Question for the marathon experts regarding fast finish long run

I don't think your HR matters.

Fast finish long runs are a means of training your body to run strong late in the race, not to simulate running conditions for the entire race. So, while your HR for the last part of the run was higher than you plan on for most of your marathon, it's probably pretty predictive of what it will be late in the race if you push the last 10K.

You would ideally run the race about 3 - 5 beats below your LT for most of the run until the later miles as you begin to get HR drift and then it will go up (assuming you are holding pace and not slowing down to account for the drift).

A better measure to me for pacing is not whether or not you can hold pace at the end of a longish run would be if you could do a 17/12 or 18/14 pace run (first number total length of the run, second number is the continuous marathon paced run during that long run).  On moderately fresh legs it is not all that hard to run your M pace for 6 miles at the tale end of a long run, especially if you have done the first 12 of the run at McMillan's long run pace for your ability level.  But in my experience it is not a real good indicator of whether you will be able to hold that pace for 26.2.

If you are borderline of hitting your goal, I think you would be better served to try and do one of the runs above, probably the 17/12 would be better.  If you can't hold 7:14 for those 12 miles on a training run, you are more than likely not going to be able to hold it after 20 of the race.

Is it a good sign you were able to pick it up at the end of your run?  yes.  Your HR being a little high something to be concerned about?  I would again say yes.  If your HR is too high during the race you very well blow up if you try to maintain it for too long.  I agree that if you want to do that pace during the marathon your better off trying it out in practice. 

I suspect this is your first marathon and for that reason I give the following advice.

If you want to be able to run the marathon fast and BQ I would suggest running at a 3:10 finishing pace for the race.      You might finish 5 minutes slower than your potential but most people I know tend to try to finish faster than they can handle.  Going out trying to beat that 5 minutes might cost you 15 in the end.  Plus if you finish in 3:10 you can race in Boston which is really cool too.  

In the context of the OP, I disagree. He was running hard at the end of an 18 miler. So of course his HR during that portion of the run is going to be higher than what he will experience during, as he put it, the majority of the marathon.

His pace for that portion was 10 seconds/mile faster than his goal pace, so he could expect his HR to be higher just because of that. But even discounting that faster pace, were he to run 7:06 for that entire 18, he would expect his HR to be significantly lower for the majority of his run, simply because early in the run he would have been fresher. As Rick pointed out, cardiac drift would account for a gradual creep up throughout his run.

It depends how hard you ran the prior 12 miles. It's definitely a good sign, no doubt, but I agree with the post above that a better indicator would be 18 miles with 14 at MP pace; such a run is the key run of the Pfitzinger Advanced Marathoning plan, and is likely substantially more difficult than your 6 mile fast-finish if you crank it up to a 3:06 marathon pace. 

Still, that's a great run, so keep up the good work. I still haven't hit 3:10 as a M33, and I ran 70mpw for 3 cycles with HM times of 1:25, and 10K times of 39:00 (several). The 26.2 distance is really tough to hold a fast pace, and is substantially more difficult than holding for 18 or 22. 

What Rick said.

I'll add a slight variation -- I like to do something like 18/15 or 19/15, two weeks before.  In fact, I did that yesterday.  (2 miles warm-up, 15 miles at around marathon pace, 1-2 miles cool-down, depending only on how long it takes to feel 'cooled down'.)  Holding pace for 15 miles gives me a lot of confidence going in that I'll maintain pace for the first 20 (then somehow hold on for the final 10K), and two weeks out is enough (for me) to recover from that run, assuming it is the final very high quality workout prior to tapering.

Anyway, this run has been a fairly decent (but certainly not perfect!) predictor for me in the past.

That is a good sign, however I would make certain that I got 3 runs in at the 18ish distance. I think there is no better method of predicting race performance than running a longer race about three weeks before (could be as few as two weeks before).  Preferably that race is a 15K or half marathon, less preferably  a 10K.  I usually run that race with tired legs (no taper) so that I know that it is going to give me a conservative estimate when I pop it into the race calculator.  I stick to that pace through maybe 15 or so miles and judge from there if I think I can tick it up or not.  For me this has worked every time and I have always come within 2 minutes of my predicted time and almost always faster than predicted.  Best wishes.