Guide for game developers who want to make their games speedrunner-friendly

2016-09-22 08:11:30 pm

Recently, there's been a spate of developers asking for advice on how to make their games better for speedrunners (including a post on here on SDA within the last week, and several posts on other speedrunning websites recently). Because I'm one of the people who often ends up answering these things, I've been trying to compile a comprehensive guide full of advice on what sorts of changes make a game better to speedrun. Some of it's general advice about what sort of game elements make speedrunning more fun (e.g. no long waiting times); some of it's warning developers against common mistakes that are easily fixable and cause large problems for speedrunners (e.g. overly aggressive autosaves that prevent segmented runs and practice, including load times in the timer); and some of it's minor points that won't have a huge impact on most speedrunners, but which will make things slightly easier and don't cost the developer much to implement (e.g. some technical advice that makes games play nicer with OBS and with autosplitting programs).

My current draft of the advice is here: https://kb.speeddemosarchive.com/User:ais523/Making_your_game_speedrunner-friendly
This (probably) isn't a final version, and I'm interested in any advice the other speedrunners (and/or game developers?) here could give me about the document; is there anything wrong? Anything I've left out? Anything that could do with clarification?

Once this is finished, it's likely to end up as a sticky somewhere, so that we have something we can point to the next time the question comes up.

Thread title:

Recently, there's been a spate of developers asking for advice on how to make their games better for speedrunners (including a post on here on SDA within the last week, and several posts on other speedrunning websites recently). Because I'm one of the people who often ends up answering these things, I've been trying to compile a comprehensive guide full of advice on what sorts of changes make a game better to speedrun. Some of it's general advice about what sort of game elements make speedrunning more fun (e.g. no long waiting times); some of it's warning developers against common mistakes that are easily fixable and cause large problems for speedrunners (e.g. overly aggressive autosaves that prevent segmented runs and practice, including load times in the timer); and some of it's minor points that won't have a huge impact on most speedrunners, but which will make things slightly easier and don't cost the developer much to implement (e.g. some technical advice that makes games play nicer with OBS and with autosplitting programs).

My current draft of the advice is here: https://kb.speeddemosarchive.com/User:ais523/Making_your_game_speedrunner-friendly
This (probably) isn't a final version, and I'm interested in any advice the other speedrunners (and/or game developers?) here could give me about the document; is there anything wrong? Anything I've left out? Anything that could do with clarification?

Once this is finished, it's likely to end up as a sticky somewhere, so that we have something we can point to the next time the question comes up.

xud9dab2

2016-09-23 05:30:58 am

heh heh heh heh heh heh heh heh heh

Low% RPGs can be made significantly less frustrating if an "earned exp turns into G/Gold/money/whathaveyou" accessory can be obtained early, at least before any bosses/other forced encounters that would give out exp. That way it doesn't become a game of juggling math and party members to ensure that levels are kept as low as possible. As a bonus that could mitigate the need to go out of one's way for extra money to buy the necessary equipment/items/what have you that a low level party is going to need.

ais523

2016-09-23 06:36:31 am

Quote from xud9dab2:

Low% RPGs can be made significantly less frustrating if an "earned exp turns into G/Gold/money/whathaveyou" accessory can be obtained early, at least before any bosses/other forced encounters that would give out exp. That way it doesn't become a game of juggling math and party members to ensure that levels are kept as low as possible. As a bonus that could mitigate the need to go out of one's way for extra money to buy the necessary equipment/items/what have you that a low level party is going to need.

Thanks for the advice. There's actually a more general point here, that doesn't apply exclusively to RPGs, that I'd missed: you don't want to force upgrades on players for doing things that they'd do naturally during the game. For example, Zelda games drop heart containers after each boss but typically don't force you to pick them up, and Final Fantasy games give out zero experience from (most) forced encounters.

Omnigamer

2016-09-23 09:38:26 am

All the things

I haven't read it in full, but I believe in at least a few of your points you are giving advice that is overgeneralized towards the execution of speedruns. For example, your section on randomness makes some pretty big assumptions that target explicitly grinding out runs. There are far more reasons why a developer might include random elements in their gameplay, and why all different types of runners would benefit from some such elements. I for one think a game without any luck whatsoever is a pretty dull game to run, to say nothing about it being a dull game.

Justin_SkyTy

2016-09-23 12:58:02 pm

Thanks for this. One question we've been discussing recently for development on our Metroidvania is which items should be tracked toward completion, particularly treasure chests. Major upgrades and equipment items will definitely count toward 100%, including those that can be bought. However, treasure chests only give the player gold once and can thus be tracked toward completion, but are much more common throughout the game, (think equivalent to every chest in a Zelda game that has rupees).

We're currently leaning toward keeping them as a separate count, so you can see how many you've opened, but can get 100% completion as far as any meaningful items are concerned without finding them all. However, chests are also a good opportunity to use as extra challenges by putting them in places the average player might not figure out how to get to, without making them miss out on anything for not finding them, making 100% more interesting to obtain. Also, if map completion is tracked as part of 100%, the player would need to reach every room anyway.

In short, we feel that tracking treasure chests separately would allow more flexibility for what an individual player (or speedrun category) considers "complete," but we're open to other possibilities. Any thoughts?

Thanks for this. One question we've been discussing recently for development on our Metroidvania is which items should be tracked toward completion, particularly treasure chests. Major upgrades and equipment items will definitely count toward 100%, including those that can be bought. However, treasure chests only give the player gold once and can thus be tracked toward completion, but are much more common throughout the game, (think equivalent to every chest in a Zelda game that has rupees).

We're currently leaning toward keeping them as a separate count, so you can see how many you've opened, but can get 100% completion as far as any meaningful items are concerned without finding them all. However, chests are also a good opportunity to use as extra challenges by putting them in places the average player might not figure out how to get to, without making them miss out on anything for not finding them, making 100% more interesting to obtain. Also, if map completion is tracked as part of 100%, the player would need to reach every room anyway.

In short, we feel that tracking treasure chests separately would allow more flexibility for what an individual player (or speedrun category) considers "complete," but we're open to other possibilities. Any thoughts?

ais523

Edit history:

ais523: 2016-09-23 02:32:54 pm

2016-09-23 02:26:20 pm

It's OK to have more than one percentage counter if you're not sure what players would find makes an interesting category. (And if you're particularly lucky, both "all upgrades" and "all chests" will make interesting categories.) Doing that's probably better than having one "main" counter.

Map completion tends to make for a poor 100% run. Having a counter for map completion is still helpful, but mostly because it'll appeal to the more obsessive of your casual players (I admit to having put effort into map-completing Metroidvanias in the past…) than because speedrunners will like it. In most of the games I know of which have a map-completion counter, speedrunners disregard it and use something else for the 100%.

EDIT: Or to put the idea another way: "all major upgrades" is a category that's often run even in games where the 100% is different, so having an easily accessible counter for that will help speedrunners verify that they haven't missed anything. Meanwhile, categories that are equivalent to "all chests" sometimes get run even when skipping upgrades. I'm reminded of the "minimum upgrade max%", which is basically about skipping all the useful upgrades you can, and (under those conditions) collecting all the useless or marginal upgrades or optional goals that you can, which is almost never used for a speedrun but which is a common casual goal for routers when they're bored and often teaches you a lot more about the game. You might also be amused at this TAS which focuses on getting all the cosmetic upgrades in a Metroidvania while getting only any% of the critical ones, and which IIRC was chosen to add some order and any% feeling into what would otherwise have been a glitchy mess.

It's OK to have more than one percentage counter if you're not sure what players would find makes an interesting category. (And if you're particularly lucky, both "all upgrades" and "all chests" will make interesting categories.) Doing that's probably better than having one "main" counter.

Map completion tends to make for a poor 100% run. Having a counter for map completion is still helpful, but mostly because it'll appeal to the more obsessive of your casual players (I admit to having put effort into map-completing Metroidvanias in the past…) than because speedrunners will like it. In most of the games I know of which have a map-completion counter, speedrunners disregard it and use something else for the 100%.

EDIT: Or to put the idea another way: "all major upgrades" is a category that's often run even in games where the 100% is different, so having an easily accessible counter for that will help speedrunners verify that they haven't missed anything. Meanwhile, categories that are equivalent to "all chests" sometimes get run even when skipping upgrades. I'm reminded of the "minimum upgrade max%", which is basically about skipping all the useful upgrades you can, and (under those conditions) collecting all the useless or marginal upgrades or optional goals that you can, which is almost never used for a speedrun but which is a common casual goal for routers when they're bored and often teaches you a lot more about the game. You might also be amused at [url=http://tasvideos.org/2448M.html]this TAS[/url] which focuses on getting all the cosmetic upgrades in a Metroidvania while getting only any% of the critical ones, and which IIRC was chosen to add some order and any% feeling into what would otherwise have been a glitchy mess.

xud9dab2

2016-09-23 02:58:15 pm

heh heh heh heh heh heh heh heh heh

Quote from Omnigamer:

I for one think a game without any luck whatsoever is a pretty dull game to run, to say nothing about it being a dull game.

When a game is so deterministic that speedrunning it is pretty much a glorified rhythm game of "hit the right button(s) at the right time for the next half hour", that's indeed not a good thing. Of course running into a situation where you have to pass a completely random (or as completely random as computer RNGs can be, at any rate) 10% 2½ hours into a run, with each failed attempt setting you back 2-3 minutes, isn't good either.

The important thing with RNG is to make sure that you can use skill to compensate for bad RNG (random enemy movement in a metroidvania game puts an otherwise avoidable enemy right in your path, but you can use fast reflexes to shoot it before it damages you), and that RNG doesn't become a major roadblock too long into a run both in terms of chances and time lost on failure.

Quote from Justin_SkyTy:

In short, we feel that tracking treasure chests separately would allow more flexibility for what an individual player (or speedrun category) considers "complete," but we're open to other possibilities. Any thoughts?

Separate trackers would be best. 100% will be defined as it will be defined, even if the game wants to call it something else entirely, so keeping track of different types of collectables separately will only make it easier for runners to keep track of what they consider important and/or unimportant ultimately.

[quote="Omnigamer"]I for one think a game without any luck whatsoever is a pretty dull game to run, to say nothing about it being a dull game.[/quote]When a game is so deterministic that speedrunning it is pretty much a glorified rhythm game of "hit the right button(s) at the right time for the next half hour", that's indeed not a good thing. Of course running into a situation where you have to pass a completely random (or as completely random as computer RNGs can be, at any rate) 10% 2½ hours into a run, with each failed attempt setting you back 2-3 minutes, isn't good either.

The important thing with RNG is to make sure that you can use skill to compensate for bad RNG (random enemy movement in a metroidvania game puts an otherwise avoidable enemy right in your path, but you can use fast reflexes to shoot it before it damages you), and that RNG doesn't become a major roadblock too long into a run both in terms of chances and time lost on failure.

[quote="Justin_SkyTy"]In short, we feel that tracking treasure chests separately would allow more flexibility for what an individual player (or speedrun category) considers "complete," but we're open to other possibilities. Any thoughts?[/quote]Separate trackers would be best. 100% will be defined as it will be defined, even if the game wants to call it something else entirely, so keeping track of different types of collectables separately will only make it easier for runners to keep track of what they consider important and/or unimportant ultimately.

Omnigamer

2016-09-23 06:21:14 pm

All the things

Right, there are good ways to apply RNG to give dynamism to a game without making it binary pass/fail on continuation. An example I like to point to is with boss cycles. Making a boss's pattern unpredictable is one thing; planning for a variety of events is all part of the routing and practice process. But if one or more of those potential patterns makes the boss invincible or otherwise unhittable, that halts progression. A better practice would be to instead make the boss vulnerable, but having it be risky to attempt. This still gives the player (any player, casual or speedrunner) an option of whether to take the risk and additional challenge of damaging the boss during such a phase.