here are doing nasty things to math. First of all, having 10's cancelled out by 1's does not mean that you will never succeed. You still have a finite chance of success, but it is a bitch to calculate for someone like me who has no training with probabilities (I can work them out though). Note: this is a flaw of almost every TN system where you are rolling more than 1 die. (d20 tests are child's play to calculate).
Secondly: Your pool size does matter if the number of successes matter. You might have the same odds of achieving 1 success as someone with a smaller pool, but (and this should be readily obvious) you can obtain more total successes(i.e someone with a DP of 3 can only hope for 3 successes [I guess this might be modified depending on how explosions were handled in oWoD]). I don't know if this matters in oWoD, but it did in SR3. Anyhow, this seems fairly academic as the system was revised and addressed some of these problems. You don't see people on these forms bitching about SR1 rules (this is just an example, I have no SR1 experience) without mentioning how they think latter systems did as far as correcting the perceived problem.

@Kagetenshi
I don't understand your hatred of SR4's modified pool, fixed TN system. Most characters (PC & NPC) have pools >= 4 and thus can succeed on all the tests I can think of (no thresholds higher than 4) [it should be noted that SR3 also had some checks that required a number of successes > 1 and thus were not possible for everybody, e.g. damage resistance]. Even if you are reduced to making a check you could not possibly succeed on, then you can spend edge to make the check. I am not familiar with occurrences where success was not at least possible (if improbable).

I will remind you that probability is very solidly within the realm of mathematics. If you want I can do the combinatorial analysis for you (after I'm done with my research paper, graded stuff comes first), but as a discrete math major, let me assure you that mathematically, you absolutely will, at times, succeed at TN 10 if 1s wipe out 10s.

Edit: the short, informal, and unsupported version is this: there exist combinations of possible dierolls such that the number of 1s, m, and the number of 10s, n, obey the inequality n>m≥0 , therefore you can succeed at the test.

~J

I think we can all agree that oWoD rules leave something to be desired and nWoD rules are an improvement.

@Kagetenshi
Thanks for clarifying your remarks about the SR systems. Just out of curiosity, what kind of system would you prefer? and is there an easy way of simulating it? It sounded like what you were describing would be a system where the numerical result of a test can be represented with a normal distribution of infinite range (I haven't taken statistics so I'm not sure if that is the proper way of saying it) and a mean result based on skill level and/or aptitude.

@Kagetenshi: Just out of curiosity, your biggest complaint about the SR4 system is that some actions can become impossible?

Not remotely. I'll clarify more when I gain full consciousness.

~J

Ok. It's hard to identify what I'd peg as my "biggest complaint", but impossibility is definitely not it. Some other complaints of similar or greater magnitude:

Glitching: adding one die to your pool will, half of the time, increase your chance of glitching, inverting the ordinary expectation.

Coarseness of the system: related to the impossibility issue but not identical. As I mentioned above, with an ordinary cap on threshold of four, the system cannot create odds of success less than 1.23% but greater than zero.

Abandonment of Shadowrun's roots/neophilia/fresh-off-of-Slashdot syndrome: the deemphasizing of cyberware, expansion of bio/geneware augmentation (an abomination which existed in third edition, certainly), the wireless matrix (which is nonsensical, particularly for the matrix as described), abuse of terminology ("hacker" and "firewall" being the two most obvious cases, as neither is remotely what the term claims it to be), ridiculous terminology ("rad" instead of "policlub", "technomancer" instead of "otaku" (though I'll grant that the now-well-known preexisting slang meaning does perhaps make the latter name undesirable)), RFID fetishism, the absolute ridiculousness of the Sakura Fubuki/metalstorm fetishism, it goes on and on and on.

The elimination of the most significant tool for tactical decision-making in the game (namely, pools).

The elimination of useful concealability ratings.

The dramatically smaller useful range of the ruleset (hard caps on attributes and skills at very low numbers, a system that cannot handle removal of those caps due to Immunity to Normal Modifiers at high pool sizes).

There's more, but homework calls.

~J

Ok. Thanks.