there is no way of knowing what the alternative outcomes would have been if he deferred all this times.
What we do know is that when you win toss and defer, it opens up the possibility (though unlikely) that the other team makes a mistake and elects to kick off giving you ball first in both halfs.
While this is rare and unlikely, when you take the ball first that rare possiblity become 0% chance. In a game that can literally come down to one play or a few inches any given day, choosing the option that increases your odds by any amount is the smart way to go more often than not.
This is one tangible reason you should "almost" always defer when you win the toss, there are other reasons that are less tangible but when you know the game and look at likely situations, they make good sense.
Typically the first few minutes is when players are most amped up... this excess excitement can be detrimental to offensive execution, but can be a benefit to defensive intensity (not all will agree, but my belief is execution is most important for offensive success over intensity while intensity is most important for defensive success over execution...not saying execution or intensity isn't good for both sides, just which is more critical) using this early part of the game to let your D go out and get a hot start while your O calms down a little and potentially gets off to an early edge in the field position battle is the wiser way to go.
Also, while rare, deferring to the second half opens door for the possibility of getting the late first half score and then getting ball right back and scoring again creating a potential 14 point swing in short amount of time. Again while this is rare, it's a possibility if you defer while it's not if you don't, and as a HC you should be making decisions that slant all the little things in your favor because it's an unknown how close any game will be.
Also, just getting the ball first in the second half is a very powerful position to be in from an overall game management position, as 50% of regulation is over and you have a much better view of what kind of game is unfolding than you do before opening kickoff. ie, are you down big are you up big, do want to run tempo and get some quick scores or is it in your best interest to eat some clock up. These are all things you're in better position to make an informed strategic decision on at the start of second half, vs using previous game analytics to guess on before the ball has even been kicked off.
Now I don't think you should pigeon hole yourself and ever be fully predictable, in other words I'm not saying you should "always" defer... I'm saying there are many more reasons to "almost" always defer.
I think only looking at stats since '19 (which covers period when Kelly had his most consistent success and best performing QB, probably slants the stats one way while I think if you put all years under Kelly into the mix it would even it out or even slant it back the other direction, but there are very few valid counter points I've heard to the things I mentioned above.
That's my take on it..