WIP shmup - criticism requested

[FailFaster]

Update: Sorry, link was broken. Should be fixed now.

You can play a recent web build here: http://failfaster.itch.io/wip-shmup-demo

So I'm working on a simple top-down shmup for Android and the web. At the moment, it's nothing special. Just five enemy types, one collectible currency that doesn't really do anything yet, and four enemy shot patterns. Consider this the soup stock that a decent game might one day arise from.

Please use this thread to let me know what you like and don't like. Also, I intend to start referencing the stickied threads in the Development section as I work on it. They're good stuff!

My name is a reference to a particular Extra Credits episode. They call "Fail Faster" the mantra of creative professionals. I've tinkered with this project for long enough. It's time for people to start breaking it so I can rebuild it, better and stronger.

Thanks!

[Ixmucane2]

The deliberately lagging mouse controls are, to put it nicely, very unpleasant.
If you are primarily targeting touchscreen input, mouse input can be forgiven, but only if movement is instantaneous. As a rule of thumb, frustration is directly proportional to inertia, sometimes worse.

Lesser areas of improvement:

• Getting pickup items from harmless asteroids but not from enemies appears backwards.
• Repetitive bullet patterns of very little aesthetic or tactical value (although not really defective). Less random enemy waves and formations would probably make bullets look vastly better.

Graphics are very good, but you should experiment with varied bullet sizes and styles.

[FailFaster]

Hmmm. Thanks, Ixmucane2. I'll see about disabling the speed limits for mouse controls next build. The current build allows you to adjust mouse sensitivity up to double the actual input speed, so maybe that'll help a little bit. Give it a try at http://failfaster.itch.io/wip-shmup-demo .

The speed is still clamped to something like 10 world units per update, but I can disable that. I was worried that using the raw mouse movement values would make dodging too fast and easy, but I'll try it your way for the next build and see what happens.

You're totally right about the pickups. Mining asteroids may technically be realistic, but screw realism. Getting better rewards from the more powerful enemies is an acceptable break from reality.

Can you clarify whether you mean the player bullets or enemy bullets? Also, what do you mean by "look better?" Are you talking about the aesthetics of each individual shot or the way the group of bullets looks? (Fair warning, if the latter: this isn't a danmaku! It'll probably never be wearing the prettiest dress at the bullet pattern ball. That said, if you have specific suggestions for shot patterns, I'm open to hearing them.)

[FailFaster]

The deliberately lagging mouse controls are, to put it nicely, very unpleasant.
If you are primarily targeting touchscreen input, mouse input can be forgiven, but only if movement is instantaneous. As a rule of thumb, frustration is directly proportional to inertia, sometimes worse.

Got the big one taken care of, at least! The current build now features completely uninhibited (as far as I know) mouse movement. If you use a laptop's builtin mousepad, you may want to adjust the sensitivity in the options menu.

Thanks again!

[Ixmucane2]

I tried the updated version, controls are as good as mouse controls can be (much less accurate than keyboard input) and the option to change sensitivity is a nice touch.
On the bullet patterns front, I remember the previous demo being the same or possibly slightly better. Games like the Raiden series have enemies which individually shoot more or less according to the same patterns as yours, but with two substantial differences:

• Enemies come in formations and they shoot in ways that make sense for the whole formation, producing aggregate bullet patterns that look coherent and structured.
• Different enemy types have substantially different shots, offering the level designer a large palette.

What's up with the bizarre player shots?

[FailFaster]

Basically it's the first phase of testing for a planned upgrade system. You'll eventually be able to buy and sell weapons in a shop menu. Right now, the weapon code is mostly there, but the shop menus aren't. So it just gives you three random weapons each time you start the game, as a test case.

I could implement arrow controls if you want, but I don't know if they'd allow for the kind of quick movements you can get from a mouse/touch interface. My concern was that playtesting with arrow keys would result in balancing the game in such a way that it's completely unbalanced on mobile. What do you think?

[Ixmucane2]

But "quick movements" are a problem, not a feature: any large and fast movement can fatally cross a bullet or an obstacle before the player can react.
Keyboard and joystick directional controls are good because of reproducible input: the exact same movement for the same fairly loose button press timing, offering easy to learn discrete choices of dependable movement options which allow very precise movement (e.g. aligning the ship with a narrow gap).
Touch or mouse input is less dependable because for practical purposes it is analog, with significant uncertainty about the outcome.
The least bad case for touch screen input is moving quite slowly, with no incentive to make large movements, so that the position of the finger is the same as the position of the player's craft and some predictability is restored; but even then obstacles need to be more forgiving.

[FailFaster]

Well, I've got a basic shop done, including an in-game preview for the firing pattern of any weapon you highlight. When nothing is highlighted, it shows your ship firing all the currently equipped weapons. This is extremely unbalanced right now, but it'll be a little better once I get stage difficulties implemented. (And for that matter stage endings, a stage select, bosses, etc.)

Also in this release, a big feature for PC users: Arrow Key Movement! I disabled half the movement code to get rid of that pesky smoothness. I can feel how much tighter and snappier the control is now, but YMMV with your local frame-rates.

The Control Sensitivity slider now affects arrow keys, so play with it to see what feels right for you. (Options screen opens automatically when you pause now, BTW.) 25% worked well for me, but 50% probably would have been fine if it weren't for the split second it takes one of these big plastic keys to depress. YMMV probably here more than anywhere else, but I hope this is a step towards better controls on PC.

As always, check out the new build at http://failfaster.itch.io/wip-shmup-demo .

[Elrinth]

Good music, but annoying loop.
The ship control is waaaay too fast for me with the arrow keys. Maybe analog stick on a controller would be sweet for this!?
You've got something going. I like it! Keep up the great work.. When it's done I wanna see it up there with Ikaruga.

[beatsgo]

-To be frank, I think for now stick with mouse movement since the ship movement is way too fast. As a matter of fact, I think you're mixing "sensitivity" and keyboard/digital controls too much. Keyboard/Digital controls should always have slower directional movement since a human can't react to that type of speeds (unless you're going for a UO POKO-like ship in SOS Fever). Also, that ship speed is wayyyy tooooo fast. Slowdown the ship and increase both enemy bullets and player's bullets.

-What style of shmup enemy health are you going for? Euro? Japanese? I'm not a fan of 6-hit popcorn ship unless that's what you're looking for.

-I think a game you should look into is Noiz2sa, great game from ABA games with good pace and feel to it. There's even a iOS release of it if you're interested, you'll understand why with that type of ship speeds mouse is the way to go for controls.

[arcadehero]

I've given the build some plays including when it was all mouse control. I personally like the mouse control a lot more although the difficulty was easy enough in that build that I made it somewhere around 33km.

Pros and Cons:

Pros:

-I like the randomization to the waves
-Nice graphics / no confusion over the shots
-store is starting to look promising

Cons:

I liked the mouse controls from a previous build, although I suppose it was too easy. On the current keyboard it feels off, I end up dying a lot more than the other builds because the ship is floating too much.

I do get that for Android you need the autofire but in my view I love having control over when my shots go out, the current standard blaster has the timing off from where it would be nice

Not sure if this build has it or not since I have not made it any farther than 900 but different varieties of enemies would be cool including more fodder types that just need one hit to shatter into pieces.

Overall I think it is promising, keep up the good work.

[Ixmucane2]

The default speed setting in this keyboard version is insanely fast; I set "sensitivity" to 29% as a rough compromise between accuracy and being able to sweep the playfield.

Regarding the upgrade design, the player starts far too weak (the basic initial enemies should die with one hit, mostly before shooting, and without suicide bullets!) but rapidly accumulates stupidly excessive firepower; there is only a brief window of decent challenge between catching one coin in a very good run and obliterating everything neglecting coins because they cannot be spent.
This is, of course, a byproduct of the flat challenge level of the single randomized stage; the player should face weak enemies when weak and strong enemies when strong, with later levels requiring the firepower that the previous levels have offered.

[FailFaster]

There's a new build up at http://failfaster.itch.io/wip-shmup-demo . The main good news is you can pick between mouse and arrow movements in the options menu, movement speed when using arrow keys is now much slower, the stages have an end now, there is a stage select menu now so you can grind an easier stage if you want to, and the shop now sells weapons more or less appropriate to the highest stage you've unlocked.

No bosses yet, sorry. It's still not what I would call perfectly balanced, but hopefully it's a bit more playable than it was before.

Let me know if it still needs core improvements. Until then, I'll be working on adding content.

[tiaoferreira]

I like this game.

Simple at all.

[Ixmucane2]

there is a stage select menu now so you can grind an easier stage if you want to, and the shop now sells weapons more or less appropriate to the highest stage you've unlocked.

It's taking shape, but now it's too easy.

Due to excessive treasure drops, the difficulty knee I mentioned during in my previous post occurs during the first halfway decent run of mission 1 (which suffices to fill all weapon slots).

Weapons should be more balanced with equally costed ones (the current completely random creation is bad) and generally much more expensive: keep the large number of small treasure drops, which is a good feature because many of them cannot or should not be caught, and inflate prices accordingly.

Cheap shots, particularly sideways from almost gone enemies, are a serious problem which would improve with better designed and less random formations.

Technical problem: equipping many Blaster type weapons vastly exceeds the maximum bullet count, causing old bullets to disappear and reducing actual weapon range to about half the screen height. Reducing firepower by adding weapons is highly inappropriate, particularly because the problem might go unnoticed for a while and the solution requires selling off weapons.

Exploit: one can buy all good weapons, exit the upgrade screen, and come back to a completely refreshed upgrade store with a new batch of weapons to cherry-pick.

[FailFaster]

Tweaked the control options a little bit. Now instead of choosing mouse or keyboard steering, keyboard steering is always on by default, and you can just enable or disable the mouse. This also means you can switch between them arbitrarily while playing if you feel like it.

Weapons now cost more in general, and the cost is always identical. (I may change this later, but only if the looting system becomes much more sophisticated.)

I'm experimenting with different enemy formations. More coming soon hopefully.

Slight graphical improvements. (Yes, I know I shouldn't be polishing right now, but some days I just can't help myself.)

Check out the current build at http://failfaster.itch.io/wip-shmup-demo

[CloudyMusic]

I see a lot of potential here. My first impressions, forgive me if you've already addressed some of these things in this thread:

- Mouse/touch movement feels great, though I'm still not sure about keyboard movement. Maybe it just feels foreign to use directional inputs but have no fire button to push. It seems like I can just blanket the whole screen with fire by shaking the mouse back and forth really fast, producing a sort of poor man's wide shot. Not sure if that's a problem, though.

- I like the background graphics. Whatever graphical effects you're applying to the ships might be kinda lost at those small sizes, though.

- The missions feel really really short, in general, and they obviously end really abruptly (which I'm sure isn't final).

- How much enemy variety is there? I think I've only seen two enemy types, unless I've seen more and they're just similar enough that I just haven't noticed. Most shmups have a balance between small, weak enemies that are somewhat harmless and big, tough ones that shoot a lot of bullets and take a lot of hits. Really exaggerating the enemy variety can keep things interesting.

- Related to the above, if you have tougher enemies, perhaps make them drop more crystals, or at least have bigger explosions. It feels good to see a big payoff when you kill something tough.

- Any plans to implement any sort of in-mission item drops or powerups from enemies, aside from the shop currency? Maybe that could make the moment-to-moment gameplay a little more exciting.

- Do you intend that the player basically has to buy upgrades to be able to play the later stages? I think those enemies eventually become literally impossible to kill with the default weapon. That could be frustrating for some people if you don't communicate that in some way.

- In the shop, I ended up buying both a 3-way spread shot and a 5-way spread shot, which seemed a little redundant. If you're going to have a bunch of different weapons, make sure they're all unique, interesting, or memorable in some way. (I haven't seen the vast majority of them, so maybe this is already the case.)

[FailFaster]

Thanks, Keres.

Keyboard movement was added late in development specifically to cater to PC gamers, so if there's anything concrete you can tell me (such as it needs to start and stop on a dime, or the movement speed needs to be 25% faster, or diagonals don't work the way you expected, etc.) that would be the most helpful feedback I could get.

Mission length isn't final, they're just short to allow playtesting of as many levels as possible, as quickly as possible. Mission endings will definitely need to be less abrupt. I haven't decided quite how to code that yet.

Right now there are two enemy models (ship and rock,) and 5 flavors of ship that all use the same model. I'll replace most of these later when I have the time to focus on art assets. I should probably at least change the colors in the meantime. I'll see if I can get on that soon.

Greater enemy variety is planned. Bigger enemy = more health = bigger explosion = more loot. Bosses someday. Rocks that split into other rocks. (Also, I should move the rocks to the space stages, come to think of it.)

I haven't given much thought to in-mission powerups. The loot system is obviously the star of the show, so anything that makes your choices in the shop less significant could detract from the game's main draw. I'll think about it.

Yes, the pea shooter (and all the weapons, for that matter,) will eventually become obsolete, by design. Think Diablo or Borderlands. My expectation is that the players will pretty much figure it out when they keep dying and see a button that says "upgrades." I'm still working on the menu flow between missions, so maybe I could set it up so you always visit the shop when starting a mission, even if "continue" was what you clicked.

I'm thinking about staggering the timing on your equipped weapons so they don't all fire at the same time. Even as it stands, they should eventually get out of synch and back again if you fire for long enough. That's just a band-aid, of course. Later on I'm planning to specialize the equipment slots to accept primary, secondary and tactical weapons. But I need more weapon types and variety before then. (Oh, and if you want to see redundancy... it gives you a free pea shooter if you load the game with no equipped weapons. Try exploiting that. )

[Ixmucane2]

I begin to be happy with keyboard controls, but more challenging levels are needed for a proper evaluation.

Neither of the bugs I found in the previous demo, the bullet count issue with Blaster weapons and the store refresh exploit, have been addressed. Although not terrible issues from the player's point of view, I think they prevent tuning weapon designs and prices adequately.

The new exponential weapon price increase doesn't seem balanced: money per level remains roughly constant (constant number of enemies and constant treasure for enemy), so buying "advanced" weapons would require an enormous amount of grinding. At level 11 of my test game, firepower became insufficient and I gave up.
There should be the right amount of treasure to afford the weapons needed to beat the next level; due to obvious positive feedback between getting more treasure, buying better weapons, killing more enemies and getting more treasure some games will be very good and some very bad, but it's OK because it's only an amplification of player skill and you can always limit the effect with artificial negative feedback (weapon gifts if players are behind the firepower curve, rank increase in ways that reduce treasure if players are doing too well).

The exponential price increases are also intrinsically unfair, irrespective of available money, because they should correspond to an obvious exponential increase of firepower (particularly in your case, where enemy health also increases exponentially and needs to be matched). This isn't the case: weapon power appears to be constant, since bullet damage isn't apparent, bullet count doesn't appear to increase, and random weapons are often obviously worse than older and far cheaper weapons.
I suggest rigorously matching prices to weapon power, with weak weapons costing very little and great weapons inducing the player to sell off everything else to buy them.

Regarding firepower, Blaster weapons are ridiculously better than anything else by virtue of shooting more bullets.

[FailFaster]

There's a new version in the works, using exponential increases for damage, health, and money drops. Check it out at http://failfaster.itch.io/wip-shmup-demo . It scales nicely up until about level 24, whereupon the shop prices start coming in scientific notation and it becomes harder to understand what you can and can't afford.

I think I may need to completely redesign how damage and prices are calculated, but I haven't decided how yet. Any thoughts?

Ixmucane2 , I increased the pool size for player shots. I don't know what specific weapons you have equipped but I hope this helps in the short term. The shop-restocking exploit isn't fixed yet, precisely because it's useful for testing edge-case weapon combinations. Probably long-term, I should make it possible for the player to choose an earlier shop in addition to an earlier stage.

Oh, and on the Options menu, there's a new bright red "erase all your progress" button. I didn't put in an "Are you sure" submenu yet, so PLEASE be careful with this.

[Ixmucane2]

Scaling exponentially damage, treasure, prices and enemy life is completely pointless: the respective ratios remain constant and there is no actual change from level to level.

Traditionally, levels are progressively tougher (with stronger enemies and qualitative changes, not with increased hit points for the "same" enemies), with variations and lulls, ending with some grand final boss. Identical levels are boring.

As a consequence, firepower should increase to allow killing enough enemies, with "enough" steadily increasing. In a powerup-based game, you would simply give the right upgrades at the right time, with some adjustments to forgive missed powerups, reward good powerup collection, and allow a comeback after dying. In a treasure-based game like this one the needed upgrades are the same, they are merely shifted to the store between levels: you need to control the expected amount of treasure in each level (presumably increasing to fit the more "valuable" enemies) and price weapons sensibly. There's enough elasticity between treasure drops and weapon prices to balance weapon prices against each other and still provide the right amount of treasure to buy the right amount of firepower.

Example: suppose the initial complimentary peashooter shoots 10 bullets per second, each doing 1 damage. Level 1 has 600 type A popcorn enemies with 1 hit point, 200 type B weak enemies with 4 hit points, 100 type C light enemies with 8 hit points and 10 type D tough enemies with 40 hit points. The player needs to deal 600+800+800+400=2600 hit points of damage to kill everything, certainly a lot (hitting with every single bullet for 4'20''). Let's expect a typical performance to kill half the enemies. How much treasure should enemy types A, B, C and D pay out? It depends on level 2 content. If level 2 has 400 A, 400 B, 150 C and 20 D the base needed damage is 400+1600+1200+800=4000 damage. Firepower should increase to about 15.3 1 damage bullets per second, say 16 to be nice.
Tentatively assigning no treasure to type A, 1 to type B, 2 to type C and 10 to type D, a 6 dps weapon (added to the 10 dps peashooter) should cost about 100+100+50=250 treasure points; it could be rounded to 40 treasure points for 1 dps. This also mean that the player would earn about 200+150+100=450 treasure points in level 2, which translates to slightly over 11 dps increase, which implies a serious toughening of enemies in level 3.

[FailFaster]

I see my problem. I was thinking about it like a JRPG or western RPG, or even some action-rpgs like Borderlands, where later levels need to continually raise the bar on how much money is required to buy the next upgrade so there isn't as much of an incentive to just grind the easiest level over and over again. Advancing to the next area means you're fighting harder enemies, but it also means better potential loot.

(In other words, in order to make Area 2 more appealing to the player than Area 1, XP and treasure rewards in Area 2 need to be significantly higher than any previous Area.)

What I was forgetting (probably because I haven't implemented it yet) is the fact that seeing the next area, seeing the new enemies, and mastering the new patterns is a big part of the draw in shooters. One area naturally follows the next, and while backtracking might be allowed, it's not the player's first instinct, especially if beating one area automatically advances you to the next.

The status quo in a shooter favors momentum, rather than inertia. Quite unlike in an open world game where the player has to take action in order to progress through an area.

Anyway. Thanks for the insight. I'll start thinking in terms of sustained DPS VS everything on the screen, rather than health-per-enemy.

[Ixmucane2]

Keep in mind that dps rates, total hit points, expected treasure and projected weapon needs are a coarse tool, and assuming certain average fractions of missed enemies and wasted shots is a coarse assumption; you need to script the levels in detail and design one wave after the other.
Usually, the standard for level flow is that a good player with good weapons and possibly some memorization is barely able to kill everything, with slight mistakes and insufficient firepower allowing some enemies to escape.

[Legendre]

The movement speed is way too fast--I am unable to precisely navigate the screen. Looks nice though, your attack pattern design seems to be on the right track. I wouldn't mind seeing more down-aimed enemy bullets in the early levels.