Not a whole lot to say here, but progress on Boss Rush continues well.  At this point, the game is basically nailed down and probably a good 90% complete.  Most of the work at this point is down to playtesting and polish, and occasionally adding a new level or two if I think of one that seems like it would be fast, easy, and safe to add.

In addition to the occasional friend or family member that I can rope into playing my game, I have enlisted the aid of some forumgoers over at TIGSource.  I have to say.  They have been wonderful fonts of information, suggestions and excellent feedback.  (Frankly wishing I had gotten around to adding more playtesters quite a bit earlier.  One more data point for the postmortem I guess…)

It’s getting close though!  I finally realized just how close when I watched a friend play on Sunday.  I think he spent a good hour playing through levels, and still didn’t get through all of them.  (And he didn’t even touch multiple difficulty levels.)  It was an interesting perspective moment, where I realized just how much I had completed, and started to feel, really feel, like “huh, it’s starting to feel like a full game now.”

It was a good feeling!   I’m feeling kind of proud of my game at this point.  Now to just finish it!

So to make sure that doesn’t happen, I need to make sure that stasis is never reached.  And one of the easiest ways to do that is to gradually shift the game rules.  This gives me two advantages.  First, if they have found something that works, then as I shift the rules, there is a good chance that it will stop working at some point.  (Provided I’m shifting the right rules at least.)  Second, I can use it to effectively put a cap on game time by shifting towards something that is, for all intents and purposes, impossibly hard.

Tetris is a classic example of a game that uses this.  (Particularly the impossibly hard part.)  It gradually gets more difficult by making pieces fall faster.  At some point, the pieces are falling fast enough that it is not even physically possible to move them to the edges of the board before they hit the bottom.  At this point, the game is effectively over.  The player can, if high enough skill, put off the inevitable for a little bit, but the writing is on the wall.  The game is going to end.

So too in Boss Rush.  The little red ships eventually gain homing missiles as an upgrade.  If they survive long enough, this gives them the ability to hit you from just about anywhere on the screen.  This is important to me for balance, since it is my only real defense against strategies that involve bugging out my ship AI by trapping it in a corner somewhere where it can’t hit you.  (I have far more faith in players’ abilities to abuse my AI than I do in my own ability to foresee all such gimmicks and plan against them.)

However, here is where playtesting gave me some interesting insights.  It turns out that, while players don’t mind being hit by missiles if the little ship got them through power ups, they detest having the ship start out with missiles.  This feels completely unfair, and they generally feel (fairly correctly, really) that the game is effectively over, but for the screaming.

“Huh”, I said.  “That’s problematic.  I really kind of need ships to start getting missiles, as part of the difficulty curve.”  So as an experiment, instead of letting the ships start with better weapons, I instead cranked up the spawn rate of ship power ups.  It becomes somewhat absurd now, with powerups pretty much pouring from the sky in curtains, like some kind of weapon-improving raindrops.  The ships still get missile upgrades pretty quickly.

And yet…  It works.  Players (at least the ones I tested with) don’t seem to mind this nearly as much, even though the net result is the same.  Some theories as to why:

• Since it’s a familiar scenario (little ship vs. big ship) changing the little ship starting loadout feels like it’s breaking the metaphor.  We’ve all played shmups before, and we know, the reward for being blown up isn’t starting up again with full powerups.  It’s being kicked back down to weakling level, and having to claw our way back up.  So maybe this implicit break in the expected structure is jarring?
• Having the ships collect items feels like there is more agency involved.  It is still theoretically possible for you to stop them, or at least delay them from getting enough to hit missiles.  And if you can kill them before they get the missiles, then they have to start over.  So even though it’s hard, it presents the players with more of a feeling of “I have a chance”?
• The abruptness of the spike.  While missiles are only level 3 out of 5 possible power ups, they represent a significant increase in danger level.  Players tended to not care as much about ships starting with more guns, but once they started with missiles, it suddenly felt like they were on borrowed time.
• The two step process “it’s raining powerups” -> “he has missiles now, oh crap I’m dead” is better for the anticipation than just jumping straight to the “crap, missiles” stage of the game.

I’m still not completely sure what the exact motivation is here.  I still don’t feel like  I completely understand what is going on in my testers’ heads.  (Which, I suppose, is why I’m running playtests in the first place…)  It might be different for each person, and might not be any of these.  Even so though, I think I can draw at least a couple of useful lessons from the experience that I should keep in mind in the future.

• A slim chance that is almost impossible feels better to players than zero chance.
• Anticipation and telescoping are good.  (Valve talked about this at GDC this year in fact.  Hearing “rumblerumblerumble” and going “oh crap, incoming tank” is a more memorable experience than just “Surprise!  Tank!”)
• Attacks that are very difficult to dodge are acceptable as a punishment (you didn’t do XXX so now he gets to attack you with YYY) but not as an unavoidable part of a game.

Good stuff!  Hooray for playtesters, I guess!