I've done a bit more fiddling with the Quickjacks, and I've learned three things (and I've been under the car again):
1) the lift points are inboard of the pinch welds. Selym's comment now makes sense to me. I think this was a contributing factor to the incident.
2) there is a lot more weight at the front lifting point than the rear.
3) orientation of the lift blocks matters (sideways vs lengthwise)


This is a photo of the left front lifting point looking aft along the pinch weld. Inboard is on the left of the photo, outboard to the right. The block is centered in the tray of the jack. The puck is centered on the block. The pinch weld is centered on the puck. Therefore, the Quickjack is centered on the pinch weld. Two things are evident: 1) the puck is deformed upwards on the right side of the photo because the body is further up outboard of the weld compared to inboard. It steps up about 3 mm.
2) the inboard side of the lifting block is compressed more than the outboard side. This is probably creating a small lateral force toward the outboard side.


Same shot, but this time the block is centered in the tray of the jack, the puck is centered on the lifting pad on the frame inboard of the pinch weld, and the pinch weld is centered on the puck. The Quickjack has been moved inboard about an inch so it is centered on the lifting pad, not the pinch weld.
Now the compression on the lifting block is much closer to the center. Dead center would be 2 1/8 inches. Here the maximum compression is at 2 3/8. Off by 1/4 inch.


I included this shot of the rear lifting block to show you the difference in compression. There is a lot less weight at the rear lifting point.
(The views of front and rear make the blocks look different. In the front, the block is up against the front of the lifting tray on the jack, so the edge of the tray partially obscures the block. In the rear, the camera angle looks into the tray, so the entire height of the block is visible).

One other thing I have learned: I was orienting the lifts blocks with their long axis in line with the long axis of the Quickjack. It's better to orient them with the long axis of the block perpendicular to the long axis of the jack (sideways, not lengthwise). There is less margin for error when they are oriented lengthwise. With the block oriented sideways, the center is at 2 1/8 inches. If the load is off center by 1/2 inch, there is still 1 5/8 inches to the edge of the block. With the block oriented lengthwise, the center is at 1 1/2 inches. Being off by the same 1/2 inch puts the load 1 inch from the edge.

New summary of factors leading up to the Quickjack being spat out from under my car:
1) Quickjack was centered on the pinch weld rather than the lifting pad just inboard of the weld. This caused the inboard side of the lifting block to be compressed more than the outboard side.
2) the front pad was misaligned (toed out 6 degrees), adding even more compression to the inboard side.
3) the pads were oriented lengthwise, reducing the margin of error for load placement
4) the small lateral (outward) force created by the uneven compression overcame the low friction between the Quickjack and the plastic garage tiles.

 

I have now been under the car again with it up on the Quickjack.

First I tested the newly-installed rubber feet by putting some of the car's weight on the jack and then trying to pull the jack out. It wouldn't budge with me pulling 100 pounds straight out. Then I put all the weight of the car on the jack, but lifted the tires only an inch off the floor. Pulled again, 100+ pounds. Didn't budge. Pushed and pulled on the car at the front quarter. It was hard to get much of a wiggle out of the car.

Then I lifted the car all the way, placed jack stands, and dropped the Quickjack into locking position. Luckily, the height of the front jack stand was perfect to take a bit of weight. Most of the weight was on the locked Quickjack. The rear jack stands were way too short to reach a solid point on the frame. I wasn't satisfied, so I jacked up the rear suspension to offer more security.

I pushed and pulled again, and it felt solid.
But I still wasn't happy . So I bought a pair of 3 ton trailer stabilizer jacks that gave me the 24 inches I needed to reach a support point at the rear. (sorry, no photo)

Now I'm happy .
1) Rubber feet with much more friction.
2) Load properly centered on the Quickjacks so there should be no lateral forces.
3) Four supplementary jack stands taking a bit of weight so they are unlikely to slip away if the Quickjack decides to take off again (which is now unlikely given the first two measures).

BTW, I have contacted Bendpak (Quickjack). We'll see where that goes.
Thanks to everyone for their comments and suggestions.

Stay safe!

 

I ran across your thread while googling quickjack failures. I have a set as well and use them on my similarly heavy SL55.

I am a structural engineer and am trying to work out what went wrong. Your testing is very helpful but I have a few questions that would help. I'll share whatever finding I can come up with on this thread.

In your post showing 6 degree tilt (#19). Can you describe the tilt as it relates to the orientation of the block during the lift? Is the 6 degrees tilt of one face to the other? Was this the block in the front that slid out?

When you started the lift did you check parallelism between the two frames? If so by measuring or eyeball?

Is your garage floor level?

What is the weight of your G37 and any idea what the weight split is front to rear?

Based on some assumptions and what I read in your posts I find that roughly 260lbs of lateral force was generated to cause the front of quickjack to slide out. How that was generated is the question. I suspect either parallelism of the jacks and/or some offset of load from the jack centerline, maybe some combination. Hopefully I can come up with some rough tolerances for these factors based on your experience.

 

With the many cars I've lifted I've noticed that any time I've seen rubber pads these have been the thinnest you can possibly run. This is to prevent the rubber to flex too much.

Because these quick jacks are universal you'll need to take a custom approach for the G. I would toss those blocks for rigid ones and utilize the pucks as pinch weld savers.

Your issue are the blocks. A 4x4 chunk of wood with a slit at the top to accommodate the pinch weld will do just fine while adding rigidity and not killing your pinch welds as well.

 

Maybe chain them together from now on once you get it positioned loop 2 chains across them so they can't slide out on you..that had to be terrifying..so glad your ok..

 

While doing some research to find out why my QJ's slid out from under my car, I came across this site. The most glaring common denominator is that I also have plastic garage tiles like the OP that allow the QJ to slide very easily. I have the SwissTrax ribbed style tiles in my garage. I think their slippery nature with the steel bottom of the QJ allows them to slip out if everything isn't aligned up very carefully.