Advancement

Only a Game does a rather in-depth analysis of XP systems from one end: how much XP to a level. I don't really much care for that approach, so I'm going to give my version:

Advancement: What are my options, as a GM?

So, you've got this great idea for a game, and you're trying to figure out the best way for players to advance their characters. Do you want to use levels? Skill points? What kind of things give XP, and how much? There's more choices than you might think, and they have a dramatic effect on a game!

There are a few basic building blocks of any advancement system.

The first building block is your choice between threshold, spent, and acquired advancements.

Advancements include XP, sure, but it also includes upgrades such as new armor, an army, and a guy in congress who likes you. It's important to remember these alternative methods, to keep things fresh.

Okay, a "threshold" advancement is usually a level. It's something which is automagically gained over time. You know it's coming, you might even know exactly when, and you are not likely to be able to change the outcome.

"You get 1000 xp!" is a lead-in to gaining a level in whatever class you are. You know you're aiming to gain a level. You probably even know when you'll gain the level. You are unlikely to try to spend your XP on, say, buying a monkey.

Also, any system where you improve a skill by using it is like this, if that improvement is based on points. For example, if you shoot a gun fifty times and gain a point of gun shootery, it's a threshold advancement.

This has the advantage of being very low-focus (although the paperwork can get irritating). It also usually has the advantage of being high-anticipation, but that's not due to it's nature. I'll explain that later.

Low-focus is good: it allows players to stay in the moment, instead of being distracted by the guts of the game system. But there are other options which may suit your game better.

"Spent" advancements are points which you gain, similar to threshold advancement points, but instead of being automatically assigned to a specific upgrade, you get to choose among a bevy of upgrades.

Money is like this: with your 10,000 GP, will you buy a castle, a space ship, or a pricey hooker? Skill points are also like this: you gain 12 skill points, and you can spend them on increasing your gun shootery to 4, or your computer geekery to 5.

These systems are typically high-focus - players have to think about what they want. This gets them more involved in the guts of the game. Many players like this.

"Acquired" advancement is when the players are given things without a point value. They may have a choice, or may not.

A player finding a +3 whoopie cushion of slaughtering is an acquired advancement. They didn't purchase it, they didn't earn their way towards it knowing it was coming. It was something given to them, perhaps as a result of some work. They may have even chosen it, but the "choice" didn't involve weighing numbers, it was something like "would you like this whoopie cushion, or the left buttock of Shiva?"

It can also be "you are now more skilled at using the crossbow" or "congrats, you now have the advantage 'breathes vaccuum'."

This is zero-focus. It draws absolutely NO attention to the guts of the game, and is useful for games which are more dramatic and freeform. However, you have to be careful or it will be zero-anticipation: players have to know what to expect in order to anticipate it, and a GM who just gives them random crap at random times isn't doing it right.

All advancement systems are built on those three options. Most games have several advancement systems which use different methods of allocation.

The second half of the system is "progression".

Usually, when a skill goes from 1 to 2, that's a much more important increase than when a skill goes from 8 to 9. Yet getting from 8 to 9 often costs many times more. This can result in a diversification of skills, rather than concentration, but usually what it means is that you have to give out a whole lot more XP.

There are alternatives. One of the most popular is a "faux exponential curve". 1-2 is a big change, sure, but 8-9 is also a massive change. Because it's not just how many dice you roll: it's some kind of new ability written into the book. At level 8 you have "burn down cities" and at level 9 you have "turn planets into swiss cheese". The difference is, relatively speaking, just as big as 1-2.

The downside is that this requires you to have designed specific powers for each new level...

Another alternative is to have a linear cost - going from 1-2 costs just as much as 8-9. Then you can keep giving out dribs of XP for the whole game, and the smaller relative rewards won't get to be absurdly expensive.

Another alternative is to have "flat" skills. Instead of having a number which goes very high, the maximum is quite low. Say, five. When you max it out, you can't get any more, and have to spend elsewhere (often on a more specific sub-skill). This works but, again, requires a lot of written-up skills.

Methods which don't revolve around skills are actually easier and often more fun. You can give out items, instead. Items often have a decay rate - they run out of ammo, or they get burned away, or they are only useful on the plane of fire. This means that you can keep players from increasing their power much at all by making them spend their items to stay alive.

Many games which use money to buy items have a million variants of the same weapon. "Pistol" comes in 900 variants, all of which are minutely different. This isn't a real great idea unless you're using ammo differences to keep the players micromanaging.

Methods of advancement which don't offer a permanent advantage are, in fact, one of the best ways to go. Skills and HP aren't the only things around.

The problem with these "temporary advancements" is that they require a lot of paperwork. In a one-shot, it's not really a problem: just use chits. But in a long-lasting campaign, it puts the "pain" in "cam...g."

The longer the "lead-up" time, the more anticipation the players feel. In short, the longer it takes to get that advancement they're aiming for, the more they'll want it.

If you give out rewards too lavishly, they mean nothing. Be stingy.

Anyway, I've started to ramble. Here's the summary:

The first building block is your choice between threshold, spent, and acquired advancements.

The second building block is whether you use a system of exponential XP, faux exponential curve, linear cost, "flat" skills, or temporary advancements.

Every game should have at least two of these systems. A list of weapons is a spent, exponential XP system (a sword+1 costs, like, a zillion times more than a standard sword). Convincing the politicians in the king's court to like you is an acquired faux exponential curve system.

Don't be too lavish: be stingy. Make them have to save up and work for it.

Think hard, plan ahead. :)