Whackjob

Yeah.  Even the rocket engines on the outside columns.   ONE strut holding them perfectly in place?  No chance.  One, all that thrust on a weak connection, and no shearing?  Doubtful.  Two, no oscillation at all?  Even if he turned off thrust vectoring on those motors, the slightest tiny oscillation by any motor would cause all the other ones to oscillate, and that whole stack would start whipping around like a noodle.  I know, I spent weeks figuring out how to tame that particular beast.  Too rigid and you get shearing.  Too weak and you get angry whipping noodle.  You need a happy medium, and that's the truss cage system.  It's the best solution I ever found, and it's a whole lot of reinforcement.  This has no reinforcement in comparison.  A picture:

THAT is the epitome, the strongest design of a truss cage I've ever come up with.  It's hard to see, but between tanks, it isn't tank-strut-tank, it's tank-truss-strut-truss-tank, and further reinforced by using equilateral triangles.  You can see it right there.  Also note that is using the smaller radii parts, before the real large ones came out.  That's the reinforcement it took for something that weighed a lot but had smaller parts.  Anything short of that would have caused angrynoodle or shearing.

I'm comfortable in my assertions, given I've spent a seriously long time building really big.  That being said, I was recently beaten soundly by Scott Manley, and I can't even be mad about that.  He's legitimate.

Agreed. I noticed the engines vectoring during the launch. They should have torn it apart.

Here's a question I've never seen satisfactorily answered: We know that node size is an indicator of connection strength, but what other factors come into play? There is some proof that connecting to something massive is good, but I've also noticed that impact tolerance seems to have an effect. I learned a long time ago that connecting wheels to girders and structural panels vastly increases the joint strength.

Can you confirm this? Obviously, you're using girders above--It's got to be more than just an aesthetic choice.

I can, in a way.  From my experience, struts increase rigidity, but not flexibility.  You can put fifteen struts between two tanks, and they'll move a huge amount less in relation to each other.  However, struts are also prone to stress failure and shearing.  A real good twist or knock or torque will unzipper the whole thing.  This is when high mass is involved, mind you.  Now, trusses, they have some rigidity, but can also flex a great deal.  By combining the two, I got the best results out of any combination.

Look again at my example picture above.  See how I formed equilateral triangles?  From the center stack, I put two medium trusses out, and put the next column.  The two trusses is strategic.  For the outside-to-outside, I put on truss on each side.  Mirrored as it is, it puts two trusses that sort of touch in the middle.  BUT, they are not connected.  Using some fancy camera work, I can put a strut in there that connects one to the other.  That link, plus the equilateral triangles, is what gives that shape its strength.

Oh, I notice you call them girders.  I call them trusses, just for reference.  As far as node size reflecting strength, I believe that only comes into play with the cylandric tanks, and even then, only at the snap-on joints at the ends.  Those are considerably stronger than what they were in the past, but seem to be only in forces that pull them apart.  Shearing forces still shear about as well, and compression is still absolutely an issue.  I have an engine cluster I play around with quite a bit.  I can't remember the motor name, but it's one of the newer ones, the large scale, it's a tank with two nozzles on the end, and the radius of the tank is the same as the rockomax 64s tanks.  The engine cluster I use has I think 24 of them clipped inside of a tank (legit parts clipping, no mods / cheats) that ends up with a cluster with 144,000 thrust.  That under the usual tank stack can create compression enough to cause detonation.

Better yet, I found a picture:

To explain what you're seeing here, the motors are inside the tank, but have that medium truss attacked to the motor, but protruding from the tank.  The fuel line runs the fuel from the tank stack to the truss, which inexplicably allows for fuel transfer in to the motor it's actually attached to.  Here is three of them in flight:

That's just fun.