If feral run speed is 9.1 yards/second, then in principle you should only be out of range for 8/9.1 = 0.88 seconds, plus a very small travel time for charge itself. Here I assume that spell ranges are measured relative to maximum melee range. Anecdotally this seems to be true -- targets with a larger melee range tend to have a larger minimum range for charge -- but I have not measured this explicitly for differently-sized targets. However, some quick stopwatch measurements with a target dummy appear to corroborate this assumption for both feral charge minimum range (8 yards) and 40 yard faerie fire. In particular, the time I measure between maximum melee range and minimum charge range is very close to the 0.88 seconds figure (again, only for the target dummy).

Realistically you won't be running directly out as you will want to maintain facing with the target in order to avoid delays due to reaction time. You may also need to add up to twice your latency -- once for the server to tell the client that you are outside minimum range and again for the client to tell the server to cast charge (I'm not sure how the WoW netcode works; this may not be necessary). Let's be generous and say you're running out at a 30 degree angle to the radius vector; this will scale run time by about 1.15, bringing the figure up to 1.02 seconds. Let's also say you have a latency of 150 ms and the netcode is such that you do need to add twice the latency. This brings the total time to 1.32 seconds. The travel time for charge is difficult to measure because it is so short (remember this is the time for charge to return you within maximum melee range, which can be much less than the total leap time). My stopwatch measurements impose an upper bound for this time at about 0.35 seconds; let's assume this is the travel time. Then the total time is 1.67 seconds.

Suppose you have 10% haste from gear and assume hit and expertise caps for simplicity. Then swings are hasted by 21% and occur every 0.83 seconds. You don't lose the full 1.67 seconds worth of swings because the swing cooldown timer continues to tick once you've stopped attacking; this means you save whatever fraction of the swing cooldown remained. On average you will run out when the swing cooldown is at half its maximum duration, so you save 0.5 swings. In total, then, you lose 1.67/0.83-0.5 = 1.51 swings. Each swing is worth swing damage, 0.0583 clearcasts, and an undetermined number of fury swipes.

Fury swipes is a 15% chance with a 3 second cooldown. With a 0.83 second swing timer there are 3 swings ineligible to trigger it after each proc. After that we expect it to trigger after 6.67 swings on average. A given swing then has a 6.67/(3+6.67) = 69% chance to be eligible to trigger fury swipes. Furthermore, if you lose a fury swipe you don't simply lose its damage, because by losing that swipe you raise the chance for subsequent swings to trigger it. Specifically, there is a 38.6% chance that a fury swipe will trigger within the next 3 swings, which would have otherwise been ineligible. Of course, we need to apply the same reasoning to these swipes: if one of them triggers then you lose 0.386 fury swipes from the next 3 swings, and so on. By this reasoning we can say that by losing a fury swipe we lose a fraction of its damage equal to 1-.386*(1-.386*(1-.386*(...))) = 1-.386+.386^2-.386^3+... = 1/(1+.386) = 0.72. Therefore, for the first lost swing we lose .69 * .15 * .72 = 0.075 fury swipe damage. Since the first lost swing cannot trigger fury swipes, the second lost swing has a larger chance to have been eligible to trigger fury swipes: 6.67/(2+6.67) = 77%. So for the second lost swing we lose .77 * .15 * .72 = 0.083 fury swipe damage. There are only 1.51 lost swings on average, so for the 0.51 lost swings after the first I will simply approximate its value by interpolation. Thus we lose 0.075 + 0.51*0.079 = 0.115 fury swipe damage.

For simplicity let us ignore combo points. The ravage option produces more combo points so this simplification only weakens it. Let us also assume that energy and clearcasts that are lost by taking the ravage option would have been spent on shred. Hence the total gain for ravaging is

(1+set) ravage - 1.51 (swing + 0.0583 shred) - 0.115 fury swipe - 10/40 shred

where set = 0 without 4t13 or 1 with 4t13, and ravage, swing, shred, and fury swipe are the average damage values for those attacks.

According to Rawr, in my gear
ravage = 34996 (below 80%)
swing = 5234
shred = 31752
fury swipe = 9444

in which case the total gain is 15272 damage without 4t13 or 50268 damage with 4t13.

I like that change from a "oh hey, I'm not double fucked for having to attack from the front" and at the same time, I really dislike the idea that a mangle spamming scrub will be better at cat dps. I feel like they could have made a better change to "un nerf" feral dps in positioning limited situations.

Maybe remove the positioning requirement on shred when you have both bleeds and mangle up on the target? I know there is a better way to solve the positioning issue. This glyph change just feels heavy handed.

Civilian, that's correct math - but at the same time it's a bit misleading as the chances of being able to do this kind of movement and tracking such that you only lose 1.5 seconds of swing time on the target (while being maximally distant right before doing so) is very small. In real applications chances are you're going to lose a lot more, especially if having to start dealing with mechanics that trigger when you're not in melee range such as some of the ranged-only attacks.

If it's a 1% gain it's essentially well within the noise of having optimal rotations, using procs perfectly and having good crit distribution for CP. Sure, do it when you can, but it's very easy to not get that production on most non-trivial fights. In any case, it's certainly not a slam-dunk boost to DPS and in fights where stampede isn't possible the net value of the 4p bonus in the end expansion is 0%. Imagine if this were the bonus that was part of 4pT12 - where the only time you can use it is when you're not a stunner and you're killing adds, because using it on Ragnaros is literally suicide? The Deathwing encounter has four floating platforms where you're beating on tentacles; I doubt you're going to be able to charge on them, either.

0.6 swings or so lost, and out of range for at most 1.3 seconds. That was basically the best possible scenario in that a swing hit just before I left melee range; if I were not watching the swing timer that would be 1.1 swings lost on average. Still substantially better than my estimate above, though of course this is just on a target dummy.

The timing between the feral charge cast and the next swing seems to be varying though, so my assumption may have been wrong that this time should just be the feral charge travel time.