One time I played a game which had one of the best text-based 2 characters combat systems I ever played (the game was awful in every other regard, so I forgot its name):

• Each side hat a number of attacks to choose from
• The damage inflicted by each attack was based on one stat of the attacker and one stat of the defender.
• Each attack caused a different stat-modifying effect on the attacker and a different one on the defender. Both effects could be good or bad.
• The status effects lasted for a number of rounds. Receiving the same status effect again did not stack. It just reset the timer.

Why did this system offer a lot of depth?

First, it is not always obvious which attack does the most damage right now, because you have to look at the status effects currently affecting the attack-stat and the opponents defense-stat for the attack. Well, when the player completely understood the game mechanics this is just a simple calculation, but very few players will analyze your algorithms so deeply and will just make educated guesses.

Second, even if you do figure out the most damaging attack, you have to look at what impact its status effects have. For example, using the attack which boosts the enemies STR might be bad, but on the other hand the enemy only got one STR-based attack and that attack targets DEX, and DEX is your best stat, so it might not be that bad. The attack also boosts your INT, which sounds good, but you still have 3 rounds of INT boost left, so you might rather want to use the opportunity to get a boost for a different stat.

I am not saying that you should copy this system. What I want to say is that if you want to design a deep turn-based combat system, make sure that

1. There are many (but not overwhelmingly many) options to decide between
2. Each decision has multiple long- and short-term consequences
3. The impact of these consequences depends on previous decisions, so the ideal decision is not always obvious.
4. Most importantly, there must not be one option which is always or almost always the ideal choice.

One time I played a game which had one of the best text-based 2 characters combat systems I ever played (the game was awful in every other regard, so I forgot its name):

• Each side had a number of attacks to choose from
• The damage inflicted by each attack was based on one stat of the attacker and one stat of the defender.
• Each attack caused a stat-modifying effect on the attacker and a different one on the defender. Both effects could be good or bad.
• The status effects lasted for a number of rounds. Receiving the same status effect again did not stack. It just reset the timer.

Why did this system offer a lot of depth?

First, it is not always obvious which attack does the most damage right now, because you have to look at the status effects currently affecting the attack-stat and the opponents defense-stat for the attack. Well, when the player completely understood the game mechanics this is just a simple calculation, but very few players will analyze your algorithms so deeply and will just make educated guesses.

Second, even if you do figure out the most damaging attack, you have to look at what impact its status effects have. For example, using the attack which boosts the enemies STR might be bad, but on the other hand the enemy only got one STR-based attack and that attack targets DEX, and DEX is your best stat, so it might not be that bad. The attack also boosts your INT, which sounds good, but you still have 3 rounds of INT boost left, so you might rather want to use the opportunity to get a boost for a different stat.

I am not saying that you should copy this system. What I want to say is that if you want to design a deep turn-based combat system, make sure that

1. There are many (but not overwhelmingly many) options to decide between
2. Each decision has multiple long- and short-term consequences
3. The impact of these consequences depends on previous decisions, so the ideal decision is not always obvious.
4. Most importantly, there must not be one option which is always or almost always the ideal choice.