Topic: Experience system

[fehknt]

Now, it's been a given that players will gain experience and gain levels and get better, like any RPG.  

I'd put forth two propositions:  1) experience is a linear sacle, every level you need 100 XP to advance, and so to make 50th level you need exactly 5000 XP.  This makes any XP transfer setup much more feasable.  Just make it so you get less XP based on the difference in levels between the character and the (to borrow a D&D term) challenge rating of the "baddie".  Add rules for multiple "baddies" and multiple people attacking.  Tada.  Ok, it's a little more complicated than that, but I think it's a good way to go.
2) um.  I may have frogotten 2.  OH YEAH!  no "hit point" or "endurance" or things like that increasing.  Make the person harder to hit or something on higher levels.  People don't magically get harder to kill as they get more experienced: just because you're good doesn't mean that you will take more shots to the chest to die, but it may mean that it's harder to hit you.
2.5) Maybe this makes items/armor particularly important? Well, in a war zone, don't you think that armor is going to be pretty important in real life?

Lets hear other people's thoughts on how advancement should work!

[Morgul]

G.U.R.P.S.

Atleast that's what we're thinking of adapting. (It was suggested by our friend Cory.... I don't know much about it right now.)

I'll post more when I've had some sleep, and can think more.

[contingencyplan]

Well, I'm not familiar with the GURPS system, since I never really got into the D&D style of RPGs.  I have a passing familiarity with the D20 system from playing Star Wars: Knights of the Old Republic, but that's neither here nor there.

However, from a game standpoint, I was kinda thinking of experience as a continuous function, rather than a discrete one.  By "continuous" I mean that you notice yourself gradually getting better, rather than having to wait until you "level up" to see a difference.  This will especially work with things like aiming weapons - the more you fire your weapon, the better you get at aiming it, and the more accurate you are (e.g., less recoil, etc.).  Naturally, there's still a hard ceiling (likely weapon-specific) to keep people from being dead-accurate with a pistol at 1000 yards.  Also, there is a system of diminishing returns; you won't keep getting the same advancement in skill each time you fire your weapon and kill somebody.  

A problem with this system: it is potentially fairly easy to abuse.  If we make firing the weapon the basis for advancement in skill with the weapon, then people will simply stand around firing their guns into the air until they have the desired expertise with the weapon.  Thus, advancement in skill would be dependent upon reaching a goal.  To continue with accuracy with a weapon, then you have to be aiming at something (e.g., a simulated target or an opponent) to gain skill when you fire your weapon - you only gain skill with successful hits.  Better hits (headshots, for instance) net larger gains.

As I mentioned in the death topic, some skills can be advanced by spending time in the simulator.  The above rules (ceilings and diminishing returns) are especially evident in the simulator, since nothing beats using a real weapon in a real combat situation.  However, for gaining an initial skill with a weapon, the simulator gives the ability to learn without a potential for dying because your skill is too low.

As far as how this affects things, we could have the recoil skill, for instance, be a percentage.  The player fires a weapon with a certain amount of recoil, and the skill is then applied to that recoil force as a percentage: If a weapon has a recoil force of 10 (whatever "10" means), and the player has a recoil skill of 20 (or .20, or 20%, or whatever desired representation), then the final recoil force will be 8 (10 - (.20 * 10) = 8 ).

Personally, I think that this will make the game more realistic.  After all, in real life, we don't gain skill as levels, but as a continuous improvement over time.

~Brian

[whitelynx]

However, from a game standpoint, I was kinda thinking of experience as a continuous function, rather than a discrete one.  By "continuous" I mean that you notice yourself gradually getting better, rather than having to wait until you "level up" to see a difference.

Definitely. I would like to see experience work this way as well, since it's much more realistic.

This will especially work with things like aiming weapons - the more you fire your weapon, the better you get at aiming it, and the more accurate you are (e.g., less recoil, etc.).  Naturally, there's still a hard ceiling (likely weapon-specific) to keep people from being dead-accurate with a pistol at 1000 yards.  Also, there is a system of diminishing returns; you won't keep getting the same advancement in skill each time you fire your weapon and kill somebody.

I'd almost think that the diminishing returns would provide enough of a ceiling that we wouldn't need to add hard ceilings... but that depends on how we build our equations.

A problem with this system: it is potentially fairly easy to abuse.  If we make firing the weapon the basis for advancement in skill with the weapon, then people will simply stand around firing their guns into the air until they have the desired expertise with the weapon.  Thus, advancement in skill would be dependent upon reaching a goal.  To continue with accuracy with a weapon, then you have to be aiming at something (e.g., a simulated target or an opponent) to gain skill when you fire your weapon - you only gain skill with successful hits.  Better hits (headshots, for instance) net larger gains.

That sounds like a fairly good idea. Also, hitting a moving target should be worth more than hitting a stationary one. (and the faster it's moving, the more it's worth... although it would have to be movement relative to the player's view... someone running directly at you, not weaving side to side, provides a very easy target.)

As I mentioned in the death topic, some skills can be advanced by spending time in the simulator.  The above rules (ceilings and diminishing returns) are especially evident in the simulator, since nothing beats using a real weapon in a real combat situation.  However, for gaining an initial skill with a weapon, the simulator gives the ability to learn without a potential for dying because your skill is too low.

I'm still a bit wary of the whole simulator thing. I like the idea as far as being able to learn things without as much risk to the character's life, but the method has to fit into the game in general.

My biggest problem with the simulators is the fact that in this day in age, simulators would be pretty damn advanced. (pretty close to real life) We have to have an excuse for _why_ they aren't as effective as real-life combat for gaining experience.

Another way to do this would be allow one character to "teach" another, passing on knowledge. They of course wouldn't be able to teach the other character more than they know themselves, but it's a way to learn new things that involves no real risk to life or health.

Dave

[Morgul]

I do like this idea, and it's been kicking around in my head for a while... Including the simulators. However, how much do we weant skill to be in the hands of the character, and how much do we want it to be in the hands of the player? I mean, sure, the character needs to develop a skill with a weapon... (I'm thinking going to a simulator, and taking a basic weapons proficiency course with that weapon)... and even training accuracy... however, some of that is the player's acuracy (which iironically will get trained while his character's is)

So, how/where do we make the distinction?

[whitelynx]

This is a good point. If you think about it, what we're doing in our basic game design is bringing more and more player-skill-based elements into an RPG-style game. (character-skill-based) The more player skill is involved, the more enjoyable it is, and the wider appeal it has. On the other hand, we want to balance the two, because it is still a simulated universe, an RPG. We can't have all the 1337 players going around getting headshots on every other person in the game and getting away with it.

Dave

[Morgul]

[...] We can't have all the 1337 players going around getting headshots on every other person in the game and getting away with it.[...]

Too true. Here's my idea: We have character accuracy and player accuracy work like this:


As the character's accuracy improved, the radius of where it might actually hit (instead of where the crosshairs are) would diminish. (obviously the picture above would be for a character who's a really bad shot.)

This way, it matters how good the player is, how often they've played thier charactr (to kno how good they are with certain weapons) and how accurate the character is.

Also, another thing to note is that to me, if the bullet hits, it hits.... I don't want to see bullets going through a character, just because they missed his 'AC', or something.

[contingencyplan]

morgul:

As the character's accuracy improved, the radius of where it might actually hit (instead of where the crosshairs are) would diminish.

This is exactly what I had in mind.  

whitelynx:

I'd almost think that the diminishing returns would provide enough of a ceiling that we wouldn't need to add hard ceilings... but that depends on how we build our equations.

The reason for a hard-coded ceiling is that regardless of how much we do a "law of diminishing returns," the function will still go towards infinity.  The player, given enough time, will be able to have 100% precisely accurate aim.

[Precision = reprodibility; accuracy = correctness.  They are mutually independent - if you throw 5 darts and they all hit in the same area on the board, then you are precise, even if none hit the bull's eye.  On the other hand, if all 5 land at random points on the board, but one hits in the middle of the bull's eye, then you are accurate.]

Thus, given enough time, the player can eliminate the whole reason for having skills - to introduce the character's effect on gameplay.  Then, we have the l337s going around and headshotting everybody.

whitelynx:

My biggest problem with the simulators is the fact that in this day in age, simulators would be pretty damn advanced. (pretty close to real life) We have to have an excuse for _why_ they aren't as effective as real-life combat for gaining experience.

It's not that they aren't effective - they're as effective as real-life is.  However, as advanced as they are, they still are no match for a combat situation.  I would still say there's a diff between firing a weapon when you know you can't be hurt and firing that weapon knowing that if you do something wrong, you're dead.  The simulator will give you experience (using the term loosely) in the general use of a weapon, while the real life experience will teach you the art of actually using it.

whitelynx:

Another way to do this would be allow one character to "teach" another, passing on knowledge. They of course wouldn't be able to teach the other character more than they know themselves, but it's a way to learn new things that involves no real risk to life or health.

I very much like this idea, especially in conjunction with the simulator.  I think that we could award more experience when you are trained by a person, rather than by a computer / NPC.

~Brian

[Morgul]

Hrm.. I like these ideas. However, we *do* need a hard limit.

Note on simulators: I think that if you spent your entire life simulating how to use a gun, you might be better at it then a veteran soldier... However that soldier has only been using it for ten to fifteen years, you've been doing it for thirty.... That's a large time difference. I think that simulators should have a quicker falloff on diminishing returns then actual combat, but that's the only difference. If you spend more time in a simulator eventually you'll get as good as someone who's been in the field for three months.... but it might take you five, or six to get that good. (It's like real life)

Also, don't foget the player aspect; They themselves are affected by the thought of thier character's death. (and the desire to avoid that). I bet that they will perform worse in the real thing till THEY gain experiance... regardless of character stats.

[fehknt]

the function will still go towards infinity

Um, not all.  It's not going to be a really often used calculation (compared to the physics stuff, for example), so it should be ok to use a "hard" calculation like a logarithm...  Which does not approach infinity in the limit as time goes to infinity.  EG: if the max is 100 skill level, something like (new skill level)=(old skill level)+ln(101-old skill level).  that way, you gain skill quickly as you learn the ovbious things about said skill, but eventually, you just can't learn any more, when you are adding the ln(1)=0.

You can get more complicated than this, but this is a simple example of a function that will act as needed without a "hard limit", there is simply a point that the function does not go above, ever.

[whitelynx]

/me votes for logarithms.

[Morgul]

Agreed.

/me votes for a TI-82 emulator to be embeded in the app to do the calculation.

/me is exhausted right now, and will likly regret this post.

--Crihs

[fehknt]

actually, I just yesterday saw a c++ game design book from a friend of mine that suggested using logarithims specifically for advancement and experience.

ok, not terribly important, just interesting.

[Morgul]

Well, it is interesting... nice to see other people suggest the idea. I really think it's the way to go.

[topher]

My opinion is that in game skills, feats, abilities should have a 40% influence and 60% on the player at the keyboard (or joystick).  I',m not a great twitch player and i wouldn't want to have a gimped character because of it.  I'd still like to take out that 12 year old who plays 23/7 working on the fastest reflexes.

[fehknt]

well, yeah, but the one who plays 23/7 probably will have more experince (character) than you anyway <:o)

Seriously, though, I'd still agree with the approximite balance you suggested -- that way, there isn't a few character "builds" that are more powerful than any other character (things can't be perfectily balanced, it just doesn't happen), and a fair amount involves the player.  That said, there's a lof of stuff in this game that isn't "tiwtch".  The RTS aspect of things is (probably) pretty much no twitch needed.