Advice on designing a game with speed running in mind

Edit history:

Scepheo: 2013-11-10 08:29:12 am

2013-11-10 08:25:26 am

Quote from Blink:

I'm not sure on the specifics, but if you can make it Hourglass compatible that will help people to TAS it. Never a bad thing Smiley

If anyone knows how to make a game Hourglass compatible, pray tell. As far as I know it means (amongst other things):
- No mouse.
- No 3D (so surface blitting, not quads).
- Whatever window you're given on game start, use that and nothing else.
- Single thread.
- No fancy DLLs, but use kernel32.
I might be wrong, but that's things that seem to cause trouble for HourGlass.

As for something useful: the quick level restarts/save reloading also means you don't want any level-intros (eg a fly-by towards the player). They're fine when you enter the level or something, but don't show them everytime the player restarts.
EDIT: Oh, and test. A lot. You'll want to be sure you're okay with whatever glitches speedrunners may find after you release. Runners hate patches. If you do have them, make it possible to revert (so no online server that tells you you need to update before you can play).
I JUST THOUGHT OF THIS: No mandatory internet connection. Those suck. If you're releasing through Steam, having it run without Steam is a big plus.

[quote="Blink"]I'm not sure on the specifics, but if you can make it Hourglass compatible that will help people to TAS it.  Never a bad thing :)[/quote]
If anyone knows how to make a game Hourglass compatible, pray tell. As far as I know it means (amongst other things):
- No mouse.
- No 3D (so surface blitting, not quads).
- Whatever window you're given on game start, use that and nothing else.
- Single thread.
- No fancy DLLs, but use kernel32.
I might be wrong, but that's things that seem to cause trouble for HourGlass.

As for something useful: the quick level restarts/save reloading also means you don't want any level-intros (eg a fly-by towards the player). They're fine when you enter the level or something, but don't show them everytime the player restarts.
EDIT: Oh, and test. A lot. You'll want to be sure you're okay with whatever glitches speedrunners may find after you release. Runners hate patches. If you do have them, make it possible to revert (so no online server that tells you you need to update before you can play).
I JUST THOUGHT OF THIS: No mandatory internet connection. Those suck. If you're releasing through Steam, having it run without Steam is a big plus.

Patashu

Edit history:

Patashu: 2013-11-10 04:02:20 pm

2013-11-10 04:01:41 pm

Quote from Scepheo:

Quote from Blink:

I'm not sure on the specifics, but if you can make it Hourglass compatible that will help people to TAS it. Never a bad thing Smiley

If anyone knows how to make a game Hourglass compatible, pray tell. As far as I know it means (amongst other things):
- No mouse.

Correct, until Hourglass supports mouse fully (it doesn't yet, Warepire is working on it but only has so much time to devote to it). The game can allow for mouse control, but it needs to be fully playable with keyboard/controller as well (Braid, for example)

Quote:

- No 3D (so surface blitting, not quads).

Are you sure? I see lots of Touhou TASes using hourglass on nicovideo, and touhou has those 3D animated backgrounds. (Though I do know that Hourglass only supports so many 3D drivers/engines.)

Quote:

- Whatever window you're given on game start, use that and nothing else.

Correct.

Quote:

- Single thread.

I think multi-threading is OK but the gameplay itself needs to be single threaded (or something like that).

Quote:

- No fancy DLLs, but use kernel32.

Not informed enough to corroborate/refute.

Quote:

I might be wrong, but that's things that seem to cause trouble for HourGlass.

One more I can think of off the top of my head:
-Needing an internet connection to start up (If it's a Steam game, then it should be able to run outside of Steam, such as Braid or Super Meat Boy which have both been made to work in Hourglass)

[quote="Scepheo"][quote="Blink"]I'm not sure on the specifics, but if you can make it Hourglass compatible that will help people to TAS it.  Never a bad thing :)[/quote]
If anyone knows how to make a game Hourglass compatible, pray tell. As far as I know it means (amongst other things):
- No mouse.[/quote]
Correct, until Hourglass supports mouse fully (it doesn't yet, Warepire is working on it but only has so much time to devote to it). The game can allow for mouse control, but it needs to be fully playable with keyboard/controller as well (Braid, for example)
[quote]- No 3D (so surface blitting, not quads).[/quote]
Are you sure? I see lots of Touhou TASes using hourglass on nicovideo, and touhou has those 3D animated backgrounds. (Though I do know that Hourglass only supports so many 3D drivers/engines.)
[quote]- Whatever window you're given on game start, use that and nothing else.[/quote]
Correct.
[quote]- Single thread.[/quote]
I think multi-threading is OK but the gameplay itself needs to be single threaded (or something like that).
[quote]- No fancy DLLs, but use kernel32.[/quote]
Not informed enough to corroborate/refute.
[quote]I might be wrong, but that's things that seem to cause trouble for HourGlass.[/quote]
One more I can think of off the top of my head:
-Needing an internet connection to start up (If it's a Steam game, then it should be able to run outside of Steam, such as Braid or Super Meat Boy which have both been made to work in Hourglass)

Warepire

Edit history:

Warepire: 2013-11-10 05:20:39 pm

2013-11-10 05:20:16 pm

Heavy Metal Powered

Don't specifically target Hourglass, just don't. It's not ready by far to handle ... lots of things. The only thing you should do in terms of targeting the TASing audience is making the game Win32, and no mandatory internet connection. The rest should be solved on the Hourglass side of things instead.

Ghostwheel

2013-11-10 06:36:42 pm

The artist formerly known as Qxy

One thing I don't think has been mentioned yet is difficulty. I think at the end of the day, speedrunners are going to run the games that they love, that made an impression on them, and that can come from a number of different factors. But people who are already veteran speedrunners looking for a challenge in a new game might be more likely to pick up the more difficult option. Additionally, I've found there to be a correlation (not necessarily causal) between the difficulty of a game and its speedrunability. The harder the problem, the more creative you need to be in your solution. Of course, speedrunning makes any game, easy or hard, more difficult because of the asymptotic nature of the goal. Most RPGs that have an easier level curve in a casaul play can become monstrous when you skip all the fights.

In the end, if you want to see if its good to speedrun, try routing it before you release it. If it seems difficult to route well, then people may have fun cracking it.

seagaia

2013-11-11 10:53:52 am

Hey guys, thanks again for replying to this, it really does help.
Sean

ais523

2013-11-12 07:52:09 am

Some more things I thought of, about timing. Your in-game timer should stop during cutscenes (which should also be skippable, but if someone chooses to watch them for whatever reason, they shouldn't be penalised). Timing should also stop accurately on saving (as in, if you save then load again, you should have the exact same time as if you didn't do the save and load); and the timer should stop during loading screens (so that people don't get an advantage from faster machines).

LotBlind

Edit history:

LotBlind: 2013-11-20 04:56:28 pm

2013-11-20 04:52:45 pm

I like to think of glitches like this: They're actually unintended features. When you "fix" them you're actually removing a feature. So you have to decide if it's a good feature or not but unintentional was never a synonym for unwanted. Bunny hopping was unintended...

Lots of really good suggestions have come up already. Here's one more: mapping. If you make maps of the game areas available to the player (after unlocking them or so) it saves them the trouble of having to map the game out themselves.

There's games with a dedicated "speedrun mode" otherwise called time attack. In addition to including the timer, this would automatically remove cutscenes and the story by default and so runners don't have to even hold a button to skip those and it also means the ground is as even as it can be for first-time runners and veterans alike.

CBenni

2013-11-20 05:51:03 pm

- Ingame timers, allowing even to see and read it in realtime from outside, so it doesnt get recorded using screen cap, but can be put on stream.
- Save the times for individual levels, both for each level as well as tracking the best SS full playthrough
- Demo recording + playback, comparable to source engine games.
- Save the glitches. (Dont patch stuff that doesnt destroy it for non-speedrunners)
- Make it only so much RNG dependant
- Allow some fast-movement method. Not having sprint and crawling through levels makes it very unappealing
- Skippable cutscenes!
- Quicksaves
- Dev console that gives a variety of commands that help both debugging the game and testing strats. For example, noclip, god mode, etc
- Prevent softlocks by not having triggers rely upon each other unless necessary
- Level selection with best times
- If applicable, completion percentage display for 100%/low%

I might find more but thats it for now.
Good luck!
cu, CBenni

Rahix

2013-11-24 07:29:34 am

Game design! Really I like it.

seagaia

2013-12-12 04:28:43 pm

Hey, could someone explain what all the terms in this video mean and which ones are most useful for an in-game status?

Se:, Be: , P.Se: ,P.Be: , Max Ord. , Live, B. Delta

I'm not sure what Max Ord. means but I think I have the others figured out?:

As far as I can tell, Se = Segment = best time in the current segment *within* a the best previous overall time, Be = best time in current segment regardless of whether it was int he best overall run, P.se, P.be are the same but for the previous segmeent, Live and B. Delta are how well the previous segment was done relative to P.se and P.be.

The white numbers on top are the times that segment was finished in the best recorded run, when you finish a segment, it shows what time your timer was at, and then the time diffrence relative to the best recorded run's time at that point.

the big green number is the overall timer for the current run, small green is segment time.

---

So under those definitions, right now for my in-game display I have the overall timer, P.Be (shown individually for each segment), previous best overall time.

When a segment finishes, the display goes from showing P.be and current segment time to B.delta and current segment time.

The record display then currently shows best overall time, then all of the Be. times.

---

What's the most important information to show? Should I just copy more or less what's in the video and that will be sufficient?

Hey, could someone explain what all the terms in this video mean and which ones are most useful for an in-game status?

Se:, Be: , P.Se: ,P.Be:  , Max Ord. , Live, B. Delta

I'm not sure what Max Ord. means but I think I have the others figured out?:

As far as I can tell, Se = Segment = best time in the current segment *within* a the best previous overall time, Be = best time in current segment regardless of whether it was int he best overall run, P.se, P.be are the same but for the previous segmeent, Live  and B. Delta are how well the previous segment was done relative to P.se and P.be.

The white numbers on top are the times that segment was finished in the best recorded run, when you finish a segment, it shows what time your timer was at, and then the time diffrence relative to the best recorded run's time at that point.

the big green number is the overall timer for the current run, small green is segment time.

https://www.youtube.com/watch?v=cxfSF4VPtuc

---

So under those definitions, right now for my in-game display I have the overall timer, P.Be (shown individually for each segment), previous best overall time.

When a segment finishes, the display goes from showing P.be and current segment time to B.delta and current segment time.

The record display then currently shows best overall time, then all of the Be. times.

---

What's the most important information to show? Should I just copy more or less what's in the video and that will be sufficient?

MrWalrus

2013-12-14 03:02:03 pm

I am a walrus

Sufficient is just having an accurate in game timer per stage and one for the entire game.
Preferably one that doesn't count loading times (Although they wont matter in a 2D game really.)
And tracking it against your previous full run stage times and personal best times per stage.

In short track full run times and individual stage times.

Keep the ingame timer simple rather than implementing everything in Llanfair.

Also overall, just make the game enjoyable and people will run it.
Feels kinda cheap if there's everything ready for a speedrun, so make the game challenging.

I have no name

2013-12-14 03:33:09 pm

Obscure games ftw

If people want the extra data they could use Llanfair to get it anyway-just putting a timer into the game that's accurate is enough to make it friendly: people like having their splits at varying points anyway Tongue

The only thing I might say would be nice to include would be ILs (time trials on levels) and if you beat your best in a run, the game tells you this.

Se: This is segment, so the length of the split in the run being compared against.
Be: This is best segment, so the best that the runner has ever done.
P Se and P Be are the same, but for the previous segment.
Max Ord is the height of the graph that shows the progress of the run as it's gone along.
Live and Best Delta are the relation to P Se and P Be.

TheBarrel

2013-12-14 06:57:03 pm

It's weather time!

To get things on somewhat familiar territory I guess, Inspiration Dave's individual level timer was simple and good. It told you an accurate time, differentiating between 100% and regular completion. It started when you gained control and ended when you lost control. Not too complicated. That's all a timer has to be IL or SS.

Like the last two comments mentioned I wouldn't worry about all the Llanfair stuff. Every runner's gonna do something different even if you implement it anyway TBH.

Xitoshi

2014-01-17 09:25:31 am

I think it's been already mentioned, but I recommend that it can have a good range of categories and even a "bingo" possibility (a series of individual objectives optional for a full run, so people can make races in which they have to select and fulfill some of this objectives, more info at SpeedrunsLive.com).

That would likely maintain interest in the game for a long time, I've seen games abandoned because of the impossibility of improvement (which in this case is also linked to randomness elements and HUGE skipping glitches).

seagaia

2014-03-27 04:19:54 pm

FYI I put up a demo for EtO today. There's one area which is the "gauntlet" type level I mentioned, but you can sort of get a feel for how this would be spedrun.

Info here: eventheocean.com
You can get the demo with any purchase of the most recent humble bundle sale - humblebundle.com/weekly

Or through my itch.io page (with this secret link - please don't share!)

http://seanhogan.itch.io/even-the-ocean-motion-demo-1/download/ftH1aU5kG5MVRvjTTs2xqP9BE9QDhoAt85882T31

Would love to know what you guys think. Lemme know if anyone's streaming it.

seagaia

Edit history:

seagaia: 2016-11-20 10:56:03 am

2016-11-20 10:51:37 am

years later the game's done! thanks for all the suggestions.

(I made a speedrunner coupon sale for 30% off if anyone wants to try: https://seanhtch.itch.io/etospeedrunnersale)

(otherwise there's steam) http://store.steampowered.com/app/265470

there have been a few speedrunners already taking stabs at it, too!

https://www.twitch.tv/lifning/v/102201140

LotBlind

2016-11-23 02:14:03 am

Grats on both finishing the game and getting a runner interested, let alone multiple runners!