I had never even noticed it (and I didn't get an e-mail notification about this thread... -_-). Thanks!

[...] I haven't actually timed any runs for a long time, I do have the idea that maybe timing my runs will make me have more fun and enjoy them more [...] [...] My goal and how I see it right now is to be a top runner for a game [...]

In my opinion, these two things don't work well together. If you want to be a top runner, and thus to keep improving your time, you must be able to gauge how much you are improving. It seems only natural to me that you'd lose interest in practicing by not seeing any returns from it.

You shouldn't submit your time to the leaderboards if you aren't happy with it. However, you should definitely keep a timer up when doing runs, if only to have a ball park estimate of whether you are doing better or worse than before (if you don't want to have splits, or feel overwhelmed by the idea of having them).

Regardless of that, I have some harsh advice... I don't feel like trying to be a top runner is a healthy goal at all. If you aren't enjoying the path (practicing/doing lots of runs to get better), then you definitely won't enjoy when/if you reach your goal. I have the impression that most people go about this the other way around: they play a game so much that they end up running it. Heck, I ran a game for 5 years before I decided to set up its leaderboard (though I'm probably a really strange person).

So... maybe stop trying to actively speedrun a game and start trying to find a game that you really enjoy playing. For example, something that I enjoy outside of speedrunning is playing Mega Man 2. I did learn Metal Man's, Bubble Man's and Heat Man's zips (though I only try Bubble's on Legacy, with the rewind function), but other than that I just play the game as usual.

Pretty much what xDrHellx said.

Try to find your motivation for speedrunning. Do you want to simply improve your time? Do you want to try to get a certain position (WR/top 10/etc)? Do you want to run at marathons? Or do you simply want to have fun replaying a game and possibly improve your time?

Speaking about myself, I mostly find playing game fun, and like to try to improve my own times. So, in a sense, I don't really practice any of the games I run. When I'm feeling like it, I'll simply attempt to do runs and see how, and if, I improved. I'll of course practice at times, but for some specific objective: to get consistent at a trick, to learn how to react to something etc. For example, right after an attempt, or at failing an attempt, I may practice a trick that I know I fail a lot a few times. However, since I take a more casual approach to speedrunning, I don't feel a need to force myself to grind to get my time down.

I agree that interacting with the games' communities is a good idea. Even outside of races, just being able to interact with fellow runners of the same game is really fun.

Looking at the games that I run (since those are the runs I'm most familiar with), I have a metroidvania suggestion and a weird one.

My first suggestion is Momodora: Reverie Under the Moonlight, a 2D metroidvania. You play as the priestess Kaho, who's tasked with dispelling a curse, or something, with a maple leaf and a Bow. Also, she eventually becomes a cat! I really liked the game casually (I think my first playthrough took me around 4 hours), and eventually decided to try to learn the any% run to take part of a tournament. There's some RNG in the bosses, but the run is mostly based on execution. Sub-30 minutes times aren't that hard with some practice, and although you do need to beat every boss without taking any damage to get a really good time, there are pretty reliable strategies for most bosses, and I think you lose at most 10 minutes by failing that on every boss. I've since stopped running this game, but I still find it quite fun to play, and will do occasional "casual runs". Oh, and there's quite an active discord for the series.

Now, for the weird suggestion... Have you heard about Catherine? It's a puzzle game about climbing a tower by pushing/pulling blocks to make stairs. The main game is interesting, but I don't enjoy the story that much... However, there's an extra mode called Babel, in which you have to climb a pseudo-random tower, either solo or in pairs. Most of the runs are done in the pairs mode (which is surprisingly easier than solo; and I heard there was a bug in the solo mode), and played in 1p2c fashion. I pretty much learned this run by playing the game a lot, after watching some marathon runs. Also, the ILs are way more active than full game, so there's also that. This game can be quite stressful at times (so yeah, maybe a bad pick), but there's something about learning its patterns and getting better at climbing that I really enjoy about it. The PC version (Catherine Classic) is kinda easier for 1p2c since you can remap the controls in the version, mapping one character to a half of the controller and the other to the other half. There's also a newer version of the game on the PS4.

That's about it from me. Youyou Kengeki Musou (another game I run, a top-down Touhou action platformer) is quite fun, but I don't know if it fits the description you asked. It's kinda short and really fun to play, though... So maybe check that out.

Good luck looking for speed games! :)

My game is called JJAT. It has some metroidvania inspirations, but it's too short and has way too few paths to be called an actual metroidvania (the 100% category is just under 5m 50s).

I'll definitely check your run out, when it's up here. :) But I must admit that for me a ~2 hour run is currently too long... the longest game that I run is ~40 minutes long (in a worst case scenario).

@polarityflow, I did exactly that, and I know of at least two other people who run their own games, as Oreo321 has already said.

So... yeah, go for it. :)

Urg... why is that in the layout, instead of being in the general options?

Thanks a lot!

Have Livesplit hidden behind the game

How exactly would someone do this? For me (on Windows 7) LiveSplit always stays on top of every other window, and I couldn't find any option to change that behaviour...

Alright, I've changed it back to keep the timer running on the main menu!

the plugin would sign each frame with the current time

This got me curious... By 'sign' did you mean HMAC, using the timestamp as the symmetric key?

But anyway, as everyone said, this method wouldn't work...

And honestly, this doesn't seem feasible at all. :(

The best I can think from a game dev perspective would be to make each run unique. By that I mean, the game would have to be deterministic, except by seeded pseudo-random events (that does not affect gameplay) that would allow verifying that the run is was done as a single segment.

For example, imagine a game that has some pseudo-random particles (that, let's say, is seeded by user input as well). You could require inputs to be submitted alongside video proof, and then compare if the video submitted is identical to the one generated by running the game with the submitted inputs... This way, spliced runs would break these particles and be analytically verifiable.

First issue: what about false negatives caused by video quality?

Even ignoring that and the fact that this would require developing the game, how do you verify that the inputs were legitimately generated by a human (instead of, say, having been TAS'd)? You could verify that the keyboard inputs weren't virtually generated (which is a thing on Windows), but then tools like joy2key would have to be banned. Also, if any form of gamepad were allowed, how do you verify that a valid gamepad was used? How do you verify that a modded gamepad (if you could identify one at all) isn't an input method specialized for accessibility?

The only way (which is a pretty big stretch) I could see this working would be by having a certified console (or even, making PC environments certifiable), with certified input methods, that in some way can't be tampered at all, otherwise it bricks.

Also, about the final split...

I just got a 15:06.79, which made me wonder how precise my auto splitter actually is... I advanced that video frame-by-frame, and I believe it's possible to see a "small pink blob" 1 frame after the timer stops. If not that, it's plainly visible 3 frames after the timer stopped.

Here's a 1-FPS GIF showing the timer stopping and the explosion starting: https://drive.google.com/open?id=1T0QzAiLoXSn6R8trXDOD_Pg4y9wPFbnP (hopefully not auto-embedded, as it's close to 1 MB...). I also rendered these frames into a MP4, since that has better quality than GIF: https://drive.google.com/open?id=1vri0bV1C6jt1i8IJ3Fk9SGc7Z255mpfL.

Would you all rather I try to adjust it further or is this good enough?

I don't have any really strong argument for removing time spent on the main menu. As far as I could tell, only Sotena has run that category, so I'd be willing to retime it and report the new time, without the time spent on the main menu, if we were to keep the auto-splitter as is...

My main reasoning for removing it is that there's nothing happening in the main menu, from a gameplay perspective. One could argue that menuing is part of games, but I can't really agree with that for an action game...

Also, I think time spent on the main menu is kinda similar to time spent on the loading screen. It feels weird to me that we remove time spent on the loading screen (which make sense, since it varies depending on the player's computer) but not time spent on the main menu (which I don't whether entering/exiting it varies)...

I tried to look for similar games to see how others do it, but the closest I could find were Celeste, which uses per-stage IGT, and A Hat Time, which does consider time spent on the hub world, but the hub world also plays just like the rest of the game... So yeah, that wasn't of much help...

Anyway, I could change it keep the timer running on the main menu. Whichever is better for everyone! :)

I've been on and off trying to find a perfect game (for me) to speedrun, get good at, stream and generally just have fun with. ... [...] even though I'm absolutely determined to pour my heart out into a game and be competitive with my times.

Before finding a game to speedrun, I think you should really try to figure out why you want to speedrun at all. Is it because you want to compete? Because you enjoy so much a game that you just want to keep playing it again and again? Because you really enjoy something about a given game (say, a movement tech) and/or community (e.g., how everyone is really nice, even just hanging out) and want to connect to others (through streaming, racing, the site, local events etc) for that? Because you really enjoy a game and want to scream at the world "Hey, this is fun!"? Or maybe you really like the concept of speedrunning and want to be part of the speedrunning community?

In my case, I kinda started speedrunning so I could compete against others...

For a long time, I enjoyed watching speedruns but didn't attempt running anything myself. The closest I got to actually speedrunning anything was beating time attack challenges (for achievements) and playing my own games as fast as possible (after countless hours of playtesting, you kinda naturally start to speedrun stuff). However, as much as I liked the idea of speedrunning, I didn't fancy actually practicing (nor, maybe, grinding) for better times. However, after watching AGDQ 2017's Mystery Game Tournament I learned about blind racing and eventually started to join other blind race tournament (mostly, SRL Mystery Game Tournament). This eventually led me to joining a Momodora Reverie Under the Moonlight tournament, which became my first "official" speedrun (or, at least, my first speedrun here on this site).

I kept playing and trying to improve my time in that game for a long while, but nowadays I mostly "speedrun it casually" (heck yeah, that is a thing... I guess?). After that, I picked two other games (Catherine and JJAT) simply because I really enjoyed playing them and they have built-in timers. So, since I would usually go back to these games and try to get better times at them, speedrunning and submitting times to the leaderboards kinda came naturally. In particular, JJAT was an interesting experience because I had to set the game's leaderboard and I had to write its guide. Only recently I've picked a game (Youyou Kengeki Musou) because I wanted to play it more and the run seemed interesting and fun.

So, I'd say that for me speedrunning comes sort of naturally: I just find a game that I really enjoy and want to play more of it. And honestly, I'm not even that competitive... I mostly submit times to document it, I guess? Sort of a "Hey, I got this time in that game and I'm proud of it!"... Heck, twice now I've submitted times in games with few runners that got me the last place.

When it comes to streams at the very least I'm looking for something that won't have me end up with 0 viewer streams since that's quite demoralizing.

My take on this is quite different. When I'm blind racing, I can't look at the chat, so I got used to simply ignoring whatever is happening there (which is usually nothing, anyway). Also, streaming is something that I find quite weird... Am I playing a game to enjoy it myself or am I trying to make me playing the game a enjoyable experience to others?

Usually, I'll simply use streaming as a lazy way to record and automatically upload runs to the internet. It's only recently that I started to interact more with people as I'm doing attempts (and much to my surprise this has been quite a fun and interesting experience). However, my default is still to start the stream with the microphone muted and mostly ignore the chat.

I guess my point here is (again), why do you want to do that?

All in all, I find this a really fun hobby, but it can be quite demanding as well... If you fell like you won't enjoy practicing, running etc, maybe try other stuff? Maybe try glitch hunting, TASing, blind racing... I highly suggest trying blind racing, but I may be somewhat biased to that.

I hope all this rambling helps in some way. :D

I made a requested, and it was added to live split!

Now, the script may be downloaded, activated and configured directly from the splits window (instead of having to add a controller to the layout).

OK, fixed these issues!

For the robot, I changed it to detect when the boss died and then split at the next loading screen. However, I didn't make this a setting (as only my runs had been done with the weird timing).

And sorry about the orb split... I should have imagined that someone would use it, since it were there. 😬

Would you like to test it before hand (specifically, the orb splits)? I tested the orb options in Hakurei shrine and in the first level, and everything seemed to work fine...

Hello all!

I've made some changes to the auto splitter:

• Stop splitting when the player dies to a boss
• Change timing to Robot split to as soon as its health is depleted
• Pause the timer while in the main menu
• Add option for 100% runs (which do not reset when entering the Level 1)
• Adjust timing of the final split

Here's the script's repository: https://github.com/SirGFM/ykm-auto-splitter.

Since there isn't any run in the leaderboard that split when collecting the yellow orbs, and since I had to pretty much rewrite the entire script, I ended up removing those options... If anyone do actually use them, let me know and I'll re-add it to the script.

Other than the improvements, the most significant change to the script was the change to the Robot split... Shortening it to right after the boss dies was the cleaner way to split it consistently...

If everybody is fine with this, I could request to add the auto splitter to Live Split, so it may be automatically downloaded by the software.

Sorry for the double post, but... erm... I may have nerd-snipped myself into doing it.

You can find it on my github: https://github.com/SirGFM/HTMLimageSelector.

Do note that this is pretty basic... You would probably need to tweak the CSS to make it look better. However, it should be pretty simple to use: download the page (index.html) open it up on a browser, configure the size of the images/grid and load every image.

Making a simple webpage (i.e., a HTML page with a bunch of pictures and some javascript to mark/unmark the pictures as you click them) shouldn't be that hard. Making it look good, though... >_<

However, this would work only as long as you were accessing the page in a browser that could be captured by your recording/broadcasting software.

I'd say most shmups aren't well suited for speedrunning as well. Usually, other than defeating bosses quickly, there's no real way to speed up the game... You can avoid going slow (by not dying, destroying enemies quickly etc), but other than that you usually have to wait on the game's scrolling speed.

However, one thing that I've learned from blind racing is that people will race/speedrun literally anything... And there's validity in that!

Even though I don't think shmups are well suited for speedrunning, it doesn't make these runs any less impressive. These games are usually pretty hard and you have to 1cc it to lose as little time as possible.

At the end of the day, you'll most likely find at least one person that likes any given game enough to speedrun it... Heck, people even run auto clickers!

It's called JJAT, and it's available on itch.io (https://gfm.itch.io/jjat). The game's definitely short, but I was hoping that having a "all rooms and collectibles" category would make it long enough as to not be an issue, as it should be longer than 5 minutes. As for playthrough lengths, a blind race took 24 minutes (having most of the branching paths blocked out) and a friend of mine took at least 40 minutes to beat it (without any blocked path, but having seen the game before).

Out of the 5 points in the game request rules, it should be probably safe in at least 4 points (longer than 5 minutes; it is a functionally complete game; a speedrun can be quite optimized from a casual playthrough and it's not a fangame), however I do believe the game is quite unknown. Some of the features that allow you to go faster when doing an actual run are:

• One character may carry the other, moving at double the speed if being carried and moving in the same direction. Using this and the teleporter (an equip-able item) you may move slightly faster;
• Momentum is maintained between screens, so you may jump to start the next screen getting on top of the other character (possibly saving a few frames, although I never timed this);
• There's a wrong warp (using the teleporter while in hitstun from the last hit) that allows skipping one puzzle and a half;
• You may skip every collectible and go directly to the last area, but it becomes quite difficult/precise.

I'll consider this for a while longer, and see if anybody else has any other opinion on this. I'll probably try to request it during the weekend.

Thanks!