The success of Old School Renaissance games has resulted in this happening a lot in mainstream forums:
Someone’s like “OSR games rule! Player skill! Sandboxy freedom!”
And so then because words like “skill” and “freedom” are coded as good and metal, everyone then tries to explain how whatever random game they like is actually secretly OSR. And then you have to be like “No, sorry Magical Owl-Touch Factory is a wonderful game but isn’t OSR” and you sound like you’re being gatekeepy and grumpy.
So it’s good to clarify things, to give us some verbal technology to talk about them:
Tactical transparency is the degree to which a common-sense idea that would be effective in the “real” situation that the game-fiction mimics would also be effective in the game.
If you need to know the system in order to be effective, the system has a low degree of tactical transparency. Magic: The Gathering has a low degree, Dread has a low degree (all risky actions are represented by pulling a Jenga block, so an idea is only as good as you are good at pulling blocks), Dungeons & Dragons 3.5 has an decreasing degree of tactical transparency as you level up (feats and spells with obscure names and tricks with actions-per-round become increasingly important). You can be skilled at these games but that skill is system-specific, it requires system mastery.
If, on the other hand, common-sense tactics that seem to make sense in real life also make sense often in the game, the system has a high degree of tactical transparency. This is an OSR ideal.
Perfect tactical transparency is impossible and would be problematic anyway: first of all it’d require thousands of rules for things like chemistry and physics (our army’s steel swords beat your bronze ones!) which would take a long time to use and consult. But the point of OSR-style player skill is to be as tactically transparent as you can be while keeping the game playable and smooth: thus the requirement for a high-trust table where all players basically agree on the GM’s good judgment and are willing to speak up when they disagree.
-Classic tactical transparency: a slow moving-monster with no distance attacks and no ticking clock (King Kong in the desert), so obviously you use missile weapons (it can’t hurt you but you can hurt it).
In many narrative systems, this wouldn’t matter—if you’re in a system where killing the monster is just a “Go Aggro” roll and it doesn’t matter whether you use a sword or crossbow, it doesn’t have a high degree of tactical transparency and so isn’t a system where player skill (or at least system-agnostic player skill) matters. Further, if thinking to use a distance weapon would only give you a +1 bonus, that is still a pretty low degree of tactical transparency--it's not giving you all the advantages you'd commonsensically have. In PBTA games, yes, fictional positioning can get you a "that won't work at all" or a "that auto-succeeds" but the whole range of options in-between is narrower, so it's less transparent. Keep reading if you don't understand.
-Fantasy tactical transparency: tactical transparency doesn’t have to be about plausibility. It can be about things that make sense by the fiction’s own rules, as long as they aren’t the just the system’s own rules. For example, one time I was playing with some people and they were trying to kill an OSR monster someone made up (Rules, Rolls, Roles) called the “Man of Wounds”. This guy is a man full of swords stabbed through him at every angle. The players keep bashing on him, landing attacks and spells, but he takes no damage. They’re getting increasingly frustrated. Someone’s aunt is in the kitchen listening to us play—she doesn’t know the system, she probably hasn’t even seen an episode of Game of Thrones—she rolls her eyes and goes “Just pull out the swords”. They try it: the guy dies. If something works by trial and error in the fiction, that’s still tactically transparent because it required a common-sense tactic unrelated to the system: when faced with the unknown or fantastic, try different things to see what works. Trial and error is a legitimate real-world tactic.
-Vagueness works against tactical transparency. For example, World of Dungeons is not much of an OSR system since, for example, all the spells (regardless of what they’re called or “skinned” as) work the same way: they just do whatever you want and have an effectiveness number. There’s very little weighing of “Ok, well this is an area fire effect, so it will result in this while this is a single-target ice effect so it will have this other result, so in this case…” . The system is not asking you to think with much tactical depth because a lot of your decisions don’t matter. You aren't deciding when a bomb is better than a beam or whether the clear advantage of the bomb is worth wasting now before you get to the boss because you get fewer bombs than beams, etc. Different kinds of games are about the different kinds of thinking they involve you in.
-Limited scales of bonus and penalty work against tactical transparency. Meaning like: if you can only get a +1 maximum due to clever fictional positioning, period, then that limits how much tactical thinking you have to do. For example, in an early draft of 5th edition D&D, it was rare to get a worse modifier than just “Disadvantage” —two disadvantage conditions stacked didn’t matter. And, in addition, being drunk granted disadvantage but also gave a minor bonus (it made you inured to pain or something, if I remember?) so it was literally better in that version of the game to be defending while simultaneously drunk, prone, and on fire than it was to just be defending while prone. If you're not the Human Torch and setting you on fire doesn’t put me at an advantage over you, the system is lacking in tactical transparency.
-When a system-specific tactic competes in effectiveness with a common-sense tactic that works against tactical transparency: like Apoc World’s rule that you get bonuses that make you more effective the more you have sex with the “Angel” PC means a tactic nobody would think up unless they’d read the book (bonus mining via fucks) is a way better way to win a battle than, y’know, planning, digging a trench, concentrating force, choosing the right weapon, etc. Torchbearer is open about how it takes over a dozen sessions to master the system, so while, yes, there is a paragraph about how the GM can give you a bonus if your plan is actually reasonably good, the system has low tactical transparency because gaming the system matters a lot more.
-Systems which abstract tactical thought itself into a character skill reduce tactical transparency: like if there's a “Read a Sitch””Discern Realities” or “Tactics” skill which gives you an abstract bonus when invoked and that bonus could matter as much or more than actually reading the situation described or actually discerning the realities from the clues given or actually thinking tactically then that interferes with tactical transparency. Like if using those skills gives me a +1 and me, as a person, taking the high ground gives me a +1, the relative value of my player skill vs the character skill in that system is reduced.
-Tactical transparency is not identical to “realism”, but overlaps with it: often a tired-ass forum n00b complaint that an abstraction in a game makes it less "realistic" is also a legitimate complaint that it makes the game less tactically transparent. See the "drunk, prone and on fire" example above.
-High-powered abilities within the fiction can work against tactical transparency, or at least against tactics: in the Fantasy Flight 40k RPGs there are some weapons, like storm bolters, which’ll just plain murdalize anyone in your way. So while you can sneak around trying to outflank and trip up the enemy, every round you spend cleverly setting up your attack is one less you spent spinning the Ballistic Skill Wheel to see if it lands on “The tyranid dies immediately”. The only skill there is reading the book enough to know that this gun is really good--and reading about every other gun and piece of armor that doesn't have a clear name like "spiky club" or "shotgun". When a game is full of powerful abilities with in-world names the best tactic is basically just about bringing them to bear asap—even though they are technically elements of the setting not the system, they still involve you less in player skill because the most skilled player will realize the best option is always just “press the nuke button”. Identifying that takes skill, just not much. -Obscure or misleading names for character abilities can work against tactical transparency: in DC Adventures you can have "Disarming Finesse" and you can have "Advance Disarm". If you're Captain Machiavelli trying to send the Delta League into the fortress and one of your team has the "finesse" one and the other has the "advanced" one, which one do you send to disarm Plasma Pig? It's impossible to tell without reading up on the system, so: not tactically transparent.
-Tactical transparency is not the same as transparency, clarity, or lack of misdirection in the fiction: If Sherlock Holmes can't find Moriarity because he's dressed as a bunny or can't hurt the Ghost Barghest of Belgium with his revolver even though the revolver should work, that has nothing to do with tactical transparency. If Sherlock can't find Moriarity because Sherlock was built with "Enhanced Perception" rather than "Perception Advance +3" or because the rules say Sherlock only can find Moriarity once he reaches Stage 5 then that is the system interfering with tactical transparency.
-Simple mechanics and tactically transparent mechanics can be completely different things. The rules of chess are simpler than those of almost any wargame. Yet the complex mesh of most wargames' rules generally end up reproducing simple military maxims: concentration of fire works, if there are rules for facing then flanking works, taking high ground works (you are above the enemy's cover so at a good angle), etc. Chess is an example of a game that is fairly simple mechanically but can be extremely tactically opaque. So a simple game system does not necessarily lead to simple choices for the player.
-5e's tactical transparency is arguable. At low levels, things that should matter in a fight do. At higher levels, feats and magic and alternate dice math make it less so, unless the GM is careful. You're picking spells: Is it better to have advantage on damage or 2 attacks? If you're trying to figure it out at all before a battle, that's not tactically transparent: that's doing system-specific math.
-A simple test: Is it worth bothering to get an opponent drunk, prone, and on fire before attempting to kill them? The more often true that is, the more tactically transparent the game you're playing probably is. There will be exceptions (no point in trying to get the robot drunk) but I'm going to make a big leap of faith and assume you get it.