No level leaderboard?
6 years ago
Kentucky, USA

I can't find level leaderboard and would really like to speedrun IL (start as small mario)

North Carolina, USA

The IL leaderboards are gone because they don't make much sense for SMW. https://www.speedrun.com/post/wdtf9

333Rich333 and VeeOfEarth like this
North Carolina, USA

@16bitPanda it is unclear whether you are being sincere or sarcastic. Would be good if you could clarify

North Carolina, USA

It isn't an issue of them being confusing. It is that they are pointless. If you want to keep track of your best level times, go for it, but maintaining a list of something like 96 exits with something like 10+ variants per exit would be pointless for the mods to handle. It isn't useful for the SMW RTA community, and so we removed it a long time ago.

Dode, 333Rich333, and Jumpyluff like this
Colorado, USA

I've got my thing going, and if somebody wants to use this and possibly add on, feel free to.

https://docs.google.com/spreadsheets/d/1bb5Jx_Rst0quDZBqOA3OW8-GqOPEG2TIKL0Lfoliqc8/edit?pli=1#gid=1521295361

Just right click the tab, and export it to your own sheet somewhere and remove my times.

North Carolina, USA

Yes Cape, No Yoshi Yes Cape, Yoshi, Wings Allowed Yes Cape, Blue Yoshi, Wings Allowed Yes Cape, Yoshi, No Wings Yes Cape, Blue Yoshi, No Wings No Cape, No Yoshi No Cape, Yoshi, Wings Allowed No Cape, Blue Yoshi, Wings Allowed No Cape, Yoshi, No Wings No Cape, Blue Yoshi, No Wings Small Mario, No Yoshi Small Mario, Yoshi, Wings Allowed Small Mario, Blue Yoshi, Wings Allowed Small Mario, Yoshi, No Wings Small Mario, Blue Yoshi, No Wings

Strats change considerably after you've finished special world because some of the sprites change. So maybe add that. Do you want glitchless variants? We can do that too. There, I've just given 60 valid variants for each exit.

ILs that don't have a purpose in an actual speedrun are largely pointless, and even if you find value in them, that doesn't put the onus on us to curate that leaderboard. If you'd like to make your own IL leaderboard with a subset of these that you find interesting, go for it bud!

Edited by the author 6 years ago
333Rich333, VeeOfEarth and 6 others like this
British Indian Ocean Territory

I guess I'll weigh in as one of the few people who almost exclusively does SMW IL runs (as opposed to full-game runs). As you might guess, I disagree with Authorblues's philosophical reasons for opposing IL boards.

I don't think the purpose of IL runs is solely to show off strats to help RTA runners; if that were the purpose then it definitely would not be worth having leaderboards. The point of leaderboards isn't to encourage demonstration, but to encourage competition. For demonstrative purposes, the wiki is far more useful than a leaderboard could ever be.

I also don't think strats which are useless for RTA right now are pointless; just in the time I've been around there have been quite a few moderate-size improvements (typically 0.5-2s or so) to category routes which originated from IL strats. These have been slow to be adopted in runs, but it's pretty normal that RTA runners, even at WR level, do suboptimal stuff to get a slight edge in consistency in longer categories, which are where the most improvements have been found. But eventually some of these improvements will have to be taken seriously because in categories like 96 exit there are very few large time-saves left to be added; much of the remaining time to be taken off WR other than mistakes is from these moderate time-saves.

The complaint about "too many categories" is also pretty silly under any slightly critical analysis. There are far more possible RTA categories than IL categories. The number of distinct RTA categories is pretty much infinite. People have speedrun "Beat Cookie Mountain, small only", "2016 exit", "17 message blocks, no cape", "99 lives", and "America%", and you can easily imagine other even more ridiculous categories like "Beat all large dot levels four times". However under some conditions which would include all the categories suggested so far there are only finitely many (if still an extremely large number) of possible ILs. Of course the number is still too large for us to ever do all of them but no less so than it is for RTAs. The solution for ILs is the same as for RTAs: you don't compete on every possible category, but you adopt a handful of ultimately arbitrary ones and compete over those. The first 2 obvious choices would be Small Only (no Yoshi) and Any Powerups (same rules as 96 exit). Just like with RTA runs, more categories would be added as they are needed to document competition, not simply because they are possible things for someone to speedrun at some point.

With that said, as a user of the old IL leaderboards, I'm still quite happy that the previous boards were removed, for more practical reasons than the ones Auth cited. I do not want them to be remade. I've identified one core problem related to the community and several auxillary problems related to the implementation of the old leaderboards. I don't think any future leaderboards will be very successful without addressing all of these.

Core Problem: IL runs are (presently) not seriously competitive in SMW, unlike other games.

If someone (e.g. Lambby) posts a new video of an exit which is faster than my best time, I don't feel particularly any more motivated to beat it than before. The people who do ILs right now in the SMW community aren't really that interested in grinding out frames or competing for the best times. We have other motivations. For me personally, I enjoy strat experimentation and documentation.

Other communities (most famously GoldenEye) have made competition over ILs serious business, but for SMW there are really only a few of us who do ILs seriously. In terms of active people who primarily do ILs, it's mostly just me and Lambby. Many top RTA runners have also done ILs of some exits, either for practice or just for fun, but rarely if ever for competition. There are games where top RTA runners have also competed for good IL times; sm64 in particular comes to mind. In comparison, SMW has no history of this; for many exits if you want to know who has IL WR there is a good chance even the person who has it would not know.

If there were 20 people exclusively doing serious ILs at a high level instead of 2, maybe there would be more competition. However, without that competition, a leaderboard is useless. A leaderboard does not particularly encourage people to compete, and so if there is no competition already happening, it serves little purpose. So the first order of business would be actually getting people to compete over IL times, particularly current RTA runners since they are already skilled at the game. I do not know how to go about this nor am I particularly motivated to do so but at the core this is the issue in my mind.

Secondary Problem: The old leaderboards highlighted unimpressive runs.

If there is no serious competition, that doesn't mean that no one will submit times. It just means that the people who do submit times will not be very serious or skilled. As a result, many of the times on the leaderboards, including some claimed "World Record" runs, were really not very good at all. I think that, as an absolute minimum, if a run is going to be listed as a IL WR, it should be at least as fast as the best time top runners would normally get. That standard was not met by many of the runs on the old IL boards. Additionally, if there are faster strats that RTA runners don't do but are humanly doable, I think a big part of an IL leaderboard should be to show off those types of strats. For almost all the runs on the old boards, I could name a faster video somewhere, and in many cases it wouldn't even have to be an IL video. There were maybe 5 really solid ILs on the whole board.

Put simply, the movement in this game is already pretty tight at the RTA level. While everyone can say "I can have better movement if I sacrifice consistency and grind out a lot of attempts" that only goes so far. A newer runner will not be able to beat a top RTA runner just by doing this; you need a good understanding of movement, the fastest strats (or at least close), and high-level gameplay to do a very good IL; even when you know what you are doing it takes a fair bit of time. Almost invariably, the people who do want to compete over ILs are newer runners who (no offense intended) don't really have the necessary knowledge or skills to beat even normal RTA times. As they improve, they tend to also become less interested in the idea of competing over ILs. I'm happy to help newer runners with an interest on these things, but a leaderboard is the wrong medium for that type of assistance.

Advertising unimpressive runs as IL WRs is silly and bad for the community. I suppose I could have recorded mediocre runs which would still beat those and posted those to the boards, but that's a lot of work over something that I ultimately don't care to compete over and the runs still wouldn't have been that good. Apparently, most other skilled players felt the same, and so most of the sub-par runs still got to keep the title of IL WR according to the leaderboard. Doing good runs, up to my standards, for all the categories on the old leaderboards (about 400) would be a monumental exercise which I have little desire to do.

Secondary Problem: The leaderboards were too difficult to use to be worthwhile.

For an RTA runner who got a good IL time in practice, submitting an IL run required the following. Take your video (assuming you were recording; if not too bad because it didn't count) and cut out the level. Upload the video to YouTube (because you probably were not streaming because it was practice). Then time the run, accurate to the frame, and submit it to the IL boards, which required as much work as submitting a full-game RTA. One of the mods would verify it much later (IL runs tended to take a long time to get verified despite being very short).

For a run that you did in practice in a matter of a few minutes, this was simply not worth the effort in most cases. In particular, the requirement to be recording the run live, and the requirement to retime the video accurately, made it mostly a lost cause. Almost all of the top RTA runners never used the IL boards for this reason even when they got good times. If it was a time you were particularly proud of you'd post a video, but submitting it to the leaderboards was an extra step no one wanted. Additionally this made it too much work to put up old times posted before the IL boards, which compounded the above problem of mostly empty boards highlighting sub-par runs rather than the best of the best. I only used the IL boards twice before deciding it wasn't worth the work, and in both cases there was no competition over the time.

Secondary Problem: Unclear how to time ILs.

Historically, IL videos have been labelled by their in-game time. The in-game timer is not particularly accurate; it fails to account for frames in which the game is frozen due to lag, screen transitions, scrolling, and powerup animations, resets on screen transitions, and does not correct for different methods of ending the level (e.g. Yoshi wings vs goal tape). That is not to say that ILs are done to minimize in-game time. Some are, but most are focused on getting the best real time, even at the expense of game time.

This has never been a problem because, as I said before, ILs are not presently competitive. The use of in-game times is little more than a label that makes strat videos easier to find and discuss. It is slightly useful for RTA purposes because during a run you can easily check the game-time while measuring real-time is harder. In-game time is still important; saving a particularly tight game-second is an effective way to show off how much faster your new strats are and a minor claim to fame even though real-time is what matters at the end of the day. But at best it is an approximation and at worst it is misleading to the point of being wrong.

However, competing for real-time in such a short run also has major issues. Invariably, every competitive exit would end up becoming a frame war, and in this game because of oscillating speeds and other nonsense SMW frame wars would never really end. This isn't something anyone currently wants. People who do ILs in the community are primarily motivated to find new strats, not to grind out every last frame, and people who do RTAs primarily would get bored long before then. Competing over real-time also adds an extra step to the submission process of timing your run to the frame, which as mentioned before was a big pain for the old leaderboards.

Secondary Problem: Some stages are just uninteresting.

This is SMW after all. There are plenty of flyover levels. Doing an IL for those levels is usually pretty easy, but because the submission process was the same time-consuming thing for all levels, this just compounded the other problems above of a lot of exits not having times at all or showcasing times that were not particularly good. There isn't much you can do to make these stages more interesting, but at least if the submission process was easier it would improve the chances of someone submitting something decent. Of course another possibility would just be to remove these exits entirely from the boards but that becomes very subjective quickly what is "good enough" to be worth including.

A (partial) solution: The SMW Practice Cart by Dotsarecool is probably the best speedrunning practice tool of any game. It has a timer that, while not perfect, is fairly accurate to real-time (including lag). It allows recording and replaying movies, which would make it possible to get a video of a good IL after the fact (at least for console players; emulator may be too succeptible to cheating for this to be allowed). It also has built-in quality levels for each exit for 4 possible categories. It also has several features (RAM viewer, speedometer, etc) which make times done with it not directly comparable with times in Vanilla SMW, but honestly I don't think any of these make things less interesting. All top RTA players these days use the practice cart extensively for practice, and many ILs are also done on it.

If instead of having IL leaderboards for vanilla SMW, we had IL leaderboards for the practice cart, this would solve most of the secondary issues. Submissions would be a fair bit easier because you would not need to be recording practice live and could instead record the movie playback. Timing would also be little more difficult than it is for in-game time in Vanilla SMW, and the practice cart's timer is, by now, a good enough analogue to real-time that there would be little controversy in using it. Sub-par runs could be filtered out by requiring submissions to be at least gold/pink tier times, which would also reduce the workload of the mods. In particular, pink tier for most exits requires extremely high-level play that is comparable to (though typically a bit slower than) IL WR times. Times worse than gold have usually been beaten hundreds or thousands of times in RTAs and at that point it doesn't seem particularly worthwhile to keep a leaderboard tally of them.

You would need to make some rules to make it work, since the practice cart has features which are not typically allowed in ILs. For instance it lets you bring Yoshi into castles and ghost houses, which might save time in some places. However this is mostly a technical detail which could be codified in the rules, just like we do with RTA categories.

That said, I do not think this would solve the fundamental issue that ILs in SMW are just not competitive right now. One might hope that simplifying the process and upping the quality requirements would get RTA runners more interested in using the boards, but I have serious doubts how far that would go. I'm not really interested in changing this either. I actually rather like the status quo where IL runs are mostly about finding new strats and executing them, not about grinding out individual frame improvements.

So if I were going to make new IL boards (which I have little desire to do) this would seem to be a better option than what we had previously though by no means a perfect one. LTTP has something fairly similar at https://docs.google.com/spreadsheets/d/1dV2ozAGgDdFZAne8rwG3h0KyOPgIN72pWeRsS1hR2QU/ and they've managed to get a fair bit of participation. I don't know if I'd call that spreadsheet competitive, but it's at least more organized and complete than what we had with the old IL boards.

Umari0, Xander479, and authorblues like this
Colorado, USA

I do agree with AuthorBlues on this topic. Super Mario Brothers had 35 entrees for the levels: 1-1 - 8-4 and -1 - -3. All started as small Mario and finished the level however they can with the fastest time on the timer in the corner. Recently, it was removed for being difficult for mods to accept runs along with the times just being downright stupid. For example, take 4-1. You would enter a pipe that gives a faster ingame time but it would be slower to take RTA. This would be the same thing. Tens and sometimes even hundreds of ILs would be sent in daily and give the mods a lot of runs to accept. If this was implemented like Super Mario Bros. where you start small and finished the level, you would have about 3 times more runs to accept every day with the level count being 96 over the 35 that are in Super Mario Bros. It would be a lot of work for nothing and I disagree on this.

Edited by the author 6 years ago
North Carolina, USA

Thank you TLD for that excellent counterpoint. I think all of the SMW RTA speedrunners are grateful and impressed by things that IL runners like you and Lambby have come up with. I guess I don't mean to downplay the value that ILs have even when not applicable to an RTA speedrun, but my criticism was more to do with your point about some levels just not being interesting. If we have a leaderboard for YI2 with Cape, that IL just doesn't make sense.

I think I would support the creation of IL leaderboards on SRC under the condition that everyone could agree on a ruleset that addressed a few points:

  1. consistent timing (IGT is a notoriously poor choice for Mario games across the board, and maybe the practice cart could solve that problem as you suggested)
  2. a reduction in modifiers. Small-only leaderboards, like what Lambby has done with his small-only spreadsheet make total sense. Cape makes sense in most cases, with "no cape" devolving back into the same strats as small-only in most cases. How do we handle Yoshi? Pre-/post-special world? Maybe no restriction in that regard and players can just do whichever option gives you the best time?
  3. clearly delineated thresholds for what is submittable. Again, a great suggestion you had is to go by the gold/pink times on the practice cart. I think some of the pink times were achieved by segmenting the stage to play more or less perfectly (Dots can correct me if I'm wrong) so gold is probably a more realistic goal.

I'm not opposed to ILs for ILs sake. But the way we had them before were messy and they were a stain on the SRC SMW page, with a bunch of random, optimized times claiming world records just because no one else wanted to record a video for YI3 with cape, etc. But you've convinced me, more or less, that I shouldn't just dismiss it outright.

Edited by the author 6 years ago
British Indian Ocean Territory

Yeah I think the main problems are practical, not with the principle of the idea. I'd start with just 2 categories, namely Small Only (where I'd also ban switch palaces and maybe even beating Funky) and Unrestricted (where the only things that would be banned are glitched item box and PI states). There will be some strats that are pretty useless, like YI2 with cape, and some that would technically be possible but wouldn't actually come up in RTA, like bringing a Yoshi into VS1 secret. I don't see either of those as a huge problem but that's just me. That's already 191 categories of ILs (98 for each except 5 of them are not possible small only), so I don't think including any more than that initially would be a great idea. If we were to branch out beyond that later, "No Cape" (obvious), "Lunar Dragon" (obvious), and possibly "Glitched" (i.e. allowing glitched reserve items) would be a few natural choices but it would depend on what people actually want rather than anything else.

191 ILs is not as impossible to manage as one might think. It's a lot, but there are a lot of SMW runners, and most of the exits already have good videos somewhere. I would certainly not want to start the boards blank, like they were last time. But I think we could find good times for most of the exits that, if not the fastest ever, would at least not be embarrassing to list. That said, some things would need to be redone, including most notably all of Lambby's small-only videos since they were done on Vanilla SMW (mostly before the practice cart even existed). So it's not completely trivial even just to put up new leaderboards if we decided to do so.

But I just don't think people care about competing over these things very much in the community. If a lot of qualified people disagree and want there to be IL boards I think it'd be workable but my experience is the opposite: the people who are really skilled and have good IL times don't care much (if at all) about defending those times or competing over ILs. For clarity's sake, that's not just me and Lambby. Off the top of my head, Link, Dode, TM, Maestro, SilverStar, Seathorne, Aaron, pogyo, Lui, Mostly, Akisto, Umari0, and reapergoblin all either have current (recorded) IL WRs, or have been beaten by only frames, and they're all primarily RTA runners. I'm sure there are others I'm forgetting or don't know about as well. If a number of people at that level would commit to participating in IL leaderboards, even just for a handful of exits each, I think we could get it to work. But so far I have not heard any indication from any of those people that they would want such boards to exist, and I don't really care much either way (and I don't think Lambby does either). So I don't really see any reason to suggest creating them at this time. That could change in the future though.

Edited by the author 6 years ago
Canada

"I actually rather like the status quo where IL runs are mostly about finding new strats and executing them, not about grinding out individual frame improvements. "

Yeah, this is how I prefer it as well. I'm probably not going to try to compete with an IL time if it already uses the best strats, even if it's a few frames off of being perfect. Finding new strats and exploring unknown territory (like Lunar Dragon) is more fun for me.