From an aircraft operations standpoint, as a general rule whichever group will have the slowest climb out should leave first. On a calm day the aircraft on jump run covers about 175 feet per second, or one mile in 30 seconds. Assuming the practical distance that a square canopy open at 2,000 feet can cover is at most about one mile, that means that the first people out would just barely make the landing area from one mile short, while the last would be able to make it back from one mile long. In other words, all jumpers have to be out in a two mile long jump run or some will land out, or a second pass will be required. In time terms, on a calm day no more than 60 seconds can elapse from when the first jumpers leave the airplane to when the last jumper exits. At busy events with several aircraft flying, second passes are not an option.
Let's take a sample jump run, where a large group will take up to 20 seconds to climb out, a 4-way 12 to 15, 2-ways six to eight, solos five, and AFF students about 12 to 15. Our load has an 8-way, two 4-ways, two 2-ways, a solo, and one AFF. That adds up to between 70 and 80 seconds from green light to last out. But it is possible to make it all on one jump run if the eight way gets out first, because the pilot figures at least 15 to 20 seconds for the first climb out. That brings us back to 60 seconds from first out to last out, and one pass. Needless to say, we don't want to do an extra pass because 2-ways want to leave before 8-ways.
(If the 2-ways get out first, the pilot can only count on a five to ten second climb out. He has to put the light on 1/3 to 1/2 mile closer to the dz than he would for the slow climb out.)
That's the timing reason why small groups shouldn't leave first. Now lets talk about separation from other jumpers. First of all, anyone who counts on vertical separation for safety is out of touch with reality. I see people in freefall at 1,500 feet and lower routinely, so just because someone plans to open at 2,500 doesn't mean you should bet your life on it. Everyone needs to open in their own column of air. Horizontal separation is the only guarantee of security. The only real reasons - and they are good ones - why students and tandems get out last are that a student is more likely to balk or ride down, and that canopies opening high can get back from a longer spot. I repeat, horizontal separation is the only guarantee of safety. Vertical separation is a nice idea but cannot be counted on since a minor loss of awareness or a long snivel will eliminate it instantly.
Now, a quick digression about fall rates. Follow these categories out or time their videos if you don't believe me. Light freestylists doing routine freestyle do not fall significantly faster than a fast falling four way. Freeflyers fall about 30% faster than normal. Small skyboards fall fairly fast, if the rider is standing, but big ones fall very slow - slower than most RW, usually about the same speed as tandems. Because of their exits, they must leave first, and because of their complex emergency procedures, they must pull high. Leaving first and pulling high defies conventional wisdom, yet not once have we had a problem with slow falling skysurfers getting out first and pulling at 3,500. In fact, as long as the first person pulls higher than the break off altitude of the following group, they are a contribution to safety, not a detriment, provided adequate time was left between groups at the exit.
We do have a recurring problem maintaining safe separation when the freeflyers get out first. Typically a freefly pair will have a forty five second freefall and open at 2,500 to 3,000 feet. Let's imagine that they are followed by an RW group that has a 10 second climb out. Now, let's say you are a freeflyer jumping a Stiletto. A Stiletto (assuming a 30 mph forward speed, which I can document is a reliable figure) covers about 45 feet per second on a calm day. If you open 30 seconds (shorter freefall plus exit separation time) before the RW group leaving after you and turn directly towards the dz (which you will, since otherwise you can't make it back from getting out first unless you cheat on the climb out, spot for yourself, and force the pilot to go around, which REALLY pisses us off) in that 30 seconds you will cover over 1,300 horizontal feet. This would put you about 400 feet from the center of a group leaving the plane ten seconds after you. In theory, that would just barely be enough, except that a good tracker can do about 70 feet per second, so if they track towards you for six seconds they are right on top of you. Furthermore, a modern canopy descends about 800 feet in 30 seconds (also documented) so if one of you pulls at 3,000 to get back from a short spot, for camera effect, or whatever - by the time you are at 2,000 you are well into the danger zone of the group that followed you.
So far, the big sky theory has taken care of us most of the time but I have heard of a couple close calls and more than once found myself directly over the freeflyers if they leave first. Having seen the consequences of freefaller/canopy collision more than once, I want to minimize the possibilities. And they go way up as soon as we add wind to the exercise. Here's why. In a 30 mile per hour breeze, the plane only covers 130 feet per second, instead of 175. In ten seconds of exit separation, the airplane only covers 1,300 horizontal feet instead of 1,750. Worse still, the RW group is in freefall for a longer time, and consequently gets blown further. Let's say the freeflyer is in freefall for 45 seconds, and the RW for 70. In 45 seconds you get blown nearly 2,000 horizontal feet. The RW blows just over 3,000. That leaves only 300 feet of horizontal separation without taking tracking or canopy movement into account! Make the winds 50 miles per hour, and the RW group drifts over 1,800 horizontal feet further than the freeflyers! Meanwhile, in ten seconds the plane only covers 1,100 feet. A 20 second exit separation will still have the RW group opening 400 feet from the freeflyers, not counting canopy movement or tracking!
Having opened right over freeflyers before, and having just heard from several expert skydivers who narrowly missed freeflyers, and having watched RW groups blow over freeflyers on windy days, I think we have a problem. You might say, make sure the groups leave longer between exits. Well, we do tell them, but if they wait 20 seconds instead of ten, that still doesn't solve the problem because Freeflyers still fly under them under canopy. So for fast fallers your only choice if you want to get out first is to always fly perpendicular to the line of flight for 30 seconds before turning towards the dz. While I am confident most of you are aware enough to do this, it brings us back to the original time on jump run problem. Basically, Skydive Arizona isn't willing to do a lot of second passes just so freeflyers can get out first.
Getting out last except for students solves virtually every problem. You control the horizontal separation, so you can ensure you won't be overtaking anyone in freefall. The windier it gets, the safer you are because you get extra separation by having slower fallers blow away from you. Students take long climb outs and pull real high, so no problem there: just get open and fly off the wind line for a few seconds to be clear of them in the unlikely event that they are in freefall at 2,500 feet.
As for the argument that the canopy separation is necessary in the landing area, I don't buy it. Opening over the top instead of short, you can spiral down to make sure you get on the turn around loads. As for congestion at the landing area, no one else on the loads seem to have any problem, although you may not always get to land right by the fence.
Please give this some thought. Unless one of you gives me an extremely convincing reason why you need to leave first, such as a safe spot for the skyball, I will make it standard policy that exit order will always be
2) freefall groups, largest to smallest, regardless of fall rate
(Note (Skr): I believe this is a typo since the real rule is: )
(2a - relative work groups, largest to smallest and then )
(2b - fast fall groups, largest to smallest and then )
(3 - AFF and tandems )
3) AFF and tandems, plus any other very high openings. The main reason for high openings leaving last is not separation, it's that they can make it back from a long spot!