I generally tend to go for either 4 or 5 min fights with light movement, occasionally I'll up the movement bit when comparing talents just to see their viability in (new) movement heavy fights, but for gear I stick to light movement.
Some people might favour simming specific fight setups, but we stopped raiding Mythic because of the logistics, so my fight lengths are already more spread out depending on how many DPS we have.
As for the stats:
Versatility is high because you have absolutely none of it.
It's easiest to explain when you compare it to crit euhm ... for Enhancement. Because they don't have any special mechanics attached to it.
(Warning; the following is napkin math)
Let's take your 22% crit. If you were Enhancement, that would basicly mean that this 22% crit increased your DPS at 0% crit (let's say that it's 400k) by 22%
So you would end up at 400k + 88k = 488k.
But what happens if you take 1% crit extra? You gain 1% of the base 400k, not the extra 88k, because you can't double crit.
If you got 1% Versatility instead, you would gain 1% on that full 488k, for an extra 880 DPS more.
Of course, if you have 1% Versatility, then adding another only works on that 488k, while crit would then work on the 400k + 4k from Vers.
In more mathy terms: DPS = (1 + Crit%) * (1 + Vers%) * (1 + OtherStuff).
The increase in DPS is then the total amplifier of the stat.
CritIncrease = (1 + Vers%) * (1 + OtherStuff)
VersIncrease = (1 + Crit%) * (1 + OtherStuff)
Noticed how the stat itself only amplifies OTHER stats?
This is called diminishing returns, which means the more you stack it, the more valuable other stats become. This is why a lot of people are shifting towards more of a stat balance (e.g. Enhancement runs something like 1 Crit : ~2 Haste : 3-4 Mastery : 0.7-0.8 Versatility)
PS: I chose Enhancement Shaman because they don't have any mechanics that benefit from crit. Elemental crit interaction is more difficult, but the basis is that for every spell, you can write overall spell DPS = (1+crit)*(1+haste)*(1+vers)*(1+mastery) even if that mastery/haste gain is 0%. Then add up each part depending on time spent casting, including the effect of Haste & GCDs on it, then balance that vs a boss model, put it all into code to automate it and ... you are where simcraft is now, after MANY iterations and thousands of manhours combined from all theorycrafters. But atleast now you should be able to understand what simc does better