Luke, the letter positions are the major thing but there are others that all add up to a positive plate identification.
The main points to plating a stamp are:
1. As has been mentioned, the precise positions of the checkletters within their squares - including whether they are sitting upright or have a tilt.
2. The shape of those letters. The major point here is whether the letters are from Alpahbet 1 or the later Alphabet 2 (these being the only Alphabets we are concerned with if the stamp is imperf). Then beyond that, there are differences within the alphabets - some Ps and some Bs have smaller loops for instance. Some letters have longer serifs, and some letters with 'fully enclosed' areas - such as both P and B - have ink in those areas (blind letters). There are some very distinctive letters from some plates - like the squat S of plate 29 and the squat T of plate 11 - and from some groups of plates - eg the tall F from plates 11 and 30 to 33.
3. The four sides. The left, right, top and bottom borders of the printed design usually can help a lot as any weakness, strength and/or duplication, recutting and recutting that has gone too far and resulted in a frameline that extends into the margin, will all be a constant for stamps printed from that plate.
4. If your stamp is lucky enough to show some of any adjacent stamp - even the tiniest little bit - then this can be invaluable help. Ideally in the plate production, the impressions should have been laid exactly the same distance apart and exactly in line. Yet this just was not the case in the vast majority of plates (indeed there was no real need for this to be so until perforating came along). Impressions can be staggeringly misplaced in relation to each other - too far or too close, too high or too low, too far left and too far right; even not parallel.
5. Any constant mark, such as marks from re-entry/fresh-entry, scratches, dots....the list of possibilities is endless, will quite quickly prove a plating. These marks are sometimes in the margins too.
So let's have look at your stamp again, shown here with the imprimaturs from plates 63 and 68:
Both 63 and 68 have the letters placed very similarly indeed. Despite this, to the trained eye, the letter positions do have slight differences, but we don't actually need to go into that because we can use the other points above to prove your stamp is one and not the other. So using the same numbers as above:
1. Letter positions are extremely close. If anything, one of the Bs has a very slight tilt to the left.
2. Letter shape - both loops of one of the Bs are blind. Both Ps have fairly large loops, though one P looks to be struck into the plate not as heavily as the other.
There's really not much difference between them at this point. The big thing is the blind B, but one needs to be careful about using this as the sole determinant because in some cases this becomes less apparent with plate use - blind letters can become 'open' from plate wear. So:
3. Both stamps have a weak and ill-defined left side. Both stamps have a fairly weak, but slightly stronger right side compared to their left. Both stamps have what I would call normal top lines, and both have basal shifts at the bottom - uneveness of the border directly below the ONE PENNY tablet (note that this doesn't extend beyond this to underneath the letter squares). The strength of this basal shift however, is slightly, yet quite visibly, stronger on one of the stamps.
By now we should have a leaning towards one of the plates. One of them has a B with open loops, and also a baseline that matches our stamp more closely than the other.
4. This is the final thing though that absolutely proves your stamp is plate 68 not 63 - alignment. Your stamp is lucky in having tiny pieces of two other stamps. From this we can see that the stamp above is placed slightly further right than it should be, and the stamp at right is placed slightly lower. This matches the imprimatur that we have 'shortlisted' based on the B and the baseline. The other imprimatur has the stamp above pretty much correctly placed, and the stamp at right is also VERY slightly lower than it should be, but not enough to be a match.
So we have a match with plate 68 (the second of the imprimaturs shown above). But we have icing on the cake:
5. The imprimatur for plate 68 shows a definite lateral margin blur, that is strongest toward the top. Just like your stamp, which is quite a worn print but still shows a good trace of it.
Luke, there is lots written about plating GB penny reds, but if there's one thing I've learned it's that the only way you can ever hope to get good at it is simply by practice, practice, practice. The other thing is that you MUST compare your results with others. I've known collectors who have sat at home for forty years and plated thousands of stamps, then promptly filed them away. How can you be sure they're correct, if you haven't bounced your results off another specialist - at least in the early plating attempts (say the first 500 stamps).