I think one thing I like to keep in mind is that I'm using exact numbers on an inexact and inaccurate system in practice. I open a pack of Egglands Best Eggs. One egg = 60 calories. But I know that not every egg is equal in size. Some are bigger than others. Some days I choose 4 bigger eggs from the package and some days 4 smaller eggs. Some packages are all small eggs and some are all large eggs. The 60 calories is based on a 50g egg, but I don't weigh the eggs prior to consumption. But when I account for eating 4 eggs, I say it's 240 calories no matter whether I ate the big ones or the small ones. It's the same throughout all my meals. I say breakfast is 800 calories every day. But I don't actually know that to be true. Only rather I've placed a number on it that seems to work out based on what happens in response by my body when I eat it.
Same goes for the burned calories from working out. Did my 80DO workout on 8/13/19 really burn 233 calories? Is Autumn correct that I have "after burn" effects from my 80DO workout such that the data my Garmin spits out based on my HR during the exercise isn't really all that useful? And even if there is/isn't "after burn", is it really appropriate in the first place to use HR based data on calorie burning from strength exercises?
All throughout the system, there are inaccuracies. But for me, I take note of what I eat, I work out, and then I see over the course of several weeks whether I feel healthier or faster. If not, I adjust.
Just taking a quick look at my calories burned based on Garmin HR data:
5.15 mi, 45 min, 135 HR, 604 calories: 117.3 cal/mi, 13.4 cal/min
15.12 mi, 120 min, 141 HR, 1496 calories: 98.9 cal/mi, 12.5 cal/min
3.93 mi, 35 min, 133 HR, 459 calories: 116.8 cal/mi, 13.1 cal/min
3.54 mi, 30 min, 139 HR, 421 calories: 118.9 cal/mi, 14.0 cal/min
6.26 mi, 55 min, 139 HR, 733 calories: 117.1 cal/mi, 13.3 cal/min
6.07 mi, 44 min, 149 HR, 628 calories: 103.5 cal/mi, 14.3 cal/min
3.06 mi, 20 min, 161 HR, 337 calories: 110.1 cal/mi, 16.9 cal/min
So I'm getting a range of 99-119 cal/mi and 12.5-16.9 cal/min. And I'm not seeing a trend for HR in either direction really. My marathons were 2276, 2440, 2568, 2055, 2900, and 2720 calories when my HR was active.
The calculator that I wrote for in-race calorie consumption uses the formula:
Total kcal burned = Body Weight (in kg) * Distance of race (in km)
**(Humphrey 2013) This is a general value and a specific value would require physiological testing.
Humphrey, L. [Hanson's Coaching Services]. (2013, Nov 23). Calculating Caloric Needs for Marathon- Updated 11/22/2013 . [Video File].
The example I used (2400 calories burned) was a number out of the air. Last week I burned (per my Garmin and TrainerRoad).
8/12/19 - M - OFF = 0
8/13/19 - T - 80DO-AAA [233 cal] + Spruce Knob (90 min; 94 TSS) [957 cal] = 1190 cal
8/14/19 - W - 55 min Easy (+Strides) Run [733 cal] = 733 cal
8/15/19 - R - Gibralter (105 min; 112 TSS) [1186 cal] + 30 min Brick Run at Easy [421 cal] = 1607 cal
8/16/19 - F - 80DO-Total Body Core [299 cal] + Mist (90 min; 108 TSS) [919 cal] = 1218 cal
8/17/19 - Sa - Longfellow (240 min; 180 TSS) [2300 cal] + 35 min Evening Run at Easy [459 cal] = 2759 cal
8/18/19 - Su - 120 min Long Run w/ FF [1496 cal] + 80DO-Cardio Flow [236 cal] = 1732 cal
Average workout cal burn per day = 1320 cal; per workout day = 1540 cal
I think it very likely goes back to the inaccuracy of the data. Am I really at a deficit of 50 calories per day? Or is it merely that all of my inaccuracies in the system I've set up has led to a 1750 calorie deficit in a week? In the span of two weeks, just about a month ago, I lost 3 pounds. So if I were eating at 1700 calories, were my calculations actually off by 500 calories per day? If the 3500 cal/lb is true, then that must be the case. (3500*3 = 10500/21 = 500 cal/day deficit)
I also know on a daily basis and day of the week basis there is definite ebb and flow. I weigh myself the morning prior to a long hard workout and I weigh X. I weigh myself the next morning after the hard workout, but before the next one and I'll probably weigh X+3. Did I really gain 3 pounds? Because I know if I give it 2-3 days, the weight will go back down to X. It's more of my body's in the moment response to that last workout. But more important to me is whether I feel healthy and fast.
You've put out numbers about a 50-cal deficit or 100-cal deficit per day without mentioning any sort of uncertainty range here. I feel like it's making this all look more precise than it really is. It seems like 50 or even 100 calories in a day is well within the uncertainty of being able to quantify what you ate or the calories you expended per day. 
I ran (run-walk intervals) what felt like an easy 4 miles the other day: felt good, could easily have kept up that slow pace for lots more miles; yeah I was warm, but the Feels Like was "only" 90-something, rather than our usual 100-something, and it felt good! But Garmin says I spent the bulk of that run in HR Zone 4. Huh. 
It sure didn't feel like a Zone 4 effort. Maybe my working HR is just higher than the average person? Maybe I'm sooooo acclimated to existing in this heat, I'm comfortable at higher HRs? IDK. It'll be interesting to try a test on the treadmill and see if I can get a real-life idea of where my max HR actually comes in. And it'll really be interesting to see where my HR falls when (IF?!?!?!?!) it finally cools down!
