Hi, Sherry! You have a few questions in there, so I'm going take a bit to do what causes a lot of G2Gers to groan: I'm gonna provide an overview...and it's likely gonna take me a couple of words. Or maybe just a tad more than a couple. <coughcough>
The mitochondrion is tiny. I mean, really tiny. As in all of our cells--with the notable exception of red blood cells--have from hundreds to thousands of mitochondria. See, it isn't part of the DNA that's in the nucleus of our cells where the rest of our DNA lives. Mitochondria are organelles that are inside our cells, but between the nucleus and the cell wall, and they provide a critically important energy conversion/production function for us.
That's why we all have them, although they're passed along to the next generation only by the mother, in the ovum. And that's one of the several things that makes mtDNA tricky for genealogy: We never test the actual mtDNA that's inherited (the germline DNA); we can only test what we find in the cells of the body, the somatic cells. That's because the ova form before the female is born, while she's still a fetus.
At least 61% of us have more than one mitogenome--a specific mtDNA sequence--inside us (Ramos, et al. PLoS ONE 8, no. 10). Which stands to reason because the average adult will have over a quadrillion happy little mitochondria working away at any given time. And every time a body cell replicates, and dies, so do all the mitochondria associated with each of those individual cells. If you live to be 90 you'll have had...well, I can't count that high. We'd need Barry Smith to estimate that. And the fact is, no copy machine is that good. Having more than one mitogenome within a single cell is called a heteroplasmy, and it's the only way that mtDNA mutates and changes.
As a result, you'll confusingly see some research papers that say mtDNA mutates very frequently, and others that say it's glacially slow. Both are true. Because we have so many mitochondria in us at all times, that heteroplasmic mutation can be very, very fast. Witness this blog post by Blaine Bettinger where he describes how he and his mother do not show as mtDNA matches at FTDNA.
So this is one mtDNA problem. It's entirely possible that the same person, testing at two different times in their lives, could be a full-sequence mtDNA mismatch to themselves. FTDNA tries to correct for this by looking at a given location a bunch of times, and then evaluating the different reads: the heteroplasmy. If there are two different letters read, one will be the "minor allele" and the other, the one that shows up most often, the "major allele." The major allele will be called, but if the presence of the minor allele is below a concentration of 20%, it's too difficult for our current DTC full-sequencing methods to detect.
Even mtDNA full-sequence testing results have to be considered with a grain of salt for this reason. You might test as a full-sequence match to a person in Siberia who is, in fact, no recent relation at all, and you might not match your own mother's results.
The follow-on problem here is that it can take many generations for a heteroplasmic mutation to become homoplasmic. Yep; really a word. Essentially, it means that over the course of generations, what began as a somatic (body cell) mutation has "stuck": it's now become the norm in the body, not the exception, and thereby becomes the "germline" DNA that dominates in the ova and gets passed along to future generations. This is the glacially slow part of mtDNA change.
In a November 2018 paper by Mikkel Andersen and David Balding it is succinctly put (though with great understatement) in the introduction: "...[While] a match of two mitogenomes without recent shared ancestry is in effect impossible, there can be large sets of individuals sharing the same mitogenome due to matrilineal relatedness that is distant compared with known relatives..." How distant are they talking? "The number of meioses [call it birth events, or generations] separating individuals with matching mitogenomes ranges up to a few hundred, and is almost never larger than 500." That's fairly consistent with Toomas Kivisild's findings that the average per-generation mutation rate of mtDNA in humans is approximately 6 x 10-7, or 0.0000006.
This is also why many feel that the FTDNA published mtDNA "22 generations to common ancestor" for full-sequence matches as a 95% confidence interval (which, BTW, hasn't been changed in several years) is way too optimistic. Based on that Anderson and Balding paper and using the Översti method for a non-endogamous population, we see that a cumulative probability of an exact full-sequence mtDNA match at about 87% intersects with approximately 190 birth events. If that's valid, the FTDNA estimation would be off by a factor of more than eight. Using those Översti numbers and assuming the average length of a matrilineal generation to be 23 years, the 87% chance in 190 generations equates to about 4,300 years. Just a bit before the genealogical timeframe. At a 50% probability, it looks to be about 31 or 32 generations, or about 730 years. Still not much in the way of evidence: a coin-toss chance within 3/4 of a millennium.
And did I mention that the mitochondrion is tiny? 
Well, so is its DNA molecule. It's made up of only 16,569 base pairs...give or take a deletion or insertion. As DNA goes, the whole thing isn't even large enough to register as a single segment using our current direct-to-consumer autosomal DNA testing. Included in that small molecule are a regulatory region and 37 genes that code for 13 polypeptides, 22 tRNAs, and two rRNAs. Whew! Those little puppies do a lot of work. These regions account for over 80% of all the DNA base pairs in mtDNA, and mutations there often mean real trouble to the host human cells. Like, survivability trouble. So most markers relevant to genealogy occur in the other 20% of the mtDNA.
That's the five minute summation of why mtDNA ain't all that great as a type of positive evidence for genealogy. But it can be useful as negating evidence--showing a hypothesis to be incorrect--although as we see in Bettinger's blog post even that isn't a sure thing. However, if two test takers aren't in the same portion of the mtDNA haplotree, it's bankable that they aren't related anywhere near the genealogical timeframe.
That haplotree is the hierarchical structure of haplogroups. If the haplotype (or mitotype) is your zip code (your mitotype will not be unique, so it isn't your street address), your haplogroup is the city you live in. In fact, if we take the total number of branches on the mtDNA haplotree, 5,433, and divide that into the global population, about 7.8 billion, we find that having the exact same haplogroup averages out to 1.44 million people. Having the same, deep-branch haplogroup is no more meaningful as positive genealogical evidence as is, say, both test takers happening to live in San Diego, the 8th most populous city in the U.S. Some haplogroups will have a far larger number, and some smaller; but that's the average.
I mentioned that the haplotree is hierarchical. That means there are parent and child branches. A child branch will display the same variants--the same defining DNA letters--as the parent branch, plus one or more variants of its own.
Your haplogroup H is the most common high-level haplogroup tested so far. But, interesting fact, FTDNA is only guessing what your haplogroup is. The guess is almost certainly correct, but you've only taken the HVR1 and HVR2 panels of mtDNA tests. Those tests look at fewer than 600 DNA letters at the start of the mtDNA molecule, and fewer than 600 at the end of it. The full-sequence test also looks at what's called the coding region, which runs from position 577 through 16024. Haplogroup H is defined by two mutations, one at position 2,706 and one at 7,028...neither of which is tested in your HVR1/HVR2 panels.
Then to be H1, you need those two variants plus another at position 3,010, where you'd have the nucleotide adenine rather than guanine. Then to be H1b you'd have to also show the correct mutation at position 16,356. To be H1bc you'd have all of the above plus two more mutations, one at position 152 and one at 10,325.
As you can see, the haplotree is like a funnel structure: the older, broader haplogroups are nearer the top, and the more recent haplogroups require additional identifying mutations and are nearer the bottom. In your question, it seems like maybe you're describing it in reverse. Haplogroup H doesn't have all the variants of its child branches, but each child branch will have the variants required to define the H haplogroup.
The mtDNA haplotree at YFull estimates that haplogroup H first appeared about 21,300 years ago; H1 about 15,400 years ago; H1bc about 12,500 years ago, and so forth. So every H1b has the requisite mutations for haplogroup H, but not every H will be an H1b.
At FTDNA, the country locations you see are simply where test takers have self-reported their earliest known mtDNA relative was from. There is no other comparative science behind that. Common haplogroups like H are spread all over the world, so you really can't read anything into those country flags. But in the upper-left of the display you can click where it says "View by Countries" and change that to "View by Variants." That will show you the defining mutations I described.
Have fun!
