Scoring systems

[WarpZone32]

Well, after being taken to school in my own thread, it's become clear to me that my game needs a more interesting scoring system. (My game features set conditions for enemies to spawn and it is not possible for an enemy to escape, so raw numbers of enemies shot down are not a good way to go about it.) I've found some helpful hints, but no clean, definitive lists of scoring mechanics. So I decided to start one.

Raw Point Values:
- Each enemy shot down gives points. (Basic)
- For enemies that spawn other enemies, take your time killing them if the clone enemies give you points (i.E. milking)
- Dead enemies spawn loot. Collecting the loot gives you points. (Collectibles)
- Score bonus items spawn randomly and disappear quickly.
- Score bonus items spawn when you attack non-combat structures. (Medals)
- Score bonus items respawn in the same spot, so the faster you pick 'em up the faster a new one can appear.

Immediate modifiers (Triggered during level) :
- For enemies that fire at you, the closer you are to the enemy when they die, the more points you get. (For enemies that fire at you, this is risky.)
- The faster you kill an enemy, the more points you get. (Conflicts with milking.)
- Scraping (earning points for a near-miss.)
- Combos (For killing several enemies at once due to some sort of chain-reaction of game physics which the player only initiated (koopa shell chains in Super Mario Bros, for instance. Could be done with over-penetrating shots or domino effects of defeated enemies in side-scrollers.)
- Bullet cancel (points for using it when there are more enemies on-screen.) (Chained neatly with collectibles in Bangai-o!)

End-of-level Bonuses. (Given once for surviving with some secondary goal intact.)
- Bonus points for leftover health or lives. (Or charge a penalty to replace these.)
- Bonus points for unused shields, bullet cancels, or other one-time use items. (Same)
- In non-linear games, bonus for finishing the stage quickly. (Time Attack.)

Long-term modifiers :
- Killing several in a row gives you bonuses if you can keep doing it fast enough. (Chain.)
- Taking down enemies (or boss parts) in a certain order yields higher point values.
- Replay mode lets you keep playing after beating the last boss, with your old score intact.
- General grinding. (Has no place in most traditional shmups.)

Meta-scoring mechanics:
- Most overt scoring mechanics could probably be chained in some way.
- Securing one objective could trigger opportunities to try for other ones. (Enemies dropping loot is a simple example of this.)
- Some scoring mechanics are mutually-exclusive. This could be an annoying hindrance to peak performance by obfuscating the goal.
- OR, you could track and rank each objective separately, enabling the player to replay the game trying to master each of the different objectives, one by one.
- Different game modes could have different sets of scoring mechanics.
- You could give the player items which he can use put a scoring rule into effect.
- These items could appear randomly, be bought and equipped between levels, could be earned over time and used at the right moment, etc.

The Last Score:
- ...is always how many times in a row you can beat the game without taking a hit. Track this value, and nobody will ever truly beat your game all the way. (Until the integer wraps. (In which case, you could theoretically use a second integer to track THAT. (And change the colors or something if they ever actually pull this off. ( Cuz I mean damn. ) ) ) )

Caveats:
- If the score wrapping too easily is an issue, use a second int to track how many times it has wrapped. (Or just use linear bonuses instead of multipliers.)
- If the scoring system is too complex, most players won't be able to figure out how to improve. (Or won't bother.)
- If the game has no time limit, the player can afford to be careful.
- If the enemies can't escape, baseline scores for completing the level are static. (Hence the need for elaborate scoring systems in the first place.)
- Grinding indefinitely is only possible in a persistent, non-linear game. (And is seldom inherently desirable, without a shop or something.)
- Traditional high-score tables only make sense in an instance-based, linear game. (And work best when comparing different players.)

How'd I do? Feel free to add your own examples below. Please describe the scoring mechanic, don't just list some game titles. (I haven't played even a fraction of the shmups you guys probably have.)

[worstplayer]

Having separate score for each system sounds interesting. I don't know about any game that does that, so it may or may not work.

[WarpZone32]

Oh, I wasn't going to use all of these at once. This was just me brainstorming.

I'm gonna start with some score pickups that disappear, then maybe add a chain or a combo slop. Too much more advanced than that would probably be overkill, I think.

We'll see though.

[worstplayer]

Score pickups shouldn't be a problem, but time-based chaining won't work in randomly generated game.

[WarpZone32]

That makes sense.

Hmm.

Scrapes ought to still be viable, though. I'll look into that.

[Taylor]

I think most arena shooters are entirely about survival based (directly or otherwise) multipliers that balloon out of control and little else because they have tonnes of non-deterministic and indistinguishable waves, and most of their players are more casual and thus don’t really care. If you’re retrofitting a score system for an arena shooter I'd say you want simple global mechanics like proximity, survival time and shot types. And you want to reward players for scoring more immediately, even if its immaterial things like the game palette changing.

Games like DoDonPachi (chaining), Espgaluda (timed cancelling mode) and Progear (variable cancelling shots) have sculpted all their stages, enemies and patterns around their score system and, more often than not, expect the player to use it. Unless you plan on revisiting all aspects of your design I would stay away from anything intricate.

Edit: Confused you with another poster from the line about respawning enemies. If this isn't an arena shooter then ignore everything I've said: past, present and future.

[WarpZone32]

No, it's me. You confused the threads, but I'm the same poster.

[Aru-san]

I just created the dumbest scoring system ever for a 15-minute game.

Scoring System

Collect medals to raise your charge ammo.
- Charge ammo and GET POINT indicator goes up in 5s, 10s, and 20s.
- Charge ammo goes up to 1000 and GP indicator goes up to 2500 on the first charge. Excess medals after maxing out GP will be added to a separate counter.

Use your charge shot to raise your enemy score
- Activate your charge shot when it's maxed and raise your charge ammo capacity and GP limits.
- Charge ammo goes up by 100 increments, GP goes up by 500.
- Charge ammo peaks at 1500, GP peaks at 10000.
- MAX out your GP to add a multipler for more enemy score.

Enemy score for killing enemies with the charge shot:
(Private matters lol)

Point counters:
Enemy score - Levels up to 33. Raise the level every 10 million points. This score counterstops at 99,999,999 when at level 33.
Surplus medals - Levels up to 17. Raise the level every 10,000 medal points. This value counterstops at 99,999 when at level 17.
Timer - 15 minutes of play time. Endgame bonus multipler will be based on your longest survival time by the minute.
PERFECT STATE - If you're on your first life and counterstopped both enemy score and surplus medals, earn 1 point per 10 frames times kills made in that time frame. Second and last life will only add per enemy kill and not per 10 frames.

Endgame result calculation:
(Private matters lol)

[Drum]

Score pickups shouldn't be a problem, but time-based chaining won't work in randomly generated game.

It will if you can find some skill-based way to freeze your combo, like, say, the combo meter only depletes while you're shooting or moving - something like that. Maybe retaining a large amount of bombs makes the meter go down slower. There are lots of ways you could pull it off. If the design isn't robust enough to handle random generation, it's not much of a design. imo.