Incentivising behaviour is bad for your game.
Axiom - Most players, most of the time, will take the most optimal option.
Incentives create optimal choices - a mechanical reinforcement for taking certain paths - whether negative (you will lose XP if you do bad things) or positive (you will gain XP if you RP well). Optimal choices, once discovered, become boring. You can incentivize the human out of the decision - there is only A Correct Choice. By removing Correct Choices, you widen the field of play - more possibilities will be evaluated, considered. In the absence of a Correct choice, everything becomes possible.
Are you interested in seeing players make choices with their characters or just slotting in to your grand design? RPGs can be more than Rube Goldberg machines culminating in your intended experience. RPGs should be more than this - and removing the idea of incentives for desired behaviour is key. Desiring behaviour in play is something we do to opponents, not collaborators. The idea of the only reason players would do something because ‘Number Go Up’ is video-game design. Besides which - if they don’t want to do something, why is that a problem? Are they not an equal contributor to the game?
When a player acts against optimal choices, we are engaged and excited - when they reject Number Go Up, refuse the mechanical temptation and take a choice as a character - doing something which means enough to them to take a firm decision. Instead, a game could be made entirely of these moments - taking the situation as your springboard, without reference to an external, invisible, intangible rule-set.
Many will reward ‘good’ behaviour with XP. You save the village, you only take prisoners, you’re nice to everyone. Being good for a reward isn’t being good - it’s just optimal play. Acts of sacrifice and kindness in opposition to ease - when actively chosen despite potential loss - are what marks ‘being good.’
Experience and Advancement
The ultimate example - and perhaps the oldest, in the field of games - is the Experience Point. How you obtain these, often, is described as ‘the point of the game’. Cementing an objective like this can lock off possibilities, and encourage ‘mechanical thinking’ - we should do action to accrue more XP. However, some schema are more indifferent - they do not dictate how they must be achieved. The classic GP for XP formula does not have to specify how the GP is gained - violence, theft, mercantile aspirations, MLM etc. But beyond this, why do we gather XP?
Advancement-as-reward poisons the well of player action and agency. It is the expression of Number Go Up - some systems attach more than numbers, but the core is the same - this is the carrot being used to drive gameplay. By removing Advancement, we ask players - why are you playing - what do you want to do, freed from the tyranny of Number Go Up. This is agency and responsibility - to find worthy and entertaining objectives in the absence of an objective measurement.
Of course, characters can still grow - simply advance them internally, without reference to some prescriptive metric of advancement. If they choose to engage with local powers, ranks and titles might work. Equipment is a concern for bandits, soldiers and guerillas alike. Seeking notable treasures (and techniques) are staples of adventure fiction - characters do not become static, but collections of experiences.
An elegant example of this can be seen in What time near dawn did they climb up the other shore, Drying their wings?
But What About Themes?
A common use of Incentives is to encourage/reinforce/enforce tone - for doing things which align to the source fiction, you are rewarded. Instead, we could talk to our fellow players about what we’d like to see and agree to work towards it without the use of incentive - why do we need our efforts ‘rewarded’? Isn’t playing fun?
We can trust out playing companions to build towards those themes - or let them drift and change in the chaos of play. Anything is better than trying to subtly encourage people like children.