Character Concepts and How I Designed a Boss

It's been a while -- but I assure you, I've been doing a lot. The past four days -- that's right, FOUR! -- have been spent designing and implementing the introductory boss fight for the game. I guess the 4 is really more like 2; I have been dreadfully sick this week and I'm only now starting to get better. But I worked nonetheless! Maybe not quite as efficiently as normal, but I still worked!

I've begun talking character concepts with the asset artist. You can see some of her work below, as well as some work from my friend Ivan:

Ideas for our hero, Super Virgin Samurai:

Some of Ivan's Doombie concepts:

Some "types" of Doombies (I'm thinking the upper-right is boss material)

...and some narrowed-down concepts for "regular" Doombies:

I'm really excited to be seeing art for this now. Now, as for the boss fights...

I am not afraid to admit: I have never implemented a combat-based boss fight in my life. The closest I've come to boss fight implementation is in an older project of mine, LOVE! Starring Maily McMailerson. In LOVE!, bosses were trivial at best; they were simply oversized regular "enemies". However this time things needed to be a little bit different; and, being the first thing the player sees, the first boss fight had to be suitably intense.

I was completely lost at first. My first plan of action was to throw a boss on a screen, have him rotate between a few attacks on a timer, and leave it at that. But as I sickeningly started to implement the casing of a boss fight, something told me this wasn't right. I implemented the spawn mechanism for the boss, the introductory "scene", and a win condition, with a shell in the middle which I'd then populate with his attacks. But I stepped back and spent a night doing some research instead, and I'm glad I did.

I read Mike Stout's wonderful article on Gamasutra about designing boss fights, and it changed my perspective on the matter. I recommend this article wholeheartedly to any game developer who is interested in designing boss fights. The parts that stood out to me the most were his two points under "A Boss, in a Nutshell":

A boss is a test.

A boss is a story.

After reading the rest of his spectacular article, I took a look at my plans for the boss fight...and scrapped them almost completely. "Okay," I thought, "What is it this boss needs to test?"

Well, he needs to test the only player-controlled mechanics in the game: moving, jumping and hitting things. He needs to be an introduction. He needs to feel tough -- maybe even tough enough to kill you on the first or second tries -- but not actually be too tough. He needs to feel like a challenge and he needs to start the game off with a "bang".

So I drew out a story arc, and I planned intended player reactions at various points in the fight on it. I noted which mechanics would be testing during the rising action, up to the climax, during and after the climax, and during the falling action.

And it looked silly.

But I think the result turned out well. It's a three-phase fight: the first phase tests primarily jumping. The second phase tests primarily combat. The third phase mixes both phases together and ramps up the difficulty a little bit. The attacks look like they HURT: I took a page out of Elama's book and added a screen shake mechanism when the player takes hits.

When the boss first throws up his shield in Phase 2 and you hit it, he sends you abruptly flying across the room. After the relatively easy game of jumprope in Phase 1, it's a rude awakening, and suddenly you know "this guy means business". It's a simultaneously positive and negative feeling -- you were waiting for a challenge and now you may get one -- but you're shocked at what just happened. Just as my rising action above shows.

On his defeat you are showered in coins and scrolls. This gives players ample loot to start the game with. A few of the charted "emotions" above aren't implemented yet; there are some story elements I have left out. There are also a few parts of the fight which are not intuitive due to lacking animations; this will be fixed in time, but for now I think it's playable.

It's a tough fight and has gone through minimal testing. It will need a lot more: I tend to make bosses far too difficult. This build release will be a "first-wave" build -- each feature I implement goes through 2-3 "waves" before it's finished, as I work on new things. The features improve each time and it's nice to see them grow. I've already toned down its difficulty quite a bit. This weekend it will see criticism. I can't wait.

The feeling of accomplishment:

Stay tuned.

Combat Revamp, Part I (Plus Gameplay Footage!)

The changes to combat have begun!

I buffed up the little guys a lot -- they do about 1/3 of your health in damage now. Health did not feel like a useful stat:

- If you have no Speed, you feel a bit slow.
- If you have no Jump, your jump feels lacking.
- If you have no Swing, you feel like you swing slowly.
- If you have no Reach, you feel like you're using a plain sword and not a super-cool one.
- If you have no Luck, you "feel less lucky" -- less 'things' in general drop less often.

However, if you had no Health boost, you still felt pretty durable. I knocked Health's importance up to be equal to that of the other stats, and I think it's worth it -- now you don't overlook Health on your gear. Before it was "Oh, cool, this sword has Luck, Luck, Luck, Jump....and Health". Now Health has a purpose -- if you get hit a lot, Health will help you out now. Before the change, anything more than 1 line felt like overkill.

I did several things to make combat more interesting:

- The HP of the enemies has been buffed slightly.
- The damage enemies do to you has been buffed -- they actually feel "dangerous" now.
- Spawners can now wait for a period of time before they begin to spawn enemies.

The third point is the most crucial here, and it factually makes combat more interesting. The other two will require playtesting to see how they fare as features. What I can now do is set up spawners so they will start spawning mobs at interval x after time y. This enables me to generate "waves" of specific enemy types.

A screenshot of a spawner's parameters in the map editor:

If a SPAWN_PRETIME is not specified, it will assume it should begin spawning immediately. It's a decently elegant solution and I'm glad I made my initial code modular enough to support parameters like this. I have a feeling there will be many more parameters to come.

I also began moving the levels over to the new "art style". While it is still programmer art, I think this will be much closer to the style the game will be going with in the future:

Level '1':

Level '2':

Level '3' (you can also see a Trail effect here):

I lied a bit when I said "Gameplay Footage". This is a raw, unedited playthrough of the majority of the test levels. It was meant to be shown as purely demonstration footage to other game developers, particularly as an analysis of the current issues with combat. However I thought I would post it up so you can see it. The music is not in-game music -- it's just what I was listening to at the time. Again, this is raw demo footage meant for development purposes -- not a trailer.

Thanks for sticking around! I honestly appreciate every pageview I receive!

Taking "Programmer Art" to a Whole New Level

Visual Effects and the Save System are now done. Sorry there hasn't been a post in a while -- I've been working hard on this stuff.

I implemented a fully functioning particle emitter system (as mentioned in a previous post), and have now finished all of the Visual effects. There are over 70, stemming from 3 main types - Glows, Trails, and Special. Glows just make your sword glow; trails come in both long and short in 22 different colors; and special create effects like that of wind around your blade. All of these are currently programmer art, but some of them actually look pretty great.

Some trails:

I have also begun touching up the in-game art substantially. There is still no artist on the project but I wanted to give the game a quick makeover. My Tilemap reader now supports a modular number of layers and can draw those layers either above or below the player ingame.

It looks much nicer but is still quite obviously programmer art:

Coming up soon: a video maybe (?!)

Implementing an EXP Curve

"Ranking up" has been implemented. Currently it has no effect on the actual gameplay, it just makes you feel good about yourself because you ranked up. Which isn't necessarily a bad thing.

I don't want ranking up to effect your stats in any way. That takes away from the loot, which needs to be all-important. So that's out. I would also like the game, to some extent, to scale with your level. Weak enemies will stay weak, strong enemies will stay strong, but they will stay strong and weak relative to your current power. Ranking up will cause you to find "higher" stats on your gear, but you will also need more of a particular stat in order to get the same benefit. Think "leveling up" in traditional MMO-style games. However I'm not sure if this feels like a punishment or a success -- it will need to be implemented in order to really tell.

I also implemented a particle system. Two major systems in one night meant sleep at 6 AM and no time for a blog post. That's been happening a lot lately, so that's why the posts have been coming in at weird times. Particles are divided up by  types and behaviors, and each can be customized through a ParticleEmitter to change its lifetime, color, etc. -- This, of course, also paves the way for things like the sword Visual Stats.

ParticleEmitters can gracefully deactivate and clean themselves up. I used this here, so when you click off the "Rank Up" screen the confetti slowly fades out. Looks nice.

Of course, "ranking up" being implemented means that "EXP" was also implemented. This is functionally EXP in its traditional sense, however I wanted to stray away from the concept of actually having "exp". I wanted the ranks to feel unattached in that regard. So while you do gain EXP from kills, it is not shown and you're not aware of how close or far away you are from the next rank. Playtesting will tell if this is a good idea (probably not, but let's try something new).

Now for the more mathy bits. I calculated the exp "curve" based on the expontential formula with a few varying factors.

exp needed for level = (e^[x-1]) / (x * f)

Where x is the current rank and f is a reduction factor, used to stunt the curve. e is itself, the mathematical constant.

In the charts below you can see the formula in the Excel spreadsheet. 

Cx = Bx / (Ax * Dx)

Where x is simply the row number and A, B, C and D match their respective column numbers.

It isn't complicated, but I think it gets the job done. I'd like to see some comments from other programmers/designers who have done similar things in their games! (If anyone reads this anymore!)

I'm thinking the next build will have saving and level unlocking -- followed, most likely, by the combat revamp. It's the last clunky system in the game -- time to fix it.

New Special Stats! Other Nice Things!

Added a lot today. Among the changes were two new special stats.

Double-hit gives you a chance-on-hit to do damage to an enemy twice instead of once.
Ghostblade gives you a chance on each swing for your arc to go through walls!

Lots of quality-of-life improvements. Surprisingly, the code is still pretty clean. Anything that isn't clean is marked as such -- I have only been leaving unclean bits in sections that are unfinished anyway -- i.e. the Quit code was a bit of a mess "just to get it in", but today as I solidified the feature I cleaned it up. I left it open so I can have more buttons if I want to -- i.e. an "Options" menu, etc.

There is now a global Quit screen in addition to the in-game pause screen. At any time, pressing Escape will bring it up. On the inventory screen, each press of Escape will clear out your item comparator in the order you filled it, and then finally bring up the Quit menu, much like most popular games do nowadays. It feels slick -- would feel even more slick with some sounds!

Quitting from the Inventory screen:

Quitting from the Level Select screen (in case you didn't believe me!):

The Inventory screen has been made a little more robust too. Scrolls can be fused more easily now, and they can also be fused in reverse ("de-fused"). The controls are much simpler and only require a click on the green or red arrows if they are available.

I also FINALLY added the Favorite box. This is a place to put weapons you really like, but don't want to equip. They have no bearing on your stats.

I thought this Knitted Tickler was pretty hilarious, so I fave'd it:

Misc. Inventory screen shot:

I've also been slowly adding more levels. I'm discovering what "works" and what "doesn't" slowly, although mostly the levels have just been for testing.

I thought this level was particularly crazy...I'd like to include something like it in the final product:

And I just wanted to show off a Ludicrous I found:

On that note, Pinks were slightly changed. From the Changelog:

- Pink items now have a 5% chance to give 3 Standard Stats, down from 30%.

- Pink items now have a 45% chance to give 4 Standard Stats, up from 35%

- Pink items now have a 27% chance to give 5 Standard Stats, up from 25%.

This means you are much more likely to find Pink items that make you say "WOW!", more or less. 3 lines and only 1 special stat is kind of a downer, but 4 lines seems so much more substantial.

All in all, lots of great changes tonight. Changed some code structuring stuff, added a bunch more to the game and fixed a bunch of bugs. Keep reading and don't be afraid to send feedback!

The Shop and Whether to Continue With the Project

I implemented a huge amount of bugfixes and additions last night. Coingasm mode is a little different now and I fixed all of the known bugs with it. Basically:


 - Coingasm appearances tweaked.
  - Coingasm mode now has a 7% base chance of occurring, up from 5%.
  - Coingasm appearances are now effected by your Luck attribute.
  - The game will "roll" from 1-100. You will then gain a bonus to that roll based on your Luck attribute.
  - You will now gain an additional bonus to your roll based on how many matches you have gone without seeing a Coingasm.
  - If your roll is greater than a certain value based on all of the above, you will see a Coingasm.

It's much nicer now -- I would feel kind of disappointed after not getting a Coingasm for 50 matches, as would others. Now there is a slight additional "bonus roll" with each match that you play where you don't get a Coingasm. After roughly 30 matches there is basically no way you can't get a Coingasm. It was a negative player response -- you would feel almost like you "activated" the Coingasm by doing something, and that it wasn't coming up anymore because you were doing something wrong.

I also implemented the store. You're damn right that's programmer art.

The event I was going to present the game at was postponed until the 23rd. I think I may just continue with it anyway, which would mean you would get to start seeing some art in the coming month. We'll see.

COINGASM! (This is the name of an actual feature)

Introducing the new best parts of the game:

GET LOOT MODE!

and

COINGASM!

Seriously, these modes are ludicrously fun. When you've defeated the last enemy in a stage, "GET LOOT!" will appear on your screen. A countdown will begin -- the numbers start off huge and "slam" down against the playfield, shaking the screen a tiny bit and intensifying as the countdown draws nearer to zero. You've got to try and grab all the loot you missed before time runs out and the level ends.

The countdown zips to 0 if you've already collected all the loot, meaning no wasted time. But this isn't the best part. The best part is the small percent chance of getting a COINGASM! When the countdown reaches zero, it will take a very short break before sending you out of the level. Sometimes, after hitting 0, you'll get a COINGASM instead of ending the level! Tons of coins spew forth from the middle of the screen and litter the level and you have 15 seconds to pick up as many coins as possible.

Damn this is awesome. Videos soon.

It's the Little Things

Doing nothing but cleaning up for presenting on Saturday. Ready for my baby to get her harsh criticism. But most of all ready to share ideas.

It's the little things that make a product nicer. I implemented a screen shake today similar to that in Elama. It isn't caused by anything just yet but it works very nicely -- you can vary its intensity and duration and all that. It's nice. Needs a little smoothing for a dropoff but other than that it's great.

People are having fun playing the builds and it gives me confidence. If the 9th goes well I'll push a build out to a large audience, primarily full of people I don't know. I know the combat is shaky, and that's ok. I don't even care about the art, which is a first for me: that can wait. Because I understand now that these things aren't meant to be rushed -- it creates tension and lowers the quality of the product and everyone's time spent. The concept, the idea, and I dare say even a slight majority of the execution are all there. And that's a great start.

Finding the balance between "development" and "production" when working on personal projects is a challenge -- always has been, but hopefully will not always be. I'm starting to learn how it works; how to let my developer side flourish while still keeping to deadlines. And to be honest, knowing there was a small meetup on the 9th helped me with that: As soon as I heard it was going down, I set a deadline. And if I continue working on this project afterwards, I will set another one for another month down the road. And I will meet it. Sure there will be bugs -- sure there will be errors. But that's all part of the learning experience. And no matter which way things go for Super Virgin Samurai, I've learned a lot in this past month. One thing I've learned most of all is that I don't need to be in college to learn; that it's only the beginning. I always complained that I didn't learn enough in college. Maybe that was the point -- to really get to know what I didn't know. To get in touch with that side and learn to deal with it and even love it.

And as I push through ugly bugs and nasty blocks of code and learn more every day about what makes "good" code versus what makes "bad" code, I still don't hate the project. Some days I get sick of it. Some days I want to strangle the damn thing. But at the end of the day, if someone plays it and it puts a smile on their face...dammit, I did my job.

FUSING IS EASIER NOW! (A dev-heavy post)

Fusing, on first implementation, was a major pain in the ass. Now I've gotten it down a bit more. Visually it is still a bit confusing and there's more I'd like to do with it, but functionally it works quite well.

Note the right side of the screen. Each scroll type can still be fused like it was before (2 of a lower-quality scroll makes 1 of a higher-quality one), but now you can defuse them as well (1 of a higher-quality makes 2 of a lower-quality). It was sort of an annoying move to fuse all your scrolls into better ones only to never get any rare loot to use them on.

In addition to this new feature, you also don't have to double-click anymore to fuse. This was way more annoying than it sounded -- it's amazing what one fewer click can do in terms of "fun-ness". I also added the ability to simply hold your mouse down over an arrow to continuously fuse scrolls. It works great, but will still need some playtesting.

Also, title screen. (For now just a "kiosk demo"-type screen)

I also increased the drop rate of scrolls slightly as the game was feeling quite slow. The "shop" feature will fix this by allowing you to purchase different-quality scrolls (and I may be implementing that before the 9th).

I added some other cute tweaks, such as swords sometimes "sticking" in the ground upright when they drop from enemies. Tactically the game feels quite nice to play, but the combat is still clunky and MapleStory-ish. That's ok though -- that will be revamped after proof-of-concept stage. The art is, of course, godawful and very, very stolen. But I'm not worried about that right now.

I don't think I mentioned this, either: You actually hold your sword in your hand now. It adds a surprising amount of attachment to your blade, as the color of your swing arc also matches your weapon. It's disappointing to get a great weapon with a color you don't like. This opens up in-game store possibilities, whether through in-game currency or outward monetization.

I've also added a bunch of levels. They're still quite basic as far as complexity goes, although some of them are definitely harder than others. But it's nice to see content and have different environments to test different things in. The level loader tool is beautiful -- not having to recompile just to test a different level is amazing, as is the ability to add new levels and even level groups on the fly.

I'm starting to consider releasing the RNG for Super Virgin Samurai as open source code. It's simple, but useful, allowing you to do little tasks such as get a specified amount of random numbers from a set with and without duplicates and so on.

Special Stats: Return of the Stats

WE ARTWORK NOW! Nah, not quite, but I did have a friend of mine (and potential game artist) doodle up some 5-minute rough ideas for our lovable Super Virgin Samurai and the evil(-ly cute?) Doombies. Check it:

Did I mention the stats? Anyway. Lots of stuff done. No really. A lot. You can now kill yourself using the K key. This is great for...I dunno...people who like to die. (It's kind of a nice way to get out of a match. Still undecided if it should cause you to lose the weapons you gained during the match).

(Programmer humor)

A few more Special stats:

Damage Block gives you a chance to negate incoming enemy damage. It's useful in a...less obvious way. It's actually saved my butt quite a few times.

Damage Reflect bounces the damage you take back to your attacker -- but you still take the damage. It's a little schadenfreude, and it's pretty fun to kill things by slamming your face into them.

Instakill gives you a chance on hit to do some ludicrous damage to an enemy, killing most weaker ones.

That's also a screenshot of another cool feature: PAUSING. YEAAAAAAH!

Something interesting to take note of is the amount of balancing I have been doing. There is quite a lot of content to be balanced even without particularly interesting levels and only 3 very simple enemy types. It's making for an interesting programming & design experience and does require quite a bit of thought to do what I feel is perfect for the game. I'm not allowing myself to cut corners with this proof of concept -- it's gotta be what I want it to be.

Certainly I can say, however -- getting a Ludicrous feels damn good.

Loot drops are a bit more varied now too; scrolls drop more often as well, and killing stronger enemies often yields more scrolls.

All in all this is shaping up to be a pretty fun project. I'm hoping it will go well on the 9th and that I'll choose to push forward with it -- if I do, you can be sure to start seeing some playable builds as well as some artwork! I will be looking extensively for alpha testers, and you might be one of them!