Welcome, Guest. Please login or register.

Login with username, password and session length

 
Advanced search

1411276 Posts in 69323 Topics- by 58380 Members - Latest Member: bob1029

March 28, 2024, 11:57:23 AM

Need hosting? Check out Digital Ocean
(more details in this thread)
TIGSource ForumsCommunityDevLogsBoss Rush
Pages: 1 ... 4 5 [6] 7
Print
Author Topic: Boss Rush  (Read 38625 times)
Montoli
Level 7
**


i herd u liek...?


View Profile WWW
« Reply #100 on: October 22, 2009, 11:03:38 PM »

And now... it's time for...
Boring Technical Post talking about how the AI works!
Feel free to skip if this topic is of no interest to you!


The AI in this game was interesting.  It was actually the 2nd thing I wrote.  (The first was a basic pattern of bullets that I liked a lot, and yet somehow never made it into the game.)  I figured, if I couldn't make a general dodging routine work, the game wasn't worth pursuing much further, so I tackled that early.

My first attempt was a total hack, which mysteriously worked far better than it ought.  It worked like this:

  • When checking for bullet-ship collisions, keep track of any bullet that is vaguely near the ship, within some set radius.  These are the bullets that the ship cares about.  (everything else is too far away)
  • Every frame that the ship updates, apply a force to the ship pushing it away from all bullets that it cares about.  Make this force inverse-squared, so it gets stronger and stronger the closer the ship gets.
  • (If any bullets are within 5 pixels, engage "panic mode" and disregard all other [non-bullet] forces.  Otherwise....
  • Also apply a force on the ship TOWARDS any power ups.
  • Also apply a force on the ship towards the point 150 pixels below the nearest boss module that isn't invincible.  (Modules boss collision shapes, which can often be destroyed)
  • Add all these forces together, and then normalize the result to the ship's max velocity.
  • Apply some basic momentum to the ship so its movement is smooth curves and not frantic jitters.  (Also to make it less perfect and more killable)

And... Voila!  Done!


Again, this actually worked surprisingly well.  No fancy lookaheads or anything, just straight up "oh, is there a bullet?  Get away!".

One minor hack in here is that bullets actually have TWO radii.  They have a hitradius (get within this and blow up) and a dangerradius (what the AI pretends is the hit radius, usually the bulletRadius * 1.5).  I had to make the danger radius much larger on lasers though, since they rotated so quickly.  (Lasers are just lines of bullets that look the same, since that means I don't have to write custom code for dodging them.)


So yeah.  This worked really well.  Far better than I expected any way.  But there were still a few problems:

  • In situations where bullets were approaching from both sides, if the ship was perfectly between them, it would just sit and take it.  This was an unsatisfying way to kill the AI, since it looked like he just sat there and took it.
  • In situations where bullets were extremely dense, the AI would act erratically.  A clump of 5 bullets stacked on top of each other (or near each other) is just as dangerous as one big bullet.  But the AI would give it x5 the weight, since there were 5 of them.  This was often enough to actually overcome the force generated by a different bullet!  So the ship would see a bunch of bullets stacked together (like, say, the results of Nidhoggr's base attack, the rows of spines) and jerk away from it, straight into another bullet...  This was also an unsatisfying way to win, since the computer seemed to suddenly stop dodging well, and just fling himself into a bullet.


So.  I eventually tore everything down, and rewrote it!

The new version is a little more complicated, but is a bit smarter about things like dense bullet zones, and symmetrical attacks.  It works like this:

  • Still only track bullets that are close to the ship.  No sense caring about things that are more than 60 pixels away or so.  They won't get here for a while.
  • Examine 9 points.  One is the point it is sitting on, and the other 8 are the 4 directions + diagonals, as far way from the ship as it can move in one turn.  For each spot, assign it a rank (-1, 1) based on:
    • Is there a bullet in this spot already?  If so, automatic negative rank.  (don't go here!
    • Is this spot in line with any bullets, assuming they continue on their current courses?  (not a guarantee, but usually good enough)  If so, penalize the ranking based on how close the bullet will pass, and how many turns before it arrives.  (if it's not arriving for a while, and not passing THAT close, then it's not so bad.  Not as bad as if it will hit this spot in one turn, for example...  (These calculations are done via some basic linear algebra distance-from-point-to-line formulae)
    • Will moving to this spot put us closer to the optimal spot to shoot the boss?  (150 pixels down from the nearest non-invincible module)  If so, positive rank!
    • Will moving to this spot put us closer to the nearest power up?  If so, positive rank!
    • Will moving to this spot put us within 50 pixels of a corner?  If so, negative rank!  (Corners are death traps!)
    • Will moving to this spot put us within 20 pixels of the edge?  Also negative!  (although not as bad as corners.)
  • After all of these points are evaluated, move in the direction of the point that got the best score.
  • repeat every frame!

The behaviors also have various weights (controlled by easy-to-tweak constants), so it's easy to adjust their relative weight.  Right now I have something like:
  • First priority is always dodging bullets.  Bullet weight is high enough to counteract everything else by itself if the ship is in danger.
  • Second priority is getting powerups.  Beats all considerations except dodging bullets.
  • Next is avoiding corners and edges, although it's not THAT far ahead of "shoot the boss" so sometimes it goes near the edge.
  • After that is shooting the boss.  Because while it's important, getting powerups and staying alive is more important.

This strategy has worked far better!  Since spots are ranked, pincher attacks from both sides no longer confuse it.  (Because it can see at least one way out and will pick one rather than sitting in the middle, undecided.)  Also, since it is no longer moving AWAY from things, but moving TO a spot that looks best, it no longer flings itself away from clusters into other bullets.

It does still have a few flaws though.  If you shoot a solid plane of horizontal bullets at it, it won't know what to do.  (assuming it can't fit between them.)  (The "bullets are springs" method has this problem too though.)

Also, if it gets stuck in the middle of a huge mess, it will sieze up.  (for example, if it is in the middle of the giant laser from tonbogiri's second form, it won't try to leave, because all places it can move are equally bad.  The spring method would at least get out of there quickly.)


Overall, I've been pretty happy with it though.  Writing a general-case solution to "how to dodge bullets" has let me basically ignore the ship as I come up with bullet patterns, since I know it will usually be able to find a way through if there is one.

ONE attack has required some special "helper" code.  The lighting bolt from Cloud-9 moves fast enough that if you click and shoot the ship directly, it can't react fast enough.  (Since it doesn't notice the attack until the attack gets in it's "I care now" radius, by which point it is too late.)  So my hack for this attack was to make the lightning bolt spawn a row of invisible bullets, with a hitradius of zero (so you can't hit them), but a dangerradius (what the AI cares about) of 10, in the path that the lightning will take. (And then they stick around for like 10 frames.)  This means that even though the ship doesn't notice or care about the actual lightning bolt coming towards it, it notices that there is a danger spot where the lightning is going to be later, and so gets out of the way.  I don't like having to write special cases, but I like the lightning attack too much to ditch (or slow down) and at least this one I was able to handle entirely in the scripting language, and without touching the core engine code (or AI code) at all.



So yeah.  That's how I'm currently handling the AI.  I hope some people found this interesting!  It's been a fun challenge, since I've never tried writing anything like that before, but it has been a good learning process.  (and I'm fairly happy with what I ended up with.  Especially since, because the AI is the bad guy, it doesn't matter if it's not perfect.  If it WAS perfect, it would be too frustrating. As long as it gives a decent fight, it all works out. :D)


-Montoli
Logged

www.PaperDino.com

I just finished a game!: Save the Date
You should go play it right now.
mokesmoe
Level 10
*****



View Profile WWW
« Reply #101 on: October 22, 2009, 11:20:37 PM »

Is this in reference to the missions that are shared across several bosses (blow up a target, kill N ships that can't shoot back, etc) or just in general?
I was referring to the copied missions
I know they are still unique gameplay wise, but making them seem more unique when looking through the missions will give a sense that the game has more missions.

Quote
Do people feel like missions are getting too far out into left field?  Should there be more "kill ships, with some constraint" style missions?  (things like the "Don't shoot the target" one?
I personally like the more odd missions. Also, the 'super turret' missions feel like  advance training missions, teaching you techniques like blocking and advance control.
Also, I think you should have another mission like 'learn to fly' where you have lasers above as well, so you actually have to control the ship, kind of like the helicopter game. (or whatever game you were thinking of when you used this control scheme)

Quote
The default versus options are basically random.  Since most of my friends aren't really into shooters (and the ones that are live in Philadelphia, on the opposite side of the country from me) I haven't had much chance to test this first hand.  If anyone who has managed to find people to play this with can weigh in, I'd love to know what people think would be a reasonable "default value" for the settings.  (I basically tried to sidestep this by just adding lots of options and hoping people could find a group of settings that was fun.)
The biggest problem is that the bosses aren't all the same difficulty to fight against. Also, humans play differently than the AI. Humans are good at staying away from areas dense with bullets, (one of the AI's biggest flaws) but aren't often very good at dodging complicated patterns. (the AI's biggest strength) Maybe you should include a difficulty meter for versus (amount of stars or something) showing how hard of a boss it is, and if you should turn the options up or down.

Quote
Sadly, flash can't handle right-mouse button pressing. Sad  I'll admit I had assumed/hoped that the numpad keys would work if you turned on numlock, but after just checking, seems that's not the case.  (what makes it even more annoying is that they DO register as arrow keys.)
I'm sure I've seen games that could do it. It could have been something else like Java though.

Quote
Can you elaborate on the bug you saw?  what part of mouse controls stopped working?
Take out Tombogiri's wings, removing attack one. Then if you click on your first attack, witch is now attack two, it selects your second attack, which is now attack three. Clicking on attack three selects attack four, clicking on four does nothing. No way to select attack two.

Quote
(also:  Your names for moves are much cooler than [most of] mine.  Mine tend to be in code labeled things like "roundBullet" and "sideShot", or occasionally difficult to translate things like "spinnyDeath" or "cone")
Thanks! Although for for a few, I couldn't think of a good one. Like Grow. Although you seemed to ignore that name and called it Regrowth, witch sounds much cooler. (personally, I liked my themed ones the best. The ones for the face and Cloud Nine)

Quote
This ship has undergone a lot of changes.  Originally the first form was supposed to be a total psych-out, like "here is the boss... OH NO HERE IS THE REAL BOSS BWA HAH HAH".  Originally he had like 2 attacks, but in order to complete the ruse, I had to give him more.  (Until now where he's basically as involved as several "full" bosses by himself.)

He's still a little weak in the first form though.  (Originally because again, the first form was so crappy, so I figured I'd get people out of it quickly.)  I can certainly bump up his hp a little though.  In fact.  Bang.  It's done.  (Holy crap, it was down at 20, the default.  For reference, each ship bullet does 1, and missiles do 2 damage, so no wonder it was dying so fast.)
I can see how he got messed up. Sort of 'lost in translation'.

Quote
First off the white background works very well; you will have some people say make it black, but ignore them.  I was very close to making my game with a grey background, I think the white does in fact work really well.
The white background is, alas, not universally loved.  I think I will keep it though, since it's how I envisioned the game originally, and I still like it.
I like the white background. If people want a black background, use inverse mode. Or Omega.

Quote
Bonus options (unicorn mode, etc) unlocked after beating all challenges on a ship on "sparta" difficulty.  (Gotta give'em something nice if they're that crazy)
Does this mean you need to donate to get unicorn mode? You're an evil man.



EDIT: You posted just before me, so:
A few AI things, in no particular order:
For the seizing up in the giant laser thing, You should make it so if it's invincible, but in a position where it is touching a bullet, you eliminate the choice of not moving, and force it to use the best move, even if they all suck.
When you use the giant laser attack, the AI always goes into a corner. You should either make a special case, or make the corners weigh more.
For the lightning attack, you should make the danger radius larger for the lightning, so the ship doesn't sit next to it, and get insta-killed when it explodes.
Getting power-ups, should rank a bit lower. Not any lower on the list, but closer to the other things. The AI can act just plain stupid when it's going for a powerup.
Also, you need a special case for shooting the queston mark, so it doesn't just keep shooting at the invincible dot underneath.


Also, completely unrelated to the AI, the ? boss needs to reference the interrobang somehow: ‽‽‽‽‽‽‽‽‽‽
« Last Edit: October 22, 2009, 11:42:51 PM by mokesmoe » Logged
Montoli
Level 7
**


i herd u liek...?


View Profile WWW
« Reply #102 on: October 22, 2009, 11:54:14 PM »

Also, I think you should have another mission like 'learn to fly' where you have lasers above as well, so you actually have to control the ship, kind of like the helicopter game. (or whatever game you were thinking of when you used this control scheme)

Play it on sparta difficulty. Evil
Logged

www.PaperDino.com

I just finished a game!: Save the Date
You should go play it right now.
Montoli
Level 7
**


i herd u liek...?


View Profile WWW
« Reply #103 on: October 23, 2009, 12:21:20 AM »

A few AI things, in no particular order:
For the seizing up in the giant laser thing, You should make it so if it's invincible, but in a position where it is touching a bullet, you eliminate the choice of not moving, and force it to use the best move, even if they all suck.
Unfortunately, in the current implementation, if a spot is actually TOUCHING a bullet, (i. e. moving there will get you hit) then it is automatically set to the lowest possible score, since you would NEVER move there.  It normally DOES choose a spot, the problem is just that in this case they all suck equally.  (and since "stay put" is the first option it tests, and nothing else it finds is ever better, it stays with that one.)

I guess I could write a special case for having it try to keep looking further away until it finds a point that is safe to move to, but I'm hesitant to have it do that broad a search.  It doesn't come up THAT frequently in the game as-is.  (Mostly just if they have the misfortune to spawn in the beam while you're using Tonbogiri's giant laser)



For the lightning attack, you should make the danger radius larger for the lightning, so the ship doesn't sit next to it, and get insta-killed when it explodes.
this is probably true.  I'll fiddle with this tomorrow.  Sometimes it manages to dodge the little zaps, but all too often it sticks next to the side and pops.

Getting power-ups, should rank a bit lower. Not any lower on the list, but closer to the other things. The AI can act just plain stupid when it's going for a powerup.
Can you give examples?  In general it I've never seen it [directly] kill itself going for a powerup.  What trouble is it getting itself into?


Also, you need a special case for shooting the queston mark, so it doesn't just keep shooting at the invincible dot underneath.
probably true.  Although I like the invincible dot!  It's the ship's one weird unique defense.  (All the other ships that have no/few invulnerable spots have some form of self-healing and/or blocking ability.)


Also, completely unrelated to the AI, the ? boss needs to reference the interrobang somehow: ‽‽‽‽‽‽‽‽‽‽

Actually, I seriously considered the Interrobang as being the 3rd form, honestly.  I didn't becase
  • I wasn't sure how many people would get the joke.  (How many people even know what the interrobang is, anyway?  Last time I brought it up in a conversation people looked at me weird and I had to wikipedia it for them before they would believe me that it was real.)
  • I was hesitant to spend more work on a boss that was originally just a joke boss.

In fact... here is the thrilling tale of....  Unknown's Origin!


I was getting ready to send a build to my good buddies at Final Form Games and had three of the four bosses done.  At that point, my plan and been to have Tonbogiri, Nidhoggr, Antonius, and the Genesis Ark as the only bosses in the game.

I wanted to send a build, but I didn't want to give away that the final boss was going to be a floating head in space.  So I put a question-mark icon over the icon I had in my boss-select screen, so as to keep the surprise.

This was at around 12:00 PM.

Then, I was like "wouldn't it be awesome if, clicking on the questionmark icon actually gave you a boss that was a question mark!  People would be like "wtf?" and I'd be like "what do you mean, I thought the icon was rather clear?  What were YOU expecting it to be?"

"Haha" I answered myself.  "that would be pretty great.  Too bad these bosses, while quick, still usually take 2-3 days to actually come up with enough attacks that I'm satisfied with.  Why, to make a boss like that, I'd have to come up with at least 4-5 attacks that fit his theme.  And what IS his theme, anyway?  Punctuation?  Not worth spending the time it would take to come up with that many fun punctuation attacks."

At this point my Muse elbowed me in the ribs, and was like "ahHEM".  And provided me with basically all of the attacks, as you see them now, flat out.  This has never happened before or since.  Every other attack has gone through multiple revisions and tweaks and versions until I was happy with it.  Yet somehow, the Unknown boss came together in one evening.

At first I was leery of putting it in, since it is fairly different theme-wise from the rest of the bosses.  (The first 4 at least, where I pretty much played it straight and held with general tropes and rules and themes of the genre.)  Then I basically said to myself "isn't this the point of being indie and making this on my own terms?  If I have crazy ideas like this, I can just do them, and not have some producer tell me that we'll never get licenser buy-in, and the publisher hates easter eggs or jokes, so take it out please?"

Later, at 8:00am, I went to bed, having delivered an early playtest build to final form games, with one more boss in it than I had been expecting.  (The next day when talking with them on the phone, I got the idea to have the 2nd form be an exclamation point and put it in along with the new form-specific attack.)

Also the next day I was like "huh.  I....  guess I have one more boss in the game than I was planning on?  ... cool?"
Logged

www.PaperDino.com

I just finished a game!: Save the Date
You should go play it right now.
Jason Bakker
Level 2
**


View Profile WWW
« Reply #104 on: October 23, 2009, 12:29:24 AM »

That post about the ship A.I. was the opposite of boring Grin
Logged

BadgerManufactureInc
Guest
« Reply #105 on: October 23, 2009, 02:10:32 PM »

I'm finding lots of useful ideas and interesting design choices from following your devlog..
I'm really looking forward to playing the final version, this is shaping up great.
 Smiley
Logged
mokesmoe
Level 10
*****



View Profile WWW
« Reply #106 on: October 23, 2009, 02:14:10 PM »

Unfortunately, in the current implementation, if a spot is actually TOUCHING a bullet, (i. e. moving there will get you hit) then it is automatically set to the lowest possible score, since you would NEVER move there.  It normally DOES choose a spot, the problem is just that in this case they all suck equally.  (and since "stay put" is the first option it tests, and nothing else it finds is ever better, it stays with that one.)
Just make it so if the spot it's already in has a bullet in it already, it get weighed down more than normal.
Quote
Can you give examples?  In general it I've never seen it [directly] kill itself going for a powerup.  What trouble is it getting itself into?
Not directly run into bullets for powerups, but it will do some stupid things. For example, If I am over top of a powerup, It will wait for it sitting right in front of me, so when I fire my cannon, it gets insta-killed. Also, If will often move beside me or behind me, where I can crush it with ease. Maybe you should add a 'staying away from the boss' priority, because thats the real problem.
Quote
probably true.  Although I like the invincible dot!  It's the ship's one weird unique defense.  (All the other ships that have no/few invulnerable spots have some form of self-healing and/or blocking ability.)
Nothing wrong with the dot itself, just make the AI able to deal with it.
Quote
In fact... here is the thrilling tale of....  Unknown's Origin!
Cool story bro!  Durr...?
Thats acually pretty sweet. I remember playing smash bros and thinking of how cool it would be to fight on a stage made up of the word RANDOM.
Logged
Montoli
Level 7
**


i herd u liek...?


View Profile WWW
« Reply #107 on: October 24, 2009, 11:50:35 AM »

Sigh.  You're right, I could do most of those things.  And you're even right that I should.  Most of them are just things that I've been rationalizing putting off, and so my kneejerk reaction at this time is "oh no, I shouldn't touch that right now!"

Bah!  Why ya gotta be all reasonable and such!

I'll look into AI seizing up today.  In theory that should be a quick fix - instead of setting the weight to -1 in those cases, set it to -2 + (distanceFromBullet/bulletHitRadius) or something.

The stupidity of the ship re:powerups though might be harder to change without introducing some specific hacks.  (Which I'd prefer not to do - a big strength of it so far has been that I haven't had to write specific code [except in one minor case] for letting it deal with things.)  The problem is that ultimately, it doesn't look more than one move ahead.  Oh, it SORTA looks further, since it analyzes trajectories and figures out how long until a bullet hits a spot.  But for its decision-making, it just always looks for the next best spot.

The intended behavior is that it will artfully weave its way through a field of bullets to try to get to a powerup.  And it does this!  But it doesn't look far enough ahead to realize "oh, by chasing that power up, I'll hit a wall where I can no longer go further and then I'm screwed".  And making it do so is a bigger task than I want to undertake just now.

Not sure what is the best way to make it handle the invincible dot elegantly.  I could make the question mark top be two modules instead of one, but then they'd have separate hit-flashes, which isn't what I want.  Hmm.  I will think on this further.


Let's see.  Other random news:
  • Have added a couple of new challenges.  Right now am focusing on adding the "Meet xxx!" challenges that walk you through using each boss.
  • I hope to have a new build out sometime this weekend.  (Probably Sunday night.)
  • Because it was the first random thread I happened to see when I got up this morning, I contributed to the show us your sketchbooks! thread, over in the design forum.  If you're curious about more random glimpses my design process, there are 3 random pages from my design notebook.  Also, as a bonus here are two more [admittedly less interesting] pages that didn't make the cut.
Logged

www.PaperDino.com

I just finished a game!: Save the Date
You should go play it right now.
mokesmoe
Level 10
*****



View Profile WWW
« Reply #108 on: October 24, 2009, 07:11:04 PM »

Just give the ship a larger danger hitbox like the bullets.

For ?'s dot, you could have two smaller sections hidden behind the main part on the two sides that the ship would target, but would never actually get hit. You would also need some sort of way to make the ship ignore the main part.
Logged
Montoli
Level 7
**


i herd u liek...?


View Profile WWW
« Reply #109 on: October 31, 2009, 09:36:08 AM »

Quick update, then back to the frantic codings.

Hand Any Key Crazy Hand Any Key

Trying to get a solid build ready in time for IGF.  I'm close.  I'm SO close.  And as I understand the rules, I have until just before midnight tomorrow night.  But I'd really like to get a final candidate done today (was hoping for last night really) so I can pound the heck out of it with last minute testing.

I'll probably send out another build today at some point, but given the short notice, I'll understand if people don't have time to give me a whole lot of feedback. :D

Anyway, onward.  Exciting times!

-Montoli
Logged

www.PaperDino.com

I just finished a game!: Save the Date
You should go play it right now.
Montoli
Level 7
**


i herd u liek...?


View Profile WWW
« Reply #110 on: November 02, 2009, 01:49:47 PM »

Smiley Smiley Smiley


That's all.

-Montoli
Logged

www.PaperDino.com

I just finished a game!: Save the Date
You should go play it right now.
Hima
Level 4
****


OM NOM NOM


View Profile WWW
« Reply #111 on: November 03, 2009, 08:28:30 PM »

I wish today was Saturday. Angry
Logged

HarrisonJK
Level 0
***



View Profile
« Reply #112 on: November 03, 2009, 08:41:28 PM »

Is that a Teenage Mutant Ninja DINOSAUR!? Also, can't wait to play this.
Logged
Montoli
Level 7
**


i herd u liek...?


View Profile WWW
« Reply #113 on: November 06, 2009, 03:50:52 PM »

haha, yes, for some reason, lately my whiteboard has had pictures of Dinosaurs on it.  Dinosaurs who are also ninja!

Anyway, after a hectic, hectic last week, I'm taking a few days off from the project, and going to (hopefully) finish up everything I need to get started on a public release early next week.  Going to try to finally start the whole FGL-look-for-sponsors-etc bit.
Logged

www.PaperDino.com

I just finished a game!: Save the Date
You should go play it right now.
mokesmoe
Level 10
*****



View Profile WWW
« Reply #114 on: November 08, 2009, 01:43:30 AM »

Umm...

Logged
Montoli
Level 7
**


i herd u liek...?


View Profile WWW
« Reply #115 on: November 08, 2009, 04:13:31 PM »

Happily fixed!  But yeah, that shouldn't have been there.

I'll try to send out an updated build this week, with most of those bugs a bit more fixed.
Logged

www.PaperDino.com

I just finished a game!: Save the Date
You should go play it right now.
Montoli
Level 7
**


i herd u liek...?


View Profile WWW
« Reply #116 on: November 23, 2009, 09:48:02 AM »

Sorry for the 2-week silence.  It was a combination of post-IGF-build collapse, (I planned on only taking a couple of days off, but it stretched into a week and a half before I could blink) boring business stuff, (making sure the official page was up to date, etc.) and minor bugfixing.

As people probably noticed, I also never actually got around to sending out an updated build, either. Undecided

Sorry about that!


Moving forward!  I have placed the game up on FGL, which means, amongst other things, that I can be a bit less paranoid about sending it out to people.  (I probably could have been a bit less paranoid in the first place, frankly...)  If people are interested in seeing what will probably be the final version, I believe this link should work!  (Although I think you need a developer account at FGL if you don't already have one.  They're free to sign up for.)

http://www.flashgamelicense.com/view_game.php?game_id=8571

I tried to make sure to stick everyone in the credits who sent me feedback or posted it in comments.  If I missed anyone, my sincerest apologies - My bookkeeping could use some work.  Shoot me an email and I will remedy the situation forthwith!
Logged

www.PaperDino.com

I just finished a game!: Save the Date
You should go play it right now.
Montoli
Level 7
**


i herd u liek...?


View Profile WWW
« Reply #117 on: November 23, 2009, 11:39:31 PM »

Made me a thing.  On the tubes.

The

.
Logged

www.PaperDino.com

I just finished a game!: Save the Date
You should go play it right now.
Loren Schmidt
Level 10
*****



View Profile WWW
« Reply #118 on: November 24, 2009, 04:42:12 PM »

That's a nice, succinct trailer. I hope it helps you get fairly compensated for all the good work you're putting in here. It was weird watching the trailer, because even though I've played the game I couldn't help but automatically imagine I was the little red fighter. It goes to show how ingrained such things become Smiley

I owe you an apology- I was going to play your game when it first went up on Flash Game License, but I went space cadet on you and totally forgot until now. Here is my (much delayed) feedback:

You did a pretty good job with the visuals. Certain aspects of the graphics are really polished. I think the explosion effects and the appearances of the ships are the strong points. I also like the missile contrails.

I feel that the GUI isn't as nice looking as the game itself. It might not hurt to have a friend who's good at graphic design go over the game and point out a few things to improve in that area. Graphic design isn't my strong suite either, but my hunch is that a little bit of work in this area could go a long way.

I think the first ship feels less well tuned than the later ones. I found myself liking the game much more once I played with the second and third ships. I don't know if that reaction is unique to me or not. In the case that it's not, it might be good to consider giving the first ship an extra layer of polish. Most people's first impressions of the game will be formed by playing with that ship, and it would be a shame to lose people early when there's so much good material later in the game.

You were pretty creative with the boss designs here. You do a good job of giving each neat mechanics that set it apart from the others. I think my favorite boss is number two. Creating and maintaining a fleet of little minions is fun. I also like boss 3 quite a bit. It's really fun to pilot; moving it around has a nice feel (especially when combined with the various attacks that are affected by the arrangement of its segments).

Best wishes as you search for a sponsor! I hope your trailer will help.
Logged
mokesmoe
Level 10
*****



View Profile WWW
« Reply #119 on: November 24, 2009, 07:39:53 PM »

The tonbogiri's laser is very overpowered. If you quickly shoot a laser on either side of the rebel ship, the ship has no way of avoiding it and dies. There is no reason to use any of his other attacks.
Logged
Pages: 1 ... 4 5 [6] 7
Print
Jump to:  

Theme orange-lt created by panic