Some things in your game accumulate — they stay or grow. Others are cyclical — they come and go.
Ideally, things in your game should grow, and thus form a ramp.
Having scores that grow during the game is a kind of accumulation, but a useful mental exercise is the following:
Take a look at your game (or any game).
Ignore the rules and scores for a second. Where are the physical components going?
If you have a hand of cards, where are they going? What about resources, and other objects? Are they being discarded, or are they amassing in front of you?
If your components aren't amassing, they probably should be.
Accumulation does all kinds of good things for your game. As well as the benefits of having a ramp, that I've already talked about, it ties the turns together, and the feeling of progress is a very strong enjoyment factor.
It's quite okay to have cyclical resources or money. However, they should be spent on something that provides permanent benefit.
In Splendor, players take three gem tokens per turn, or they spend gem tokens, to buy cards. These cards provide points, but they also provide a discount on all future purchases. The gems are a cyclical resource — they're constantly leaving the player, but all the bought cards accumulate.
As I'm growing tired of crap euro games, I'm increasingly looking at the game being played, or at the back of the box. If each player doesn't have their own board or tableau, the game is probably cyclical and terrible, and I don't want to play it.
I like Agricola, so I made my own farming game.
You ran a farm, and periodically, "guests" would come along. You could
feed them, and provide housing, clothes, or a fire, to keep them warm.
This was how you got points. This was open-ended, so you could satisfy
the guests on any scale you liked. Players would often kill off all
their animals, or create a great bonfire, rather than build things like
buildings or fences. This degree of freedom was great, but after cashing
in half their farm, a player was left rather deflated and aimless.
They were basically back at the start of the game. What would they do
now?
Allowing players to make the game highly cyclical was a huge
mistake. They shouldn't have been able to neglect and undo their
accumulation to that degree. I didn't understand this mistake at the
time, and it killed the game.
Cyclical play is a huge trap that adversarial games fall into, as attacking the enemy is typically cyclical, not accumulative.
Radlands has no accumulation. It gets lots of fun from other sources, but the absence of a ramp made it much harder to design.
I've played numerous "battle" game prototypes, where the player plays a variety of cards, in order to fight enemies on the board. I get the cards, I play them, and the enemy is defeated. It's nice that I'm closer to victory, or have gained some points, but I still feel exactly the same as I did before. I'm not any stronger. If I'm spending my resources, I want to be rewarded with an upgraded gun or something.
In 7 Wonders, each card taken is either a resource or prerequisite, that stays for the rest of the game, and pays for future things.
In a bad game, a player takes resources, and spends those resources to achieve victory.
In a good, accumulative game, a player gains resources, spends those resources to create systems, and those systems create victory. (The systems might just create more resources, which create victory. That's fine, too.)
I want to build an orchard, not just be a fruit-picker.
Systems are fun. Victory is not. It should just be the natural result of the successful creation of systems.