I really don't know. In my own testing it makes no difference whatsoever which side the cuts are on. I have asked others about this and can vouch for at least three other makers that have confirmed this as well as Sal Glesser of Spyderco in his own testing. These locks can and do slide off the interface contact area for either way its done but this has a lot to do with the pitch angle of the contact also. There are theories behind why its best on one side vs another but nothing that can prove one way is better. I note that some makers like Ken Onion and Chris Reeves among others will do it both ways at times. Strider believes it has some way of controlling the energy transferred down the lock under stress. I have not seen this in my testing and it appears to me not to make a difference what side it is on or how many relief cuts there are. I've done it both ways.
Emerson, and many others do it inside and have for years with no issues. I happen to feel it looks better inside tucked out of the way so thats how I do them 99.9% of the time for my own. In my own testing, its the contact area and more specifically the angle of that contact ramp as well as the way the lock contacts the blade, ie amount of lock contact for surface to surface area has more to do with when the lock defeats and how than the relief cut itself does but its much more than this too really. Many people don't realize that longer knives tend to put more pressure on the contact of the lock and blade at the pivot. Oh, they know about leverage but they don't associate it with a folder much I don't think but believe me it affects them. 100 pounds on the tail or lanyard end of a Military is more weight on the pivot than it is on a Strider PT. The amount and way the lock contacts has to do with how well the lock wears and how also but again the length of both the blade and the body play a role in this also. Indenting is very commonly seen in certain knives vs others in the industry and its all tied into these same points mentioned.
One of my biggest pet peves with many makers is how incredibly thin they take these cuts to spring the locks. Don't take this wrong because the length of the lockbar as well as the width in back at the cuts has a lot so say about how much you can leave behind when you make these cuts, but I will say it seems some get a bit carried away and take more than is necessary out of the lock. I've measured some as thin as .032 thickness on .125 or thicker slab sides the lock was made out of and yet the lock would probably have worked and been fine at .055 to .060 or thicker had they stopped there and tried it out. On other makes you can find knives with similar length and thickness lockbars with lock cuts left thicker that work fine. All I've ever been able to figure for this thinness is that they want them to be easier to push to release the lock, or they are concerned with how the lock can affect blade centering or something. I really don't know what the reasoning behind making one this thin is.
Chris Reeve takes his down to between .050 and .060 thickness at both his cuts with a thin spot in the middle perhaps .010 or so thinner still and this is based on my own measurements of his knives, two of which I own myself so that is where I put mine usually. Anywhere between .045 and .060 seems to get by for most without worry but again I don't hear a lot of factual support to back up one way being better over another. Its all conjecture but based on what I've seen the locks do in my shop under stress it appears to make no diff at all.
I have noted also that a number of makers do one, two or sometimes three cut reliefs in their locks, while others use just one thin one and others still use one massively wide one. There also seems to be a wide margin of difference in how thin they are in the same models for productions and customs alike because I've had SnG and PT Striders, Emerson CQC12 models and many more from all manufacturers and makers from Terzuola to Mayo and more here for low rider pocket clips or other such things that allowed me to get to check them out. During these times I've measured on several occasions knives of the same model and got different measurements. Don't get me wrong, its not real extreme but some would be .034 and others .039 and yet others .044 thickness. It seems they just eye it up or something. I really don't know how a lot of them do it either. Some looked milled out, some maybe done with a laser, and others are almost like they were sanded/ground down instead.
Again it all has more to do with their own personal beliefs I think than anything I've found in testing my own and many sent to me for evaluation by manufacturers.
On integral folders I've hung free weights from using a cable through the lanyard with the blade open and locked and clamped tight in a wooden vice this relief area is the weak link that will literally fold in under enough stress and at times even kink in ruining the lock but in a very high % of all knives I've tested the lock will slide of the interface before the stress is allowed to get that great. Keep in mind that a lot of these folders I tested took weights more than likely beyond what the normal 200 pound man would be able to put on one but others failed miserably at 65 pounds on the tail of the knife. Ironically the ones that fail a lot of the time are the bigger higher leverage models that look a lot tougher than some that hold an unbelievable amount of weight. Again there is no rhyme or reason to how or when they fail either and in use with the grip holding the lock it is feesible that a lot more weight could be applied before the lock kinked in. Personally if I was looking at two massive frame locks I'd pass over the one with the .032 thin relief and get the beefcake looking one myself.

The thickness matters to me more than which side its cut out on.
STR