Hitboxes: lazy/bad programming or best idea ever?

[Erppo]

People are talking about scale but there is an other, possibly more obvious argument about why small hitbox is different. When your hitbox is tiny, you only need to do a single tap to dodge an aimed bullet. When your hitbox is large, you need to make a larger movement. This is an essential difference in gameplay. If you wanted to replicate the same style with a large hitbox, the only solution would be insane ship speeds which obviously leads to other problems.

[KidQuaalude]

Yes. Good point.

[gray117]

That wouldn't look as cool though.

^ This, the game looks crazy and prompts curiosity and attracts plays. Once you realize what is going on you're probably having enough fun with the play that you're perfectly happy with it.

I think realism/realistic is not exactly what author meant since there's a lot about the typical shmup that is unrealistic ...but perhaps rather it seems unbelievable/illogical to your visual conception of the game and it takes you out of it somewhat? ... I can sympathize with this, and I feel there might be a middle ground between the cave style hitbox and energy bars or hits of varying damage [direct hit vs. glancing hit on wing etc.]. I think it would be interesting to see this kind of done (brought back?) by someone.

But I wouldn't necessarily criticize the small hitbox games for their mechanics or chalk it up to bad programming... Perhaps if I were being uncharitable I'd mark it up to cautious/lazy design?

I suppose design wise the thing you get with a small hitbox is an obvious near miss feeling, combined with no break in hectic play AND combined with a continuous 'all or nothing' feeling? It's kind of the reverse of a typical fighting game, especially when you can get a KO after a tap to the shins

[DarkSlayer]

defiantly not lazy, it is a design decision
and a technical decision: perfect collision checking is expensive on the hardware (especially more years ago)

so my guess is they make the hitbox a square within the sprite for performance reasons then saw the advantages to gameplay that a different sized hitbox can bring.

[Obiwanshinobi]

Yeah, have you ever though about how, when you get to boss screens where the screen is no longer scrolling, how the hell does the ship move backwards if the thrusters are on the back?

Raiden took care of it. The flame gets blue, indicating that some Invention™ is at work there. Future and all that.
Small hitboxes get on my nerves only when bullets are slow, as if somebody tried to stop me with rose petals. Not a problem in Dangun Feveron.
Calling typical bullet hell shmups "manic" sounds all wrong to me as well.

[Thorham]

and a technical decision: perfect collision checking is expensive on the hardware (especially more years ago)

Old systems (Commodore 64 old!) often have hardware based sprites with hardware based collision detection, making it essentially free. Even when you do the detection with the CPU it's not expensive on old systems (Amigas and later). On todays machines it's a piece of piss.

[louisg]

and a technical decision: perfect collision checking is expensive on the hardware (especially more years ago)

Old systems (Commodore 64 old!) often have hardware based sprites with hardware based collision detection, making it essentially free. Even when you do the detection with the CPU it's not expensive on old systems (Amigas and later). On todays machines it's a piece of piss.

Yeah, I believe even the Atari 2600 had pixel-perfect collision detection in hardware. IMO it can be somewhat less desirable than more selective collision styles, even for games with sparser bullet patterns, for the reasons already given.

Though, I don't think many shmups are doing any kind of crazy interpolations for the hit detection, either, are they? I saw someone brought that up. It probably would be reasonable on fast hardware if everything is a hit-point and not a hit-box (line intersect algorithm), but once you get into more complex shapes, AFAIK it gets very complex very fast.

[Drake]

Small hitboxes get on my nerves only when bullets are slow, as if somebody tried to stop me with rose petals. Not a problem in Dangun Feveron.
Calling typical bullet hell shmups "manic" sounds all wrong to me as well.

I would call bullet hells and manic shooters two different categories with distinct gameplay styles. It seems oddly common to equate the two but it doesn't make sense at all. They can overlap, but I definitely would call them separate entities.
That being said, would you rather a larger hitbox where there are slower bullets and likely more of them all over? Such a concern seems similar to OP's problem; that there's a sense of immersion breaking with an "unrealistic" aspect of the game. I find these sorts of attitudes interesting because I don't have that sense at all (unless the game is praised for something I find fault with obv).

This is generally why I find OP's complaint absurd. I can understand not being able to consciously see the design aspects that go into hitboxes and other such decisions, but I find the claim that it's bad or lazy programming completely insane, and so far there hasn't been any legitimate criticisms, just the equivalent of "I don't get why it can't be larger therefore it's bad". Surely you should still be able to tell the difference between hitbox and shipbox and why the former would be necessary simply by playing many different types of games?

Because it seems that it's not exactly a set-in-stone science, doesn't it?

The thing is that it is a set-in-stone science, in that the reasons for hitboxes are mostly technical and are objectively required in order to fit with certain styles of games and play. From the way I'm reading your arguments, you are the one attempting to appeal to some subjective aesthetic; in other words I find that you're attempting to bring the other side of the argument to the level of your own.

[boagman]

Because it seems that it's not exactly a set-in-stone science, doesn't it?

The thing is that it is a set-in-stone science, in that the reasons for hitboxes are mostly technical and are objectively required in order to fit with certain styles of games and play. From the way I'm reading your arguments, you are the one attempting to appeal to some subjective aesthetic; in other words I find that you're attempting to bring the other side of the argument to the level of your own.

Then I'll rephrase, if you really want to parse it out: if it's set-in-stone, as you say that it is, then why isn't it therefore one-size-fits-all? By your definition the technical aspects call for the method...then why doesn't it follow that it should be that way ere the sun rises? Why aren't *all* hitboxes one stupid pixel in size? You paint me as plodding, but that ain't exactly the truth, now is it?

I guess it sort of depends on where you're willing to settle, especially in terms of aesthetics. I find the argument "because it looks cool" to be an argumentum contra se...looks at the expense of play, or at the very least, logic. If my shots hit a plebian enemy ship on the tip of his wing and destroys it, shouldn't the inverse also be true? And if not true, shouldn't the player darned well *know* where the line is drawn?

Actually, to borrow your phrasing, I'm not attempting to bring the other side of the argument to my level...I'm attempting to bring the other side of the argument into congruence with what it claims to be. I'm not seeing that as of yet. I get at least part of what's being posited (there's room for more than one style of shmup, yes), but the original focus of my query hasn't really been addressed to my satisfaction.

Perhaps that just isn't going to be possible.

[Jeneki]

I wonder how games with grazing mechanics would work with a full-sprite hitbox. The feel of glancing blows triggering effects would be lost if they trigger when a shot is just sort of near you.

[Thorham]

If my shots hit a plebian enemy ship on the tip of his wing and destroys it, shouldn't the inverse also be true?

Of course not, because it's a game, and the designers can do anything they want with it.

[Udderdude]

Hixboxes suck, hitcircles are the future :3

[Muchi Muchi Spork]

I've always looked at it as making at least some sense because the center hitbox part of the ship is the thickest part of the ship so the most likely to be hit and the skinny little wings could be maneuvered more easily around bullets by the skilled pilot. You guys lack imagination and now I'm going to send pink bullets toward your hitboxes for it. FIVERAWN FIVERAWN

[ptoing]

I highly doubt there is ANY arcade shmup with an exact pixelperfect hitbox, mainly because this is quite a bit more processing intensive as opposed to a square/rectangle which can be defined by 2 points. Also whatever you do, you will always have some form of "hitbox".

[Thorham]

Hixboxes suck, hitcircles are the future :3

Easy, use circle tables

[Squire Grooktook]

Why aren't *all* hitboxes one stupid pixel in size?

I already answered this like 8 times. It's because the size of your hitbox effects the fundamental way you dodge, as well as what you can dodge, thus resulting in a completely different gameplay style. No matter how hard you try, you'll never be able to recreate the gameplay style of Dodonpachi with a Gradius style hitbox. On the flip side, Gradius with a Dodonpachi sized hitbox would have to make use of completely different threats to challenge the player, and would also have a fundamentally different feel and play style.

One play style isn't necessarily better then the other. Small hitboxes and big hitboxes are both fine. Whatever floats your boat. This is why we have different sub genre's.

...but perhaps rather it seems unbelievable/illogical to your visual conception of the game and it takes you out of it somewhat? ... I can sympathize with this, and I feel there might be a middle ground between the cave style hitbox and energy bars or hits of varying damage [direct hit vs. glancing hit on wing etc.]. I think it would be interesting to see this kind of done (brought back?) by someone.

The thing is, shmups are not space ship games. The players craft could be a space ship, or a loli, or a dragon, or a super saiyan, or a two dimensional representation of the human soul. What it REALLY is, is an avatar of the player. So designing a game based on the realism of piloting an aircraft is similiar to inertia: It hurts what the genre is really about.

But I wouldn't necessarily criticize the small hitbox games for their mechanics or chalk it up to bad programming... Perhaps if I were being uncharitable I'd mark it up to cautious/lazy design?

Cautious or lazy are the last two words I'd describe it with. A lot of work goes into designing these games, and cautious? Really, a pixel sized hitbox must have been a really brave and unorthodox design choice when it was first introduced.

[shmuppyLove]

Hixboxes suck, hitcircles are the future :3

2D is old-hat bro, hitspheres are where it's at.

[Drake]

I wonder how games with grazing mechanics would work with a full-sprite hitbox. The feel of glancing blows triggering effects would be lost if they trigger when a shot is just sort of near you.

You wouldn't be able to set up a coherent system. The larger the hitbox is, the fewer bullets you're able to purposefully concentrate around the player, and the fewer bullets you can set up in the space around the player. You can't use streaming techniques for graze since you can't easily dodge streams of perfectly-aimed bullets, which makes the whole notion of streaming blocky and might it as well not exist. The closest you could get is letting the player sit near patterns they acknowledge won't kill them.
Really you should be able to get an idea just by playing a game with a shipbox and trying to graze (with no incentive other than imagined points).

I highly doubt there is ANY arcade shmup with an exact pixelperfect hitbox, mainly because this is quite a bit more processing intensive as opposed to a square/rectangle which can be defined by 2 points. Also whatever you do, you will always have some form of "hitbox".

But a single pixel is just a 1x1 rectangle in game logic, and you can still get more precise than that.

Then I'll rephrase, if you really want to parse it out: if it's set-in-stone, as you say that it is, then why isn't it therefore one-size-fits-all? By your definition the technical aspects call for the method...then why doesn't it follow that it should be that way ere the sun rises? Why aren't *all* hitboxes one stupid pixel in size?

Set-in-stone methodology. Hitbox size depends on the game and its playstyle; what the designer intends to make you do when playing is everything. It naturally scales depending on what surrounds the player, or rather, the general hitbox size is determined as a function of the play scenario. If you play a lot of shmups, it should become very obvious when the hitbox size is chosen poorly, and after a long enough time players can even begin to tell if a hitbox is a pixel bigger than it should be. Finding a lower bound on hitbox size is more difficult, but follows the same rules.
I see where you're getting at with different-sized bullets, since clearly some bullets are larger than others, but allowing slack is still important. Just because you have smaller hitbox sizes than the whole sprite doesn't mean the natural conclusion is that everything should be as small as possible. You have to find a balance in your constraints.

I guess it sort of depends on where you're willing to settle, especially in terms of aesthetics. I find the argument "because it looks cool" to be an argumentum contra se...looks at the expense of play, or at the very least, logic. If my shots hit a plebian enemy ship on the tip of his wing and destroys it, shouldn't the inverse also be true? And if not true, shouldn't the player darned well *know* where the line is drawn?

I'm of the opinion that aesthetic should never diminish gameplay. Intrude on gameplay and integrate with gameplay definitely, but you should never give up gameplay over aesthetic. As for the question: no. The player should darned well know where the line is drawn because people are not idiots. This is something you don't, or shouldn't, even consciously process when playing. You learn these things by playing the game, and even if you don't know exact measurements, you should be able to easily tell things like, say, where your hitbox is.

What, exactly, do you think hitboxes claim to do? What problem do you think they claim to solve? Your opening post only has your own qualms, there is no said claim besides "the best idea ever", which is definitely not the case. To that extent, you don't seem to have an "original focus of your query". Are you sure you aren't simply attempting to bring the opposing argument to a strawman formed by your viewpoints?

[Ed Oscuro]

I wonder how games with grazing mechanics would work with a full-sprite hitbox.

Just as the internet is a bundle of tubes, super hitbox is a bundle of smaller hitboxes. Not exactly boxes within boxes, but it evokes the charismatic feel of a Eagle 5 flying Winnebago.

[ED-057]

that the hitbox should be the scale of the ship. But then either the ship is huge and takes up too much of the screen to be fun (and there are plenty such terrible shmups out there) or you want everything scaled down until your ship is literally a tiny dot and you're squinting to differentiate it from the enemy bullets.

I'm not clear on how it is that learning where your one-pixel hitbox is on your sprite and then navigating it through similarly-small gaps in bullet patterns is doable, while at the same time having the actual object graphics be reduced to such a small proportion of screen area would be too hard to see.

If sprites from classic low-res shmups were used to make a shmup that ran at native resolution on a modern HDTV or PC monitor then the number of objects fitting on the screen would be vastly increased. Would that permit implementing "the bullet hell experience" or not?

[Ed Oscuro]

I wonder how games with grazing mechanics would work with a full-sprite hitbox. The feel of glancing blows triggering effects would be lost if they trigger when a shot is just sort of near you.

You wouldn't be able to set up a coherent system.

Sure you could. Grazing can (and I assume does) use nested hitboxes. "Full sprite hitboxes" are only conceptually different in that they'd be more precise and vastly slower. Grouped boxes or spheres can be used to similar effect, as a compromise between a single fast but visually unintuitive box or radius, and pixel-perfect accuracy.

If sprites from classic low-res shmups were used to make a shmup that ran at native resolution on a modern HDTV or PC monitor then the number of objects fitting on the screen would be vastly increased. Would that permit implementing "the bullet hell experience" or not?

I always thought the Atomiswave Metal Slug was odd looking.

[gray117]

But I wouldn't necessarily criticize the small hitbox games for their mechanics or chalk it up to bad programming... Perhaps if I were being uncharitable I'd mark it up to cautious/lazy design?

Cautious or lazy are the last two words I'd describe it with. A lot of work goes into designing these games, and cautious? Really, a pixel sized hitbox must have been a really brave and unorthodox design choice when it was first introduced.

Sorry you've got me slightly out of context - the small hit box is a fine piece of design.The issue going forward is perhaps whether going forward that is all we see from design - if I was to be uncharitable you could chalk the lack of variation on the design side to be over cautious/lazy - simply sticking to re-trodden ground, or what was successful last [for someone else] last time. Of course such decisions are rarely so cut and dry - but it I think would be interesting to see the hitbox premise kind of re-explored in some manner.

Similarly, although the disciplines are often quite entwined, I think it's pertinent to identify the hitbox size as a design choice rather than a programming limitation.

[Ed Oscuro]

I think it's pertinent to identify the hitbox size as a design choice rather than a programming limitation.

Well, nowadays you have that luxury. Limitations drive design, though.

[Obiwanshinobi]

Just realised why Gradius V hitbox doesn't bother me: projectiles don't look like cherry, plum etc. blossom petals and - the game being a vertical scroller - don't look like snowflakes falling either.
Coupled with the kind of haze I like in my games (that gives me the impression of remoteness), those nuances allow me to willingly suspend my disbelief.

[KitesAreFun]

How about "It isn't necessary"?

Well that depends doesn't it? I think the smaller hitbox is a design choice that catered to another design choice. Probably as games became more dense/manic/crazy - someone, somewhere, decided that having a largish block on screen as a hitbox was making what they wanted to make, impossible. So to make the impossible possible, they made the hitbox small so they could make the game they wanted. Some games play the way they do contingent on that small hitbox.

Of course you could argue that they could do just as well by speeding up the bullets, or enlarging the patterns etc. But then you'd have two different games, and they might be equal in difficulty (contingent on many factors) but they certainly would PLAY a lot differently, no?

In my opinion, and as a practical matter, I think the small hitbox with a larger visual cue was, and still is, a fundamentally good design choice, which is probably why it's still here.

Edit:

Looks like I may have replied too soon in the post.

[KitesAreFun]

People are talking about scale but there is an other, possibly more obvious argument about why small hitbox is different. When your hitbox is tiny, you only need to do a single tap to dodge an aimed bullet. When your hitbox is large, you need to make a larger movement. This is an essential difference in gameplay. If you wanted to replicate the same style with a large hitbox, the only solution would be insane ship speeds which obviously leads to other problems.

Agree.

[Obiwanshinobi]

Rationale for bullet hell as we know it must be strictly economic. Bear in mind that the subgenre originated when arcades fell into decline and so did graphics of coin-op games. DP-DDP-DOJ cycle illustrates that deterioration. To keep shmups showy (arcade games need such allure) cheaply, bullet hell was born.
You know all too well that after Einhänder, Radiant Silvergun and a couple of ST-V shmups there was nothing nearly as elaborate graphically (for economic reasons, no less). Look at Dragon Blaze, then look at Mahou Daisakusen and guess which one's graphics took more effort.

[boagman]

I don't want to come off as a chicken here for starting a conversation and then not replying to anything for a couple of days. Suffice it to say that I'm probably a bit too dense to quite grasp a hold of why there needs to be a large graphical ship with a small hitbox in just about any place. I guess to me, if you're going to do the craziness of bullet hells and what not, the ship should be relatively to scale in terms of the stuff it has to avoid. But that's just me.

Anyway, the discussion itself seems to have taken a nice direction all on its own, and far be it from me to quash that in any way shape or form. I'm sure that there's room for multiple viewpoints, just as there's room for multiple types of shmups. We may like what we like, but I can respect the need for something different in the gray areas, I guess.

I still support the correct scale of hitbox-to-ship-size, but maybe that's just a hurdle I can't get over due to my tastes or abilities.

[Drake]

You wouldn't be able to set up a coherent system.

Sure you could. Grazing can (and I assume does) use nested hitboxes. "Full sprite hitboxes" are only conceptually different in that they'd be more precise and vastly slower. Grouped boxes or spheres can be used to similar effect, as a compromise between a single fast but visually unintuitive box or radius, and pixel-perfect accuracy.

I'm not saying that you couldn't make up a system technically, just that it wouldn't be very good design-wise, by virtue of what grazing is supposed to accomplish. The very point of grazing is both to reward the player for playing a bit dangerously by moving close to dense groups of bullets (or taking advantage of pattern design), and to evoke a feeling of satisfaction from said hazardous play. You wouldn't be able to do that effectively with a shipbox for a few reasons. Immediately, the proximity and density restrictions that come with a larger hitbox does limit what you can throw at the player without your pattern turning to garbage. As Erppo mentioned, there's a near-removal of grazing by streaming aimed bullets due to the distance the player needs to now cover to avoid the bullets, which limits how much can be done in that regard. A player will also naturally put a lot of distance between them and the bullets when they have no perceived safe-area or leniency-area around their hitbox (which is also one benefit of smaller hitboxes), so you'd need to show or at least imply where they can graze without being unreasonably close to death, so they even know that it can be done "safely". Meanwhile, grazebox size doesn't scale well, relative to the player hitbox and bullet sizes. A 5x5 hitbox and a 64x64 grazebox is much more lenient of a system than a 30x40 shipbox and a 96x96 grazebox, and clearly using the same scale is impossible. It's especially unruly if the shipbox is indeed rectangular at that size; that just wouldn't work at all. Additionally, the goal is still to get the bullets within the grazebox and not the hitbox, but because bullet sizes don't really change that much regardless of the hitbox size or pattern density (rather than scaling proportional to them), this creates another constraint on the relationship between the two.

[ciox]

Yeah agreed, the only way around that problem is having a tiny ship sprite, which is basically a loophole to have a small hitbox again.
The system I like is Psyvariar style with a really large graze radius that you can also increase to the max with a special move.