When a given torque is applied to any fastener (or any screw of any type), a percentage of that torque is converted into linear force (the 'tightening' force) and a part of that torque is wasted overcoming friction between the mating threads (and converted to heat but we really do not care about that). As the thread friction is reduced the linear force goes up. When we tighten a fastener what we really want is a given force (or stretch) applied to the fastener head, or the sealing ring of a spark plug in this case. So when the threads are coated with anti- seize the friction goes down and the force applied goes up for a given torque. So to maintain the same force on the sealing ring of a spark plug after applying anti- seize (or any friction reducing material) we must reduce the applied torque.
To expand on that a bit, we only use torque to measure the value of 'tightness' on a fastener because we can directly and easily measure the torque applied. What we really want to measure is the force applied to the fastener but that is difficult to measure. But measuring torque is a 'best guess' situation anyway because anything that changes the thread friction alters the torque / force relationship. It works best for things like head bolts because the threads are clean and a lubricant is usually specified (motor oil in most cases). It works really pretty poorly for things like lug nuts because of the dirt, rust, deformation, flaking plating, etc., etc. that is really all over the place.
The way that works extremely well is to measure the stretch actually applied to a fastener. The stress / strain relationship (stretch to force applied) of steel is very well understood and documented. Chevrolet small block connecting rod bolts are best tightened this way; a gauge is put over the bolt before tightening and set to zero: when the bolt stretch reads 0.007" or 0.009" (I do not remember which), the bolt is <correctly> tightened regardless of the torque applied to the nut on the far side of the fastener. Obviously something like this is impossible to do with a spark plug though. As an aside, come critical fasteners are made with a built- in strain gauge which is used to <correctly> tighten them.
As a bad 'rule' of thumb, machined steel threads are considered to be 50% efficient (50% of the applied torque is used to make force, 50% is wasted as thread friction). A very good extreme pressure lubricant may raise that number to 80% or 85%. So here is what I do with the spark plugs in my C-14: I use anti- seize which is a fair extreme pressure lubricant, I do not bother looking up the recommended torque for the spark plug, and then I install them using a 14mm box end wrench on the end of the Kawasaki supplied spark plug tool. I tighten them too exactly "reasonably snug", being very careful not to over tighten them. Yeah, I know that is funny but it is the truth. Bolt tightening torque is not a very exact thing under good circumstances anyway, and anything w/in 40% of correct is more than close enough in most cases. Case in point: there is a very specific torque specified for the rear axle bolt of a C-14. So we tighten that fastener to exactly the correct torque…. and then tighten or loosen it slightly so that the cotter pin holes align.
I bet that is way more than you wanted to know about tightening spark plugs.
Brian
Clarification please. My intuition tells me reduced thread friction would call for more torque. Is My intuition wrong again 
