@Malicant: I'm fine with
Octopiii's entry. I believe that it does, as you all have agreed, actually help contradict
Hermit's argument given that its success relies on extensive cyberware.
But that's not to say it isn't a mage and it's certainly not to say that it isn't an impressive attempt. I think that
Octopiii would agree that it doesn't match a Samurai's capabilities, however.
For a start, there's the simple issue of staying power. He's well-armoued with 12 points of Ballistic, but that's still 17 dice to soak vs. Betty's 21 and that's mainly due the form-fitting body armour which puts
Octopiii's entry on the absolute limit of what can be worn without penalties. Betty can handle a bit more so given that I had about 3,100

left (plus a spare BP if I want it), I'm going to get Betty some Form Fitting Body Armour too at 900

for the half-suit. Betty's now up to 25 dice to soak against the replacement's 17. I hope
Octopiii does not feel I am being unfair. Certainly the replacement build could be modified to soak damage slightly better but it would be at the cost of something else. One of the advantages that Samurai have over magicians is that they can spare a few BP for rounding themselves out better. Anyway, this is a difference of 8 dice or 2-3 boxes of damage on each attack. I think we can say that Betty has sufficiently more staying power than the replacement that it can be counted as a point to Betty.
As an aside, this has just driven home the effect of the new Attribute scores. It makes me queasy just thinking about the cost of raising Betty's Body attribute now. Which is one of the first things I'd have gone for after boosting her skill list slighty if only so I could get more armour on her.So that's staying power in terms of damage taking. What about staying power in terms of damage output? I'm going to contrast the two builds in single target and area effect attacks. Betty will obviously use her assault rifle and grenade launcher while OMA can use Stunbolt and Stunball.
I'll note at the start an advantage that each has over the other. Betty does Physical damage so she'll have greater trouble with subduing opponents without harming them. OMA does Stun damage so has trouble when in a "kill and retreat" scenario, e.g. assasinating a target that is being ambushed, being driven away, etc. Both these scenarios come up in my games so the different characters would fare better in different situations. We're talking about the magician replacing the Samurai however - not them having advantages and disadvantages - so it's worth mentioning anything OMA has trouble with that Betty doesn't. Obviously the spells could be replaced with Physical versions, though that's a slight increase in drain and more pressure on
Octopiii's carefully calculated averages.

But anyway, moving on. We're looking at pre-Errata rules but for interest I might make notes about how things would be different post-Errata.
A decent sec-guard, with Body 3, Reaction 4 Will 3, Dodge 3 and an armour vest (6/4)
(SR4A example grunt).. Obviously not a threat, Betty will roll 16 dice, averaging 5.3 hits. Guard will average 1.3 hits, so net hits are 4 meaning that's 10P damage to soak. The guard will probably reduce that to 7P. (I'm using regular -1AP ammo here, so the guard only got 2.6 hits on soak and I rounded up in his favour). OMA uses the Stunbolt of course and gets 3 hits on average. Now OMA has a Magic of 3 so presumably we're going to overcast. If we don't, then OMA does on average 5S damage. We'll ignore that and assume overcasting to Force 6 which is what
Octopiii suggests. Much better now, with OMA still averaging 3 hits against the guard's 1, but getting 8S damage in return. Ignoring the Stun vs. Physical discussion, that's actually better than the Samurai. Drain will be 2P but with 10 dice to soak, he'll normally cover it. Neither party has dropped their opponent in the first attack however so they'll have to go round again. No problem for Betty - her recoil is easily compensated and her first attack was a Simple Action. So in her second Simple Action she'll just fire off another shot and very likely drop the guard doing even more damage now he has a reduced defense pool. It's no no problem for OMA either, but unfortunately her spells are Complex Actions so she has to wait for the next IP giving the guard a chance to shoot back (he only has a bad headache at this point) or raise the alarm. But there's another risk also. If you have 10 dice to soak 2P each time you cast your spell, how often, on average, will you fail? The chance of
not getting your 2 hits with 10 dice is 10.4%. So the chance of that occuring in two attempts is 19.72%. Basically there's a 1 in 5 chance of having taken 1 to 2P damage from drain in the two casts of Stunbolt. I'm going to discount that for now but we'll come back to it.
I'll mention that the magician was overcasting. Probably the equivalent to that wouldn't be Betty firing off a single shot, but doing a narrow burst. Her recoil comp will easily cover it so we could bump her actual damage up to 9P with a decent chance of 10P. It's quite possible for Betty to drop the guard in one pass which it is actually pretty difficult for OMA to do without spending Edge.
I'll also mention why OMA isn't using Manipulation magic to Mind Control the guard into shooting himself. Firstly it takes at least as long as OMA first has to get control and then secondly make the guard act. Worse, by a strict reading (and many GM's would enforce this), it will actually take a whole new combat round because in the first IP OMA takes control and it's not until the second IP, when the guard's own IPs are all used up, that OMA gets to issue an order. What? You thought Control Thoughts let you grant ordinary security guards extra IPs?

Also, the drain is worse even though partially compensated for by increased drain dice. Finally a minor point - although sustaining the spell doesn't matter in our abstract situation where both the Samurai and the Mage hopelessly outclass their single measly opponent, it is vaguely detrimental in more balanced situations.
So lets quickly scale this up to something more significant. Each of our builds has three IP so lets give them three guards each to face off against. OMA could keep using Stunbolt but at two IP per opponent, this is going to get her shot or the alarm raised or whatever. So she has two options - Mob Mind or Stunball. Mind Control is more appealing in this situation than it was before so we'll consider it, but lets look at Stunball first. We'll ignore casting at Magic rating - it's not going to work, and go straight to overcast. It's 4P drain with 10 dice to soak. OMA's is going to take damage for this even on average. What's he going to get in return? It will be the same on average as the Stunbolt was on the single opponent. That's 8S damage to each of them in the first IP. No-one's down but everyone is hurt. In the second IP, they can all be finished off (the chance of failing to affect one of these guards twice is low enough that I'll discount it for now, but it's not out of bounds). Two IPs, three guards unconcious, two lots of 4P drain to soak. Likely result of drain? 2P damage taken. Chance of actually soaking it all is less than 34%. Switch this to Mob Mind and you at least don't allow them their own actions even if you have to wait till next round to actually get rid of them, but your likely drain is actually worse than casting two successive overcast Stunballs.
Now we contrast with Betty. Betty also has two options - grenade or autofire. We'll do grenade first. We assumed that the three guards were within the six metre radius of the Stunball. We'll do similar for Betty. She's likely to get 5.33 net hits with the grenade launcher which at -2m scatter per hit, compensates for the average scatter (or it would completely eliminate it at -1m per hit if the grenade were merely thrown) So that's 10P at AP-2 for a HE grenade. That's a likely 8P damage for the guards on average. Probably we'll see some scatter around that average. Maybe one takes 6P, maybe another is dead. Anyway, we're seeing something comparable to the Stunball in a single IP but without a cost in drain. With the second Simple Action, Betty can probably bring down a second target with a Short Burst (at -2 dice for switching targets). We're quite probably seeing two targets dropped in the first IP. Second IP, Betty will probably drop any remaining targets.
I wont run the numbers as they're similar to earlier, but if Betty sticks to basic automatic fire without grenades, she can probably drop all three targets but it will use up all her IPs, to OMA's two.
Analysis? Stunball is better. Firstly it scales past three targets which Betty's grenade followed by burst fire does not. Also it's silent. But it will push the magician very close to or onto that first significant Wound modifier. Betty can keep firing all day long (and with her superior damage soaking, can take punishment for longer too).
Is OMA out-samuraing the samurai, yet? The build certainly isn't outclassing Betty yet as Hermit was insistent Magic did. Is it duplicating the role? Quite well all things considered. It's better at flattening a large number of opponents, but it takes some punishment for doing so whilst the Samurai has a lot more sustainability. So far, this is just confirming my "same but different" thoughts.
Aside: Under the new Errata, OMA will take more drain than before. Noticeably, actually. Likely an extra couple of points in our Stunball example, meaning it put her on -1 dice wound modifiers for taking out those three guards.But the above is a narrow set of circumstances. How do different things affect it? Well as noted already, if we move past three opponents, it starts taking the true samurai longer to deal with them all if using gunfire or launched grenades, though hand-thrown grenades are fine. But in juxtaposition, the more fights the magician has to engage in the worse things get. There's simply no way that the magician can keep up the same level of damage output turn after turn. Do other GM's class subsequent drain damage in a single mission as the same set of wounds so can't be repeatedly healed? Or is that just me?
What about better armoured opponents? Boost the guards' ballistic armour values up to 12 and the samurai has a harder time taking them down, though still not too awful. Remember that in all the above I was letting the magician overcast but not the samurai use anything other than regular ammunition. But better armour obviously doesn't hinder the magician at all. Against mentally tougher opponents, the magician suffers a little but again not too bad.
What about counterspelling? Suppose those three guards have a wage mage or a guardian spirit protecting them. Lets say they get a counterspelling of three dice? It only knocks the magician down by a box or two of damage when overcasting (though it makes non-overcasting pretty useless), but there's an annoying chance now that the magician will fail to do any damage at all. If the counterspelling rose to 5 dice however, this magician is in trouble. He's rolling 9 dice against 8 dice and needing net hits. That's close to a fifty-fifty chance of damaging and even damaging is going to be doing 7S damage now.
Visibility can be a factor, but this magician has low-light vision so except when battling in complete darkness, it's less of an issue (though again, the Samurai will be less affected).
What about drones? I can't even run the numbers on this one as OMA literally can't affect them. Other than to try and pick one up with Levitate, there's nothing he can do. He has Physical Mask but his chance of affecting the drone even under the pre-Errata ORs is very small. The samurai, needless to say, isn't affected.
As to the supplementary roles, OMA has a very limited skill set. He has no social skills at all. No athletics skills. He has hacking programs but no computer skills so would make a very poor hacker. I presume the Mental Manipulation spells are meant to substitute for the lack of social skills, but that's not going to fly in most games. It might get you past the guard but it's going to get you nothing but dangerous enemies with your contacts and Johnsons. To boot, the cost of attempting to fill these secondary roles and compete with the samurai in terms of IPs with cyberware, is that OMA is actually not a very good mage. OMA has no Binding skills and frankly crap Charisma making her a very poor Spirit summoner. She also can't deal very well with magical opposition, e.g. spirits and counterspelling enemy magicians. She is better than the Samurai in some of the ways a magician should be better than a Samurai, but not nearly as better as a proper magician should be. A decent comabt mage should be able to drop a room full of security guards and this one actually struggles to do that. Now that's fine because the challenge was to duplicate the samurai. But the point of the challenge was to show that this couldn't really be done very well without giving up a lot. OMA can more or less duplicate the Samurai but is more vulnerable, less able to sustain damage output, has certain circumstances that render him unable to perform his duty (e.g. counterspelling, background count, drones, vehicles) which the true samurai isn't subject to and has no roundedness to the character - e.g. supplementary skills such as Etiquette and Pilot Vehicle.
EDIT:Also, note that OMA has an Edge of 1 against Betty's Edge of 4. Definitely helps with the general theme we're seeing of Magicians being powerful but limited by odd circumstances and durability to a Samurai's steady and robust consistency and adaptability.
I'm open to criticism on all the above. I also have to compliment
Octopiii on getting closer to meeting the challenge than I expected people to get. But I consider this a partial success at considerable cost, imo.
K.