Friday 17 January 2014

868-hack update

Okay, 868-HACK has been updated on iOS. The update clears the high score tables and starts fresh. The old scores can still be viewed through gamecenter under "legacy". If you don't update, any new scores you get will remain separate. If you're in the middle of a streak, updating will reset that so scores can't be carried over that way (sorry!). Your local score list is also cleared, so if you want to keep that record take a screenshot before updating.

Since release I've been learning more about the game and tweaking the balance. Most of the changes are smallish, but they add up and make a difference to how easy it is to score. Many of the things I've learnt have come from observing the game in the wild; I couldn't have known to account for them before release.

At some point during development .STEP allowed stepping off the edge of the level to wrap around to the opposite side: this felt very cool as a glitch in an abstract digital space, but was far too useful for skipping levels (because it makes the corners adjacent) so it didn't stay. So at release I felt like .STEP was slightly underpowered because it was weaker than it had been; it was more expensive than .WAIT / .PUSH / .PULL as a parity flip and it was less useful for grabbing lots of points on the last level and getting out safely than before. I WAS WRONG. It's still extremely effective, you just need to make sure to save up enough energy. I've nerfed it a little: it now spawns 4 enemies instead of 3 on acquisition, and it can no longer be used to instantly siphon (since while that was a neat interaction it's a more interesting problem to find a safe space to siphon while stepping through a level full of enemies). This means it's now harder to get a high score because you're more likely to have to deal with the enemies you alert, and it's harder to get a streak because you're less likely to be able to skip the hardest levels.

I wrote before about the question of whether .SCORE was overpowered. I don't believe that it technically was because I was able to get scores as high without it as with it. But many others did believe that it was and that belief was shaping the way they played in a way they were unhappy with - it doesn't always matter if a belief is true if it has genuine effects anyway. This belief reinforced itself: players convinced that it was the only way to get high scores would only try to get high scores when they had it, so the high score list filled up with .SCORE-based scores, convincing anyone else looking at it that this was the way to get high scores. James Lantz describes the process of constantly restarting the first level to get a good run - this is mostly a problem with .SCORE because it is clear from the first level whether you'll be able to get many points out of it, whereas other ways of scoring big aren't obvious until later. Meanwhile I would almost never pick up .SCORE myself because it just wasn't worth the risk. Also I noticed that many of the high scores weren't quite as high as they could have been because they'd forget to dump their remaining energy into .SCORE on the last level when it was safe - which is admittedly a boring requirement. I'm attempting to fix these concerns in a way that remains balanced by making .SCORE grant one less point with each use, while spawning one less enemy on acquisition. This means it's harder to get high scores with it (you'll need to take even more risks or rely on other methods of scoring as well), every use of it is an interesting gamble (you can't use it safely at the end because it's worth 0 by then), it less encourages restarting lots to get a good run (because it has less variance), and it's more viable during streaks (because it's less dangerous to pick up).

SPOILERS: There is a secret level. It's expensive and risky to get there but sometimes it can be exploited for an advantage. There was a design error which is now fixed, in which it was possible to recursively access it and maybe push that advantage too far. It remains a possible way to get extra points, but not so many as before.

SEMI-SPOILERS: During streaks difficulty modifiers get applied to provide an extra challenge and hopefully end the streak eventually. I didn't want streaks to continue indefinitely because then the optimal way to play would be to get one point per game for a thousand games rather than taking any risks, so I increased the chances of failure as you go.
(It seems like bad form to just present an impossible run, but with a finite discrete game you can't just keep increasing the difficulty without it ending up impossible. The best I can hope for here is that when you eventually lose it'll be because of a risk you took that you could have done differently: this feels less unsatisfactory even though in the limit it's still outside of your control.)
When I made the game I was enamoured with the idea that a streak might last a very long time, that there could be some elite hacker known only by a mysterious pseudonym somehow *STILL RUNNING* against all odds. And this happened exactly as imagined, nobodyweknow has an amazing streak, 1228 points in 47 games at time of writing. But there's a side-effect to this that I hadn't anticipated: it's putting other players off even trying for streaks, leaving them just grinding at the (less interesting) single-game scoreboard. Additionally, nobodyweknow has said that at the level he's playing at some of the runs seem pointless because they're not offering enough challenge - the difficulty modifiers, originally conceived just as a way to kill off streaks, have become a core part of what makes the game varied and interesting. So I've significantly ramped up the difficulty increase to hopefully end streaks quicker and more consistently challenge experts. This has a more dramatic effect on scoring than the single-game changes because of how much the points can add up across a streak.

All of these combine to make scoring more difficult, but in ways that are somewhat ambiguous - I can't just recalculate old scores to fit the new scale. There may only have been one score that exploited the old secret level (my own) but I can't be sure, and certainly there were many that exploited the old .STEP and .SCORE. Probably all the old scores would be beaten eventually even though it's harder - just by someone eventually being more skilled and having a luckier run - but until then they'd be hanging around in the lists and I worry this would seem unfair to new players. Also it matters because you can examine what progs a score was obtained using, so this would be providing misleading information to anyone hoping to learn from it (similarly, you can examine what modifiers were applied to games in a streak). So all up I think it is best not to have the old scores hanging around. I hope this is okay with everyone!