Maybe have a simple bar graph to show how your rating is to a particular group. But have most of the true effect visible in the gameplay. If your liked, then people will say Hello as you pass. If your loved, then they will get out of your way and compliment you. If your worshipped, they will do the same above except bow down as well. If neutral, they do nothing. if your iffy, they give grunt of disgust. And so on. It should be fairly clear by what you can do and how NPC's interact with you which should give a good idea as to where you are at. The more variation and the more common these interactions occur, then the more the player will notice.

If I go into a village and there's 50 people.
Scenario #1: Say I spent about 10 minutes there and 4 people say hi and the 1 person i traded with is pleasant to talk to. My rating is good with these people. If this is about an average amount of interaction anywhere in the game, then I'd be checking that graph a lot.
Scenario #2: I spend 10 minutes again, 52 people say hi (some a couple times as I walked by a few times), 1 person i traded with is pleasant and 3 others i walked by asked if I wanted to do business with them. I'd probably look at that graph a lot less. I'd know just by how the NPC's are acting around me as to how they feel.

--------------

I'd like to see a rating system which has different levels.
Like Shock, Actual, and Short-term.

This is what the different systems do and how they interact.

Shock is immediate response from NPC's as to what you're doing. You stab a dood, they react. You stab an enemy they cheer. This may be the mechanical feel but it has absolutely no long-term effect. They may cheer for you but the next day or even a few minutes later they won't care. Another thing, if you go to another town, your shock rating is ignored... it's only valid in a small area around whatever you did.

Actual is your actual rating. This is how NPC's and whatnot view you as. If you're good or not. If you enter an area you have not been to before, your actual rating is what is seen here.

The Short-term is a little complicated. Basically it's similar to some more modern health bars in some games I've seen. Say you have an "actual" rating of 200 (0 is neutral) towards group "foobar". you perform an action like giving them food which gives you 20 points. Your short term will go up 1 point per minute and after 20 points, it will last 3 hours at 220 points. Before start dropping down 1 point per minute until it reaches the actual rating. Short term may affect a little larger area than Shock. Like maybe the surrounding towns but a town far away won't care what happens here. (Explaination is that the news hasn't reached them yet or it's just too far for it to matter to them.)

The twist is the actual rating will only move 1 point (or maybe based on a more complicated formula) toward what the short term rating is per hour. So again, you got an extra 20 points for short term. so for 3 hours you are at 220 and this brings your actual up to 203 points before the short term is back down again.

It's a little complicated but each action will have a short term and long term affect.

believe it or not, i have more complicated systems in my head [wink]

Another simpler example which I just thought of is the honour system of Americas Army. Each match has about 7 games. If you shoot a teammate you get ROE (very bad) which makes your score go to the negative. Each kill / mission completion brings your score up. Your honour is only changed after each match, not each game. So this means your overall style of playing is what affects your honour.

I don't know, having 52 people talk to me as I go by would make just normal traveling take forever, unless hi is all they say and/or you're allowed to move while the text they're saying is up (or if the speech is actual audio, of course). It'd take even longer if every now and then someone wanted to babble forever, although I think that would be a nice feature.

Also, you have to consider that no matter how good your rating is, some people will only barely tolerate you at best. I hate games where every person in existence gushes about how cool and 1337 and godlike you are. Problem is, would differences in people make it hard for you to tell how good your overall reputation was? (Done with moderation, this could further disguise the system behind everything. If a guy you've known to steal and beat up his wife starts patting you on the back, you know you're in trouble heh.)

Excuse the random rambling, I'm not awake yet. :P

I think the best solution is to make a system that is more elborate then a simple numerical value. You could make a matrix and title system instead which reflects the qualities that the "class" is suppose to represent. For instance in your example:

Warrior
Reputation[rep]
Righeousness[rig]
Devotion[dev]
Humility[humi]
Humanity[huma]

So these are all linked together and its how the player balances these factors that determine their charater growth. There are also white and black marks that reflect action or deeds performed by the character that have an effect on the character matrix depending on the mark it may or may not be permanent. You could also use how they player balances the matrix to determine character growth for instance a matrix weighted heavily towards righteousness might make the character a paladin or similar thing.

The actual mechanics of how it works could be something like this:

Reward level of deed compared to current level of quality equals change in quality level and update to last quality related action performed.

so if you have action:
feed poor(+1 Rep, +5 Huma,(if all food given +5 humi))
and the current qualities levels are
rep: 100
rig: 70
dev: 10
huma:10
humi:10

then
comp(rep, rep bonous) = 0 change
comp(huma, huma bonous) = +5 change
comp(humi, humi bonous) = +5 change

So since the characters reputation is so high feeding the poor no longer increases it, but their humility is still low enough that it increases because of their action. If the player wanted to increase thir reputation they would have to perform bigger more noteworthy deeds.

At the system time this system incorperates a decay factor so if the only act of humility the player performs is to one time give their last ration to a begger then their humility would slowly decreases over time since they are no longer humble.

As for white and black marks these reflect significant deeds or actions that player has performed in accordance with their code of ethics. For instance if the warrior killed an innocent they would recive a permant black mark called "Slayer of the innocent" which would permently effect their qualities it may limit their humanity to half their maximum quality level, and decrease any reputation reward by 25%. Because of the taint of killing an innocent that travels with them where ever they go.

Depending on how you handled character growth, you could have each quality in the matrix have the same basic maxium level perhaps 100 per character level, or you could use a weighted pool system where the total pool points is equal to 100 per character level and the values for each quality is percantage of that pool based on their actions. So increasing reputation would decrease the other qualities which would make balance as important as indvidual actions.

Those qualities could then relate to character stats, npc reactions, skills, cultural relations and so on.

Hope that helped a little although it may seem a little random.

I think this issue has more to do with player psychology than game design, though game design can mitigate the tendency of a player to "play the system".

Over at www.indie-rpgs.com, there's an excellent almost academic treatise on rpg game theory. In effect, there are 3 kinds of players: gamists, narrativists, and simulationists. Gamists prefer games with definable winners and losers, narrativists place the highest emphasis on telling a story, and simulationists favor a realistic exploration of the setting the game is in (this is incredibly glossed over and can be nitpicked...I highly recommend going to the website and perusing their articles. It has some of the best layman RPG design I know of).

Gamist type player will attempt to play the system. The system exists for them to exploit to gain every possible for their ultimate goal....to win. Winning may be succeeding at an objective, or more likely, to increase power, prestige, experience, money, etc etc. Most sports games, wargames and strategy games attract the gamist types.

Narrativists on the other could damn the rules altogether if they get in the way of weaving a good story. For them, it is the telling of the tale that is important, not how the story itself is weaved (and the story is ultimately weaved through the system rules). So grandstanding, being flamboyant, or being tragedian are often hallmarks of this player type. Being larger than life is usually the goal of this group.

Simulationists are perhaps the rarest and most demanding, for they are a mixture of the two. The system is important in that it reflects the versimilitude of the world around them. If you can't plausibly deny the inputs or outputs of the system, it will turn this group off. But unlike the Gamists, the rules are not there to exploit, but to create the constraints for what is possible in the world. Simulationists want to explore possibilities, both physical and emotional. It is the emotional exploration of the game reality that makes them similiar to narrativists. If a character becomes enraged, they want to know what factors it was that enraged him as well as what bonuses and penalties it would give him. "Why?" is the hallmark of a simulationist.

Game design can be done in a such a way as to avoid these possibilities. The easiest is simply information hiding. In real life, we don't know our exact scores in strength, intelligence and dexterity, so neither should our characters. Nor should our character know exactly how wounded they are (what if its internal bleeding?) nor should they know exactly how skilled they are. Information hiding is also a great way to simulate over or under confidence. By hiding the specific concrete values of traits, it becomes very difficult for players to play the system. Instead, they will have to act on their wits, judgements and emotions to decide what to do.

Unfortunately computer RPG's can't do as much as PnPRPG's can about this situation. GM "fudging" or obfuscation of task resolution chances are a great way to make the player who insist that their character has such and such a bonus because he's got the advantage "Combat Sense" despite being temporarily deaf and blind from a concussion grenade going off is simply wrong (by nodding his head, saying "mmmhmm....ok now roll the dice. Oh gee, looks like you missed Ted".

Truly getting into the world and the culture of the system is what role playing is really about. And RPG'ing on the computer is just light years behind PnPRPG's in this regard. I probably sound like a broken record, but I really really recommend people who want a true roleplaying experience to give up their precious computer for awhile and play an PnPRPG session. Even IRC PnPRPG sessions aren't quite the same because of the lack of so many social cues (like tone of voice, accents, and body movements) as well as just missing out on the social experience of hanging out with some people and eating really bad food.

When ever I start playing a game, I tend to be good because there MIGHT be a hidden system keeping track of my being good or bad and preparing for some retribution. On the Kolberg's Moral Scale ( http://www.aggelia.com/htdocs/kohlberg.shtml ), this is bit of Stage 1 and Stage 2, namely PUNISHMENT AND OBEDIENCE and INSTRUMENTAL EXCHANGE, respectively. If you want people to appeal to higher morality, you have to compel them, not make deals with them.

I think the only real solution to this would be to just make the game and setting so immersive that the player *wants* to live up to those ideals. I know it's vague, but I believe that's the only way you can get players to not just "gaming the system". Those ideals would have to be an integrated part of the entire game world, not just the player character. If everyone in the game expects you to behave like this, and others act like it too, then the player might be more inclined to follow suit.

Just hiding the feedback, removing numbers and stats won't help much by itself. As have been said, it's really a question of psychology more than game design. If you just delay feedback or obscure numbers, all you're really doing is confusing the player, making it harder to figure out what to do. True, that can encourage the player to play the game differently, (Or at least, discourage from "gaming the system". But that in itself isn't enough. You also have to encourage some other playstyle, otherwise the player would just get frustrated... He's not allowed to just go straight for maximising his stats, and he has no particular reason to follow those cultural ideals.

Actually, if you think from square one that the player is just going to be gaming the game (which, the vocabulary alone should imply it), then it possible to make a game where the player thinks hes gaming it, but its also teaching him some lessons.

For instance, lets take the case of some farmers that have a possessed goat out in the field eating the crops and the farmer's children. The farmer runs up and asks for help. Lets make a chart of options and their effects to a Reputation scale.

OPTIONS                          POINTSIgnore the farmer.                -3Ignore the farmer, but help.      +3Promise to help, but walk away.   -5Promise to help, and help.        +5

Now you have a scenario where the player will attempt to assist everyone for the +5s all over the place... but lets penalize the player for that. How about when the reputation gets to high, everyone starts asking for help on all the most asinine activities. My cat is stuck in a tree, the garbage in my house smells bad, Bobby called me a stupid-head. Now the player is getting exploited by the town he was so eager to help.

Lets look at the reverse, the player helps nobody. Eventually, his reputation would plummet into the region where nobody wants any help from him. For every good action he performs, his reputation may go up, but he'll still get dirty looks and people being angry with him.

One last thought, is that your player would definately scheme a system like this, but it doesn't neccessarily have to be a universal scale. Pretend we have a By-Town reputation, and the player can get a bad reputation in one town, but be a total hero in another. That'd lead to the interesting gameplay that an antagonist could wander from town to town and drag down that reputation while you're away.

Is it still a game? Yes. Will you have players trying to strategize it? Yeah. But are they performing the good deeds anyways? Yeah. Stop and wonder for a moment by what grounds we normally dictate a person evil. If they spend al of their time working at a Soup Kitchen, are they evil? What if they then go to a brothel after, and not to evangelize?

Considering that this will apply mostly to RPGs, and games where there are many stats that can be improved, why not just set a hard limit on improvement. If there is only a finite amount of XP (or whatever) in the game, and nowhere near enough to fill out all the player's characteristics, then they will concentrate more on deciding how to 'spend' it. This may mean certain contrivances (why do the players have the choice of how to spend it in the first place?), but it has the bonus that the game can be replayed in many different ways, depending on how you divide XP among stats. No more killing infinitely-respawning 1XP monsters for hours on end.