The 16.2m/s/s mentioned is based on simple acceleration. We have a known distance - the distance between EI and Landing - and we have a known velocity delta (about 8 km/s). This allows us to calculate the average rate of decelleration required to have near-zero velocity by the time we are over our landing target. In your example, that was about 16.2 m/s/s. This happens to be achievable by most of the winged vessels in Orbiter, so you are fine there.
You seem to have a grasp on the basic theory, but haven't yet had that "Eureka" moment where you see how it all applies. Part of the problem is that you keep looking for a single "best" answer - and there isn't one. Even with a given re-entry (same vessel, mass, conditions, etc) there are many answers that are equally correct. Obviously, there are incorrect answers as well. Perhaps I should explain it as there being a "range" of correct answers. How large that range is depends on the vessel, condition, pilot's skill, etc.
You seem to believe that one particular answer will be better than all others - but that isn't the case. Granted, some answers in the range will be better than other "correct" answers, but there is still a whole range of "better" answers, not "one best" answer.
Another concept that can be confusing is the way ANT is measured. Since this is an angular measure using degrees, we tend to see it as an angle rather than a length. After all, we usually use linear measures (ie, meters) for distance. Yet, while this is confusing, it's the best way to handle it.
Let's look at two points on a single re-entry. One point is Entry Interface, and the other is the runway. Like any two points, there is a distance between them. What is that distance? It depends!
There are several valid ways to measure that distance, and each will generate a different result - all of which will be correct. We could measure straight line distance in three dimensions (AFAIK, this is how MapMFD measures distance to a base). We could measure the actual distance our vessel will travel, which will be longer. We could measure geographic distance (ground track) which may be shorter or longer, depending on the overall length.
We use angular measures for the distance because it simplifies the equation by ignoring altitude. But make no mistake, this is a measure of distance. When you use an ANT of 10, it is like approaching a stop sign at 100 km, and not hitting the brakes until you are five meters from the intersection. "10" is not within the range of correct answers.
In Orbiter, thanks to it's non-dynamic and utterly predictable atmospheric model, it's possible to calculate a precise re-entry before we even make the DEO burn. But what is the point? It's a lot of work for something you just don't need since you can easily adjust during the aerobrake. Not only that, but it wouldn't work in real life because you can't accurately predict the flight path since you won't have perfect data on the atmophere yo will be traversing.
The Shuttle re-entry profile involves a "safety margin". The shuttle maintains maintains more energy than it predicts it will need (ie, carries "excess" velocity). Then it uses the S-turns to "extend" the distance the vessel will travel so it can shed that "excess" energy. This allows the shuttle to react to the differences in atmosphere - if the air is thin and dry the turns are made "deeper", increasing the distance to account for the lower rate of decelleration. If the air is cold and thick, the turns are shallower, reducing the distance traveled. As the shuttle gets neared to the base, the "safety margin" is allowed to shrink some. Finally, the HAC turn is used to shed the last of that "extra velocity" as well as align with the runway. The more "excess" velocity the larger the cylinder. So, in real life, while the predictions are made before DEO, there is a "slop factor" built in.
As with any "range" of anwers, there are limits. The "low" end of the range is determined by how fast the vessel can decellerate, which is determined by the amount of drag the vessel can create - and also by the heat tolerance of the hull. A higher heat tolerance means I can transfer energy from kinetic to heat at a higher rate, meaning I can decellerate faster and have a shorter distance traveled during the aerobrake.
The high end of the range is determined partly by the amount of lift the vessel can generate. More lift allows for a higher "limit" to the range. In Orbiter, you can DEO a shuttle well over a full orbit in advance, and use a "skip" re-entry to increase the distance the vessel can travel. In Real Life, that doesn't work. Insulators are subject to "heat soak". Insulators don't prevent the transfer of heat, they only slow it down. After a while, heat will "soak through" the insulator, resulting in extreme temperatures getting through the insulation. There is a limit to not just how hot you can get, but for how long you can maintain the high temps. So, while a gentler re-entry will produce lower maximum heat levels, it will generate heat for a longer period of time, perhaps enough to "soak" the tiles and result in failure.
Also, IRL, the tiles have another limitation. A "skip" re-entry subjects the tiles to thermal cycling, they will get very hot, then very cold, then hot again, etc. The ceramic tech we use can't handle that - it weakens the tiles and will lead to fracturing and failure after only a few "cycles".
None of the vessels available for Orbiter model "heat soak". This means that the "correct range" in Orbiter is MUCH larger than it is IRL. It also means that the "perfect answer" you are looking for is realistically impossible. In Real Life, that "perfect answer" would likely result in a failed re-entry and death. And while Orbiter doesn't force us to account for heat soak and weather, most of us like to pretend it matters and do things the way they would be done IRL. In other words, we include the safety margin and account for it on the way down by changing our descent rate or adding S-Turns. This is how it is done in Real Life.
In short, the "perfect answer" only exists in theory - in practice that "perfect answer" will be incorrect - or at least no more correct than any other answer within the range.
