Taking your Golem party examples i will take it from the perspective of someone playing the game, not designing it.
Assuming the given skill as a 75% reduction for 3 turns useable by only one party member, and assuming also that using this debuff switches golems AI to prioritize magical attacks.
It would create a nice tactical strategy, this is the necessary strategy to clear this encounter then? (making golems hard physical hitters and average or bearable on magic) in this case the debuffer is again a required member that when lost puts the party in a huge disadvantage, makes that party member a revival priority and becomes a necessary resource dump of MP recovery items to mantain the debuff, in the case the debuff stacks on renewing the valid quantity of turns gained grows exponentially as long as you have enough resources to keep adding to the stacks.
Assuming the debuffer attacks in the middle-late turn phase, after 2 golems and 2 party members got their turns.
Turn 1 golem 1&2 attack (which lets assume doesnt kill anyone), debuffer reduces 75% atk, in case of a normal non adaptable AI the other 2 golems do little damage, all got 3 turns of debuff, turn ends 2 stacks remaining.
Turn 2 golems 1&2 do magical attacks (which assuming this is the valid strategy, are endureable) you recast the debuff which, assuming it stacks, now adds up to 5, turn ends, stacks lower to 4.
Every turn you gain 2 stacks/refresh (in case it only renews) and the rest of the party will deal with the enemy, as a player i wouldnt give two... things about on locking down one party member to lock down the enemy party, and if the rest of the party cant survive or quickly kill the golems whitout this debuffer guy, well guess who's the bigger crutch now.
Now this is only in your "magical/physical golem" scenario. Will all enemies switch attacking types when affected by this debuff? We fall on the crutch again as this one guy is the one controlling the enemy behaviour, attack pattern and type of reaction, worse yet, it will fall on repetitiveness and sometimes implausible situations (debuff a bandit, will the bandit switch to fireballs now?) so like it or not, some encounters will be completely shut down by this debuff or all encounters will change behavior depending on it (unless we fight a party of 4 mages making the debuff useless), even in the case the debuffer attacks after some enemies you will only have to endure the first turn which if you cant whitout the debuff, we fall on "necessary" ground again.
Note im not even taking into account boss battles anymore, as from my point of view this skill's existance makes everything but boss battles revolve around it.
In the case you want to keep it on such absurd numbers, make it one turn and on cooldown/high cost so its a brace/one turn barrier and the player can regain momentum on a critical situation, on a free cast, these high amounts make it a balance nightmare and something i and many other players would gladly abuse.
When a move becomes needed to make the battle possible a turn is being expended every so often to make the game playable. Its a mechanically flawed design.
Edit: one turn high cost 75% single target debuff..oh man im using this. Gonna call it topple.
You guys both misunderstood my post, and I believe it's because you started with a faulty assumption that the debuff was "necessary" to reasonably win the encounter.
Please start with the opposite assumption - assume that the battle is well-designed and can be won using standard damage techniques alone with no debuffs. Now, with that in mind, introduce the -75% ATK state as a possible option for one of the characters to use, and go through the battle again. You'll see that the outcome didn't change that much, meaning that the debuff is balanced. (If you use a -30% ATK state instead of a -75% ATK state, you'll find that using the debuff is absolutely not worthwhile.)
Use of Turns is a helpful shorthand that becomes necessary once there are a lot of factors at play, but since this example is so simple, we can actually go through turn-by-turn and crunch the numbers.
I'm going to start with the assumptions that Golems each have 400 HP and deal 60 damage with an attack or magic spell, whereas Characters each have 400 HP and deal 100 damage with an attack or standard damage spell. You can choose almost any numbers you want as long as they're not ridiculous, and this example will still hold.
With Debuffs:
Turn 0: All characters have 400 HP; all golems have 400 HP.
Turn 1: Character 1 debuffs Golem A. Characters #2-4 attack Golem B (no point in attacking Golem A who was debuffed), dealing a total of 300 damage. Golem A attacks Char 1 for 15 damage; Golem B and Golem C attacks Char 3 for 60 damage each (120 total); Golem D attacks Char 1 for 60 damage. Characters now have 325/400/280/400 HP. Golems have 400/100/400/400 HP. Golem A has a debuff.
Turn 2: Char 1 debuffs Golem C. Char 2 attacks Golem B, killing it. Chars 3 and 4 work on Golem D, dealing 200 total damage to it. Golem A and Golem C each deal 15 damage to Char 2 (30 total). Golem D attacks Char 3 for 60 damage. Chars now have 325/370/220/400 HP. Golems have 400/0/400/200 HP. Golem A and C have a debuff.
Turn 3: Chars 1 and 2 attack Golem D, killing it. Chars 3 and 4 start on Golem A, dealing 200 damage. Golem A attacks Char 4 for 15 damage. Golem C attacks Char 1 for 15 damage. Chars now have 310/370/220/385 HP. Golems have 200/0/400/0. Golem A's debuff wears off at the end of the turn; Golem C still has a debuff.
Turn 4: Chars 1 and 2 attack Golem A, killing it. Chars 3 and 4 start on Golem C, dealing 200 damage. Golem C attacks Char 4 for 15 damage. Chars now have 310/370/220/370. Golems have 0/0/200/0. Golem C's debuff wears off at the end of the turn.
Turn 5: Chars 1 and 2 finish off Golem C before it can attack. Chars still have 310/370/220/370.
Party has cleared the battle, having lost a total of 330 HP (and having spent MP on two debuffs).
Without Debuffs:
Turn 0: All characters have 400 HP; all golems have 400 HP.
Turn 1: Characters 1-4 attack Golem A, dealing a total of 400 damage and killing it. Golem B and Golem C attacks Char 3 for 60 damage each (120 total); Golem D attacks Char 1 for 60 damage. Characters now have 340/400/280/400 HP. Golems have 400/0/400/400 HP.
Turn 2: Characters 1-4 attack Golem B, killing it. Golem C attacks Char 2 for 60 damage. Golem D attacks Char 3 for 60 damage. Chars now have 340/340/220/400 HP. Golems have 0/0/400/400 HP.
Turn 3: Chars 1-4 attack Golem C, killing it. Golem D attacks Char 4 for 60 damage. Chars now have 340/340/220/340. Golems have 200/0/400/0.
Turn 4: Chars 1-4 attack Golem D, finishing it before it can attack. Chars still have 340/340/220/340.
Party has cleared the battle, having lost a total of 360 HP (and having saved MP by not using debuffs).
As you can see, in this very apples-to-apples example, the balance and the outcome in using the -75% Debuffs, versus using all of Char #1's turns to attack the Golems instead, ends up being very neutral. (If you had reduced the ATK Debuffs to -30% instead of -75%, the party would have lost
492 HP in this battle, meaning that these MP-costing Debuffs were actually worth far less than the value of the turn it took to cast them). The player was able to win the battle about as easily in the first example as the second, meaning that there is no "crutch" here - just a choice between two different styles, one more offensive and one more defensive.
The very close outcomes mean that,
in this simple example, the Debuff is very well balanced at -75%, even though intuitively you would think that this percentage would be way too high.