Bad statistics make a weak argument, the Illusion issue in SR4A Options

[pbangarth]

In several threads, people are arguing that the OR changes in SR4A make Illusion spells that affect objects unusable. Examples given to support this argument often use a dicepool of 14, saying this dicepool has a roughly 50% chance to affect OR 4, and is out to lunch for OR 6.

This would appear to be based on the idea that 14 dice X the 1/3 chance for each die comes up with 4 and a bit. I think this is seen as being halfway, as people say "on average, 4 hits". I thought about this a bit and figured people are not taking into account all the possible outcomes better than 4 hits. I wasn't sure of the math, so I asked a mathematician on Allexperts.com for a formula for "n dice needing at least h hits". He came back with three methods for calculating the answer. I am still communicating with him, as the first and third methods given currently (as of this posting) use 1/6 rather than 1/3, but I used the second method (long and tedious!) to come up with the following:

Dice pool of 14:

4 hits or better: 73.88 %

6 hits or better: 31.02 %

My conclusion from this? Illusion spells are NOT being nerfed/jerked around/made useless by the SR4A OR changes. They HAVE gone from near infallibility to being effective against even THE WORST STUFF almost one third of the time.

PS. I am extrapolating that the dice pool 18+ characters will still be in the 75% range for OR 6.

EDIT: This arithmetic applies to all spells' dice pools by the way.

[Draco18s]

After much head banging and trying to remember my permutations, combinations, and probability math that I learned WAY back in 8th grade I finally arrived at the same solution you have here. That was rather difficult, for the sake of not having to replicate this and having to learn, the steps are as follows:

1) Determine the number of wanted events (4+ dice at 5 or 6)
2) Divide by total events

Step 1 is the hardest. Here's how I did it, in excel I made collumns for 4, 5, 6, etc. dice (the number of successes) and found the total number of permutations that I can have out of 14 dice:

14!/(N!*(14-N)!) N being the number of dice above. 4 dice should have 1001 permutations.

Line below that I did
1/3^N * 2/3^(14-N)*P where N is the number of dice above, and P is the number of permutations in the last step. This gives us the probability of rolling this exact combination (N successes and 14-N non-successes) in any one of the permutations.

Summing gave me my total probability for "N or more" successes.

[BlueMax]

I think you are both wrong. The chance of successful illusion is 100%

Rob: I'll cast Illusion X
Jim: It didn't fool my sensors, try again.

Lather, Rinse, Repeat till 6 successes

Jim: Bob? Bob? where the frag are you?

BlueMax
/see post icon
//see me
///if you can

[Demerzel]

Boy, I wish I still had the PDFs I made of the probabilities hosted on the web somewhere (Damn DSL didn't come with webspace, I knew I shouldn't have switched). I put together some tables of Threshold vs. Dicepool and it have odds of success.

PM me and I'll email them to ya if you like.

The original post was here:
http://forums.dumpshock.com/index.php?show...;hl=probability

I still have the PDFs I just haven't taken the time to make sure they are hosted and relink them.

[Draco18s]

Just a quick note:

DP 12:
26.97% for 4
7.89% for 6

DP 16:
83.40% for 4
45.30% for 6

DP 18:
89.83% for 4
58.77% for 6

DP 20:
93.95% for 4
70.27% for 6

[Shinobi Killfist]

I think you are both wrong. The chance of successful illusion is 100%

Rob: I'll cast Illusion X
Jim: It didn't fool my sensors, try again.

Lather, Rinse, Repeat till 6 successes

Jim: Bob? Bob? where the frag are you?

BlueMax
/see post icon
//see me
///if you can

yet another reason i prefer the resisted test. Intuition+perception for people, sensor+clearsight for cameras and drones.

[Werewindlefr]

They HAVE gone from near infallibility to being effective against even THE WORST STUFF almost one third of the time.

71% with 14 dice is pretty far from near infallibility. Also, spells that only require 1 net hit (improved invisibility, for instance) are reasonably easy to defend against: ultrasounds, MAD, etc. 71% chance to succesfully cast a spell which is relatively limited in power? I think that's a good score, and that requires 14 dice, which is a nice dice pool already (my players, not being power players, throw 10 dice without edge for spellcasting tests, and sometimes even less.)
Other spells, such as flak/chaff, affect all sensors on a device, but they require several net hits to be efficient. I mean, to reduce the sensor rating by 2, you'd need 8 hits? Yeah right.
30% chance to fool only the visible light part of a camera, with 14 dice? Yeah right #2. I'll just play a hacker/mechanic instead.

4 hits is a lot, you don't need that very often against metahumans unless they have counterspelling backup, and even then it's a high threshold.

[suppenhuhn]

If i remember correctly 400 BP characters are supposed to be experienced shadowrunners.
14 dice for a new magician means soft maxed magic, soft maxed spellcasting plus specialisation and a totem or focus that adds another two dice.
To me that looks pretty much like a damn specialist in his field and i don't think being successful on a rather trivial task three quarter of the time is to much to ask for a specialist. Failing two third of the time is unacceptable though.
Also note that he now has to overcast for this spell to take effect.
Worst stuff for improved invisibility are things like ultrasound scanners though which aren't nor ever have been affected by this particular spell.
They also are quite cheap.
Back to your example, the average mage has a magic of 3 and spellcasting of 3 which gives him a chance of 1/3^6 to successfully fool a sensor.
Not everyone is a munchkin but thanks to such horrible rule changes anyone has to be apparently.
No magician at my table throws more then 15 dice and in my opinion that should be enough to somewhat reliably cast standard spells.

Frankly such high thresholds for such standard spells is probably the worst thing that could happen to the mundanes because someone that throws 20 dice at a camera (because he needs to) will throw similar amounts of dice when casting manabolt, mind control and summoning spirits.
That is something I don't want to see at my table, simple as that.

[pbangarth]

71% with 14 dice is pretty far from near infallibility.

73.88% against the top level of object in the old Object Resistance Table (SR4, p. 174) for an easily reached dice pool, without use of Edge.

Against OR 3 objects such as cameras, the percentage success was 89.47%. Did I overstate 89.47% as 'near infallible'? Would you bet big on those odds? I would. Yes, alright, I accept chastisement for an overreaction to overreactions. Let me rephrase 'near infallible' as 'really, really good odds'.

[pbangarth]

Back to your example, the average mage has a magic of 3 and spellcasting of 3 which gives him a chance of 1/3^6 to successfully fool a sensor.

Maybe I don't understand what 'average' is supposed to represent, in the context of running the shadows. If a mundane with AGI 3 and Pistol 3 wanted to hire himself out in the shadows, how long would he last?

Why should mages get preferential treatment?

[suppenhuhn]

Maybe I don't understand what 'average' is supposed to represent, in the context of running the shadows. If a mundane with AGI 3 and Pistol 3 wanted to hire himself out in the shadows, how long would he last?

If he's not an idiot he could last quite long actually.
The guys that bite the bullet are mostly inflated egos with some huge stats.
Also note that his odds of hitting your average sammy are way higher then the ones of the mage trying to be invisible to some average surveillance gear.

[pbangarth]

If he's not an idiot he could last quite long actually.
The guys that bite the bullet are mostly inflated egos with some huge stats.
Also note that his odds of hitting your average sammy are way higher then the ones of the mage trying to be invisible to some average surveillance gear.

(emphasis mine)

Well, I hear you there!

Let me put it another way. If a magician with Attribute and Skill at 3 each enters the shadows expecting to cast spells and make them work, should he have a higher expectation of success than a shooter with the equivalent stats?

[Zurai]

(emphasis mine)

Well, I hear you there!

Let me put it another way. If a magician with Attribute and Skill at 3 each enters the shadows expecting to cast spells and make them work, should he have a higher expectation of success than a shooter with the equivalent stats?

A shooter with a dice pool of 6 is far, FAR more likely to affect a typical target than an illusionist with a dice pool of 6. And assuming his SA pistol has any recoil comp at all, his second shot will be even more likely to hit and cause damage.

The shooter is firing 6 dice against (most commonly) 3-4 Reaction, for an average of ~1 net hit, which is all you need.

The mage is casting 6 dice against OR4 to OR6, for an average of -2 to -4 net hits.

[Werewindlefr]

73.88% against the top level of object in the old Object Resistance Table (SR4, p. 174) for an easily reached dice pool, without use of Edge.

Against OR 3 objects such as cameras, the percentage success was 89.47%. Did I overstate 89.47% as 'near infallible'? Would you bet big on those odds? I would. Yes, alright, I accept chastisement for an overreaction to overreactions. Let me rephrase 'near infallible' as 'really, really good odds'.

I'm pretty sure a camera is OR 4. OR 3 would be wires, capacitors, self-inductances, transistors, not complex assemblies. A camera in Shadowrun is closer to a computer than an early 20th century camera.
(OR 3: Advanced Plastics, Alloys. Electronics equipment is a vague term, but it's probably components and simple circuits, since computers, drones and vehicles are OR4)

[Demerzel]

6 dice attack against 4 dice defense gets 1 or more net hit 54.1% of the time, if you would like to be precise.

However, how many tricks does pistols 3 agility 3 get you compared to magic 3 spellcasting 3?

In this whole debate, the thing that has bothered me most is the comparison of bolts to bullets. I mean really, if you wanted to kill people then you should be a gunslinger. It's clearly superior to the task. But next time you find yourself on the roof of a building rigged with explosives that you know you only have 30 seconds to get off of before it collapses below you I'm going with the guy with 3 spell casting and 3 magic to levitate my hoop over the guy with 3 pistols and 3 agility to shot me off the roof.

I may be able to help you with a great deal of statistical analysis. I can tell you to three significant figures the odds of any attack pool up to 24 against any defense pool up to 20 by looking it up on this table I made. But I can't tell you if this basket of apples is greater than, less than or equal to that bushel of oranges.

[suppenhuhn]

(emphasis mine)

Well, I hear you there!

Let me put it another way. If a magician with Attribute and Skill at 3 each enters the shadows expecting to cast spells and make them work, should he have a higher expectation of success than a shooter with the equivalent stats?

The point is there are plenty of counters versus improved invisibility and the like that have a 100% chance of success and I don't mean any elaborate mumbo jumbo with 3 naked witches dancing around a bonfire but simple and cheap things like scanners that happen to work differently then cams and are thus immune to such spells.
Also mages can't cast bursts of invisibility and getting a focus isn't as trivial as getting a laser sight.
Additionally, as I mentioned before, most of the other spell categories get completely out of bounds when you throw such big dice pools and happen to not play pink mohawk. 20d6 in manipulation or combat are spells of almost epic proportions yet such dice pools are needed if a mage wants to do some low profile stuff with illusions.

[Mordinvan]

(emphasis mine)

Let me put it another way. If a magician with Attribute and Skill at 3 each enters the shadows expecting to cast spells and make them work, should he have a higher expectation of success than a shooter with the equivalent stats?

Well try calculating the odds of getting 6 successes on 6 dice. Its about 1/3^6 ~ 1/729 ~ .137%. Yep will have to cast the spell that means a 99.8628% chance of failure.
Shooting at a security guard generally doesn't have a threshold of 6. So might want to think about your comparisons a little more.
Our mage would have have to cast over 430 times to have a 50% chance of doing it once.
Think about it for a second and ask yourself it that is even close to reasonable.

[Zurai]

Not to mention he'd be forced to max-overcast to do it.

[Cain]

Not to mention he'd be forced to max-overcast to do it.

Well, he could spend Edge, but he'd still need to have a good roll. And Edge doesn't last forever (unless you're a Mr. Lucky build, that is).

[Zurai]

Well, he could spend Edge, but he'd still need to have a good roll. And Edge doesn't last forever (unless you're a Mr. Lucky build, that is).

Edge breaks the "you cannot get more hits than the Force of the spell" rule? I don't see that anywhere in my rulebook. A mage with Magic 3 is forced to overcast to even be able to use enough hits to affect cameras (OR4) or drones (OR6). Drones require a maximum overcast of 6 Force. Doesn't matter if you roll 50 dice - if you're casting a Force 3 spell, you get 3 hits, max.

[Cain]

Edge breaks the "you cannot get more hits than the Force of the spell" rule? I don't see that anywhere in my rulebook. A mage with Magic 3 is forced to overcast to even be able to use enough hits to affect cameras (OR4) or drones (OR6). Drones require a maximum overcast of 6 Force. Doesn't matter if you roll 50 dice - if you're casting a Force 3 spell, you get 3 hits, max.

I don't know the SR4.5 page reference, but check p172 of the BBB4.0. "This limitation does not apply to Edge dice that are used to boost a spell." It's at the very end of the Force section.

Still, being forced to use Edge is still a suboptimal fix. You'll be burning through Edge very quickly if you're using it on every routine illusion spell.

Edit: p182, SR4.5 has the exact same quote. Edge dice can break the Force cap.

[Ryu]

Feshy has written something very useful for SR statistics:Diceroller with statistics function

[AllTheNothing]

After much head banging and trying to remember my permutations, combinations, and probability math that I learned WAY back in 8th grade I finally arrived at the same solution you have here. That was rather difficult, for the sake of not having to replicate this and having to learn, the steps are as follows:

1) Determine the number of wanted events (4+ dice at 5 or 6)
2) Divide by total events

Step 1 is the hardest. Here's how I did it, in excel I made collumns for 4, 5, 6, etc. dice (the number of successes) and found the total number of permutations that I can have out of 14 dice:

14!/(N!*(14-N)!) N being the number of dice above. 4 dice should have 1001 permutations.

Line below that I did
1/3^N * 2/3^(14-N)*P where N is the number of dice above, and P is the number of permutations in the last step. This gives us the probability of rolling this exact combination (N successes and 14-N non-successes) in any one of the permutations.

Summing gave me my total probability for "N or more" successes.

I have to congratulate with you Mr. Wyrm, even the simple idea of going to search the formulas was enough to give me an headache, your post is very usefull; thanks for your commitment.

[Rotbart van Dain...]

In several threads, people are arguing that the OR changes in SR4A make Illusion spells that affect objects unusable. Examples given to support this argument often use a dicepool of 14, saying this dicepool has a roughly 50% chance to affect OR 4, and is out to lunch for OR 6.

Nope. We aren't even arguing that is that impossible to get 6 hits.

The point is that Hits (not Net Hits) are limited by Force, so to evade the standard Security Drones, most characters will either have to overcast or use Edge.

The other point is that it's silly that a Meta Link's Camera is more likely to spot invisible mages than a Security Camera - we'll end up with security design glueing Meta Links to the walls.

[Malicant]

73.88% against the top level of object in the old Object Resistance Table (SR4, p. 174) for an easily reached dice pool, without use of Edge.

Easily reacehd dice pool? Are you mad? I have a mage (Magic 6, Spellcasting 7, Powerfocus 2) who is by definition of his skill better than the best of the best. Yet, he will have trouble to fool technology with spells designed to fool technology. How does that make sense?
Sometimes it is not enough for rules to work (and this does only work for min/maxed cheese casters and even then barely), they need to work for regular people inseide the gameworld too. At least unless Fluff does not change to support "this spell will only fullfil it's function when you achieve mastery of magic at least of draconic standard"

DP of 14+ easily reached my ass.