JOW is 50% chance, or 37 mana per wand attack, right?

If that's true, then the speed is easily the most important part of wanding back your mana, much more important then stats. The [Consecrated Wand] from a duskwood ally only quest will get your mana back the fastest, as it's the only 1.2 wand in the game (30.8 mana/s). Next best is the [Wand of the Seer], a netherstorm quest reward available to both factions (29 mana/s). It's also the highest damage 1.3 speed wand. If you have trashed that for some reason, the next best is the Sporeggar rep one [Sporeling's Firestick] (28.5 mana/s). Assuming I got JOW right, it will easily beat any epic wand with regen stats, based on its superior speed.

None of these are shadow wands, there isn't much of an excuse to use one when the firestick is clearly superior and easily obtainable.

Imagine if you could use a scope on a wand!
Stabilized Eternium Scope

Apologies. I quoted WOWWIKI, which is obviously out of date. If you'll point me to the correct description I'll be glad to correct my post.

Thanks, Dennis. Very useful information. You've convinced me, and I'll be spreading the word within my guild as well.

Any guild with a warlock in it relies on ISB for additional shadow damage, every lock in your raid is assuming at least 30-40% ISB uptime if they're casting shadowbolts otherwise they'd more than likely be better off casting incinerate.

SW: P does not consume charges, SW: D and MB do but they certainly don't consume charges on the order of once every 1.3-1.5 secs for a measly 20% of a couple hundred damage.

If you've never played a warlock it's more than likely you had no idea what ISB was or how vital it is to any build spamming shadowbolts (read: all serious warlock raiding builds). It's a big deal though, and keep in mind that wanding with a shadow wand also hurts the shadow priests DPS as well which, in turn, means less mana back for the shadow priest's group. So do the raid a favor and pick up a non-shadow based wand.

Umm, no. But I say this only because I've been testing it. Data first, then explanations.

70 Night Elf priest, naked except for wand, shooting at Dr Boom (level 68) till the wand broke - he died once during this and I waited till he respawned and continued the process. Wand is "Wand of the Netherwing" (sorry, don't know how to link) which only adds to Stamina. Spell Crit (paperdoll) from Int is 3.05%. Melee/Range crit (from paperdoll) is 5.08%.

During the sequence only three results were obtained: Hit, Critical Hit, and Resist.

Total attacks: 8569
Total Hits: 8121
Total Crits: 278
Total Resists: 170

Miss rate: 1.9839%
Crit rate: 3.2443%

Observations and tentative conclusions:
1) The crit rate is significantly closer to 3.05% than to 5.08%. This tends to support the idea the spell attack table is used instead of the ranged attack table. Supplemental but by themselves very weak evidence is: a) the miss is a "resist"; and b) crit damage modifier is 150%.

2) The slightly higher rate is interesting. It alone would imply level may influence crits. What I'm discovering, however, is that this is not necessarily true over other tests. As in...

Additional: I made a similar test - there are enough differences it cannot be considered conclusive, but it tends to support the above. This time my night elf priest was fully geared and had a 6.59% paperdoll spell crit rate, still a 5.08 ranged crit rate, and used the calming spore reed. Out of 2959 attacks, 189 were critical hits - giving us a 6.3873 crit rate. Note two facts. First, this approximates the new spell crit rate. Second, it is LOWER than the paperdoll crit rate. I find it curious. I have had similar wild divergences using other wands - but the crit rate tends to approximate the paperdoll spell crit rate regardless of where I apply it.

Untested: do spell-damage enhancing talents (such as shadowform and shadow-weave) affect the performance? And do wand hits proc other items and talents?

I forgot 3) - the miss rate is consistent with spell attack table. Specifically, even level attacks for spells are 4% miss, while for melee/ranged they're 5% miss. Historical patterns of both indicate that each level difference (for at least two levels) is 1% per level. The ~2% miss rate is consistent with spell table at two levels of difference.

Wands are spell attack table, not ranged.

I'm curious whether wands are capable of activating other on-hit trinkets. Since wands proc the physical bonus rather than the magical bonus of the Darkmoon Card: Crusade, perhaps wands are able to proc other physical chance on hit trinkets, but not any magical chance on hit trinkets.

JoL health restore will trigger on melee attacks only.

If the damage is not classified as a spell, does that automatically mean the game looks at the damage as physical? That is to say, does this mean that the wand damage is counted as physical when it is dealt, but is counted as a particular school of magic when it hits the target?

When you perform a damaging attack, it's classified as (attack/ability/spell), with spell further split into schools. When the damage lands, it's only classified by schools. A wand attack counts as an attack, not a spell, when launched so it doesn't even get into the school breakdown. It's not that it counts as physical, it's that since it's not a spell such a distinction doesn't exist (it doesn't count as a physical spell like taunt or demo shout). When it lands it counts as the appropriate school of damage.

I think.

I went to Dr. Boom and used the Wand for a while, +spell damage does nothing, but wands do crit with regards to your spell crit rate.