@Marcus, as I was writing my browser refreshed and I lost most of it, I will try to answer you again.
My system is no the same as Anarchy where the Glitch Die alones determines if there is a glitch. Here the Glitch Die is a condition die, if it is 1 then there could be a glitch, otherwise there is no chance of glitching.
This could be rewritten as:
"If you have 1s equal or more to the numbers of hits, roll 1 extra die, if it is 1 you glitched." But since I play a lot of The One Ring, Dragon Age, and a few RPG and boardgames other that have a special die roll with your pool, I choose to describe it as a condition die because it was more familiar to me.
You say that Shadowrun is a classic system. I played second edition in the 90's with target numbers (the same way as storyteller system was in the 90's), and it as confusing as hell (with failures happening with a single 1, exploding 6 where standard, etc) this was changed in the next edition, and changed again, and again. The system is changing, the dicepool grew bigger, than was limited, now is bigger again, but with limits. And the system is improving, but the odds are still a mess, who knows how the sixth edition will be? Specially now that we already have Anarchy with an alternative glitch system?
There nowhere in RAW (at least I never found) that says that the minimum pool is 1. Defaulting with atribute 1 could result in a 0 die pool, as a test with negative modifiers. You could say: "but then the player wont do the action to risk a critical failure", but what if this is an opposed test? Now there is a difference between a simple failure and a having a glitch to go with it.
About the half and 1/2 this is being strict, I'm not saying that someone would do that, but if you apply the Rules As Written, there is a rounding done there, and the RAW says to round up unless told otherwise.
You said: "No one rolls a die pool of 1, or a pools of 3, why b/c players know better". Take a look at the previous answers here in this thread. I begin my discussion talking about pools in the 12-16 range as the usual for players in stuff they are good at, and 6-10 in stuff they might have to test sometime. And a lot people here told me: "Ah, but in my game players roll 4 dice pool a lot, you never know the modifiers, and if they are forced to roll, etc, etc, etc". The smaller pools will have mostly the same odds they would have with RAW, the larger ones (the ones players do roll) will have more chance of glitching. Why? Again, because without that there is NO glitch in pools 10 and above, it simply doesn't happen. If the odds are below 1% in a game where each player makes what? 50 rolls at most in a session, it means that they will have a glitch (non critical) with a 10 dice pool ONCE every 2-3 sessions, IF they are unlucky and roll a lot of 10 size pool, and then there is always edge to go away with glitch.
The bell curve of success is still the same, the hit cancelation trims a little bit of it (mostly by turning a few of the glitches in critical glitches) but otherwise you wouldn't notice any difference in the ammount you success you had before vs what you have now. One very important point that is not apparent, is that for 1 hit to be cancelled there is a minimum of 4 dice involved, so it is quite impossible to a large ammount of success to have even one hit cancelled.
Finally you said that there is no jagged curve, I invite you to look at my table of odds:
https://docs.google.com/spreadsheets/d/1jwn8YAhCc39QDQ9mjxEBHVlLnb0APk-pi2UXPtXdM1U/edit#gid=0This is RAW and my system, below the table there is a graph.
Dark blue is RAW critical
Light blue is RAW glitches
Dark Purple is my criticals
Light Purple is my glitches.
Now I dare you to tell me the blue ones looks like something done on purpose... :/
If you want unlikely glitches and a better curve, you could use my system with a d12 as glitch die.
EDIT: Ops, when I lost the text I missed the explanation of dice sizes.
You can have any size of die for the Glitch Die, from a d4 (more glitches) to a d20 (glitches as rare as RAW). They all have nice probability curves, and all of those fixes the issues with RAW. With a d8 you will have just 75% of my glitches, with a d12 it will cut my system by half, and with a d20 it will be as rare as RAW with 30% of my glitches.
About why I choose the d6:
- A Shadowrun player has lots of d6 and for sure a few of them are diferent to work as glitch die, otherwise just throw it separately.
- The odds for the d6 are nice enough for my taste, not too much, not too little. And since glitches can be anything, it happens enough that I can use it to insert things as minor or as important as needed to the scene.