Nobody here who can build the timer mod for Windows ?

Considering that loading times can represent 2-3% of the RTA time of a run, somebody who'd like to try to do a run on 0.9.3 on Windows might be put off from even trying if he don't get them deduced, while @Alistair_Findlay and me do using the linux builds.

Having received no answer (only that like from UJ), I tried to change the categories for full-games (I only changed 0.9.3 atm) ; with "any%" and "all challenges", removing "novice%", "intermediate%" and "100%. That last category is not fully covered by all challenges, but until someone publish a run for it, it seems rather pointless. Especially as the fastest way to get the 5-laps achievements is to watch a replay of a 5 laps ghost on OMC (yes, it triggers the achievement), and as it's not covered by the timer mod to remove loading times.

I put a "difficulty" variable to store the info and still have expert/intermediate/novice runs differentiated in the leaderboard, however it is also added for any%...

So tell me how you prefer it : now, or before ?

The ability to class the levels of the IL list in groups/categories.

In https://www.speedrun.com/stk/individual_levels; we have three type of ILs : GPs, time-trials races and egg hunts. It would work better (less rules duplication, and no need to have 20 times the same variables to only apply them where they makes sense) if we could set rules and variables which applies to all the levels of a specific type.

More FTL-hate, just look at this (it'll open at the beginning of the FTL) :

I won the race, but I could as well have lost considering how it went.

In this try, yes. In the previous (yeah, I also failed another attempt before because of this), I had done two other races before.

Any% is quite short, so it's less annoying than when it happens later in a full-game run, but still broken.

I think we need a FTL hate-thread.

@Alistair_Findlay already has a video up about this.

Here is what happened to me tonight :

I was roughly on par with the Any% WR time, even maybe a few seconds faster (I finished Candela City 9 seconds earlier, then did Gran Paradiso rather than Green Valley so it's hard to compare), then FTL happened. I didn't race carelessly : I applied the "try to get a lot of nitro at the beginning to use when you're caked" technique.

But... Just watch it yourself. There are two races. Each make me hate the FTL mode. Having to follow-the-leader means that if you're skilled enough to go faster than the leader (and skilled STK players go A LOT faster) you're harmstrung. You can't use your skill to gain time, and when all karts are packed close, you can't avoid any bad item.

New features :

For the moment, I've kept the number of replay file you can compare to two and it's not yet possible in normal version to switch view from kart to kart as I do here (I use a debug feature for this).

It's not that it matters much, but having a fully inactive mod don't make much sense to me.

I noticed that there are still items in egg hunt mode, and it irked me that a lead dev want to keep them as a "playground for kids" (as if there wasn't battle/soccer/race with 0 AI to use items randomly...).

So I just added egg hunt mode to the IL leaderboard. I'd like to have some nice runs to show the next time the argument come up.

There is currently a way to switch kart view in debug mode, I'd need to port this and assign it to some button.

More than two ghosts (files, because one file may contain more than one kart), but how many ? Three, four ? The UI would not be able to accommodate much more. There are diminishing returns here.

How should the live difference be defined if there are multiple karts played ? Difference to the kart before/behind ? Difference with the first kart ?

You can find several replays here : https://github.com/Alayan-stk-2/stk-code/blob/Replays/data/replay/Replays.zip

Hacienda, XR591, Scotland, Cocoa Temple, Abyss, Green Valley, Shifting Sands, and 2 Zen Garden ghosts with a <0,15s delta.

I've finished cleaning up the board. Now, there is only 0.9.3 ; 0.9.0 to 0.9.2 and 0.8.x in the ILs. Much much cleaner, and the columns aren't crammed onto each other anymore.

I've also rejected a bunch of old times which were imo legit (though some had add-on karts) but which didn't have a video. No video, no place in the leaderboard, period.

Tell me if there is anything wrong.

Btw, I also question the utility of having novice% & intermediate% for full-game runs. Don't really see the point of having a "low%" category. It's not as if it was difficult to get a low% (as it is in some games), it's just plain boring.

I'd have as categories (not names to be used, only the idea) : Any% (can skip challenges) ; All challenges (any level of completion will do) ; All challenges on max difficulty.

For 0.9.3, the fastest way of doing all challenges is to do all challenges on expert, which obsoletes the "all challenges on max difficulty" category

For the future 0.9.4, this distinction will make sense as for SuperTux% you'll have to redo some old challenges, making it take more time than just having all challenges. However, any% and all challenges at any level would then be equivalent...

Maybe we could have a "no GP skip" category, though I'm not sure if it really makes sense.

Here is what I have done : 1)Have a difficulty variable for all IL runs. 2)All 0.9.3 runs have been assigned to the correct difficulty variable. (Non 0.9.3 runs are currently all "SuperTux") 3)All 0.9.3 runs have been moved to a common "SuperTuxKart 0.9.3" category (I put "SuperTuxKart" because the column title needs some longunbreakableword to be large enough)

All IL runs for a version are displayed on the same leaderboard with difficulty mixed. If a player has runs on multiple difficulties, his PB for each difficulty is displayed in the leaderboard (they're set to not obsolete each other), but it will do weird things in your player profile list of ILs.

TODO : 1)Move 0.9.2- runs as appropriate, on the model of what I did for 0.9.3. Separating 0.9.0 runs from 0.9.1/0.9.2 runs would be useful. Once all the runs have been correctly moved, empty IL categories can be deleted. 2)Determine if it's better having reverse mode as subcategory (I've put it like that for the moment) or having it as a general category (upside : will be shown in the general IL timetable. Downside : less logical organization, will create clutter when there will be several versions with both normal and reverse ILs)

I've found a way to get subcategories for ILs. However, moving time around from one setting to another is always risky, so it will take some time to do it properly.

EDIT : After some more testing, it is the best time of the first subcategory which is shown on the main IL leaderboard.

I personally like to have the time down to the ms ; I'd let them like that. If a situation should arise where two times are too close, we'll be able to handle it with more care, requesting for ghosts and retiming.

It would be good if all IL WRs were required to post a ghost (that way, others would be able to train against it, or even check it if they wish so). I'm not sure how we should organize it.

[quote]But still, how active or interested I am doesn't give me the right to super mod.[/quote]

It doesn't give a right, I just thought it'd be nicer to have a supermod who actively runs and verify the game, and is known to be trustworthy.

[quote]I'm actually confused about one thing, why are the times milliseconds, and not what the time says in the IGT in ILs? I've verified very badly since I'm not sure where the last decimal comes from. @runners did you retime your ILs?"[/quote]

Replays times are saved down to milliseconds, and the values come from there.

However, the game, when playing a replay file is not perfectly precise, mainly for the lap times which are also flawed in normal play. If you play several times a replay file, you can have some variations in the time displayed. In theory, you could squeeze some milliseconds simply by replaying again and again the same replay file.

If someone tried to cheat his milliseconds times, he'd practically never gain a place from it, and on the long run, over 10-20 ILs, it would be possible to spot an anomalous pattern anyway (that'd be a very stupid way to cheat).

The main issue that could arise is two runners having a very small difference, it would not be possible to know for sure which one is faster, as it would not be possible to rely simply on the ms time given by the runner.

The good new is that my improved ghost mode allows to compare ghosts, including from 0.9.3 and older. When the timing issues of the race result GUI will be fixed, it'll be possible to feed two replay files to 0.9.4 ; and to watch them together, which should allow to find if there is anything suspicious and which is the fastest.

We have IL categories as misc at : https://www.speedrun.com/stk/individual_levels

But they still show up on the main IL leaderboard.

To clarify my point about the "stupid kid" : there is 25 ILs per difficulty per version. If we add reverse ILs, that could be 200 ILs per version.

That's enormous, and who's going to submit for all that ?

If some new runner were to come tomorrow and to submit a WR for the At World's end GP in SuperTux, I'd say why not ? That'd be good to have a new runner, even if not very skilled, and as the category he's submitting for is legitimately interesting, it would be up to other runners to beat his time if they think they can do better.

Now, have some new runner come and post times for the 75 or so expert/intermediate/novice ILs which currently don't have a time. They would be WR by default, as they would be uncontested. Now, would that mean that this runner is the best STK's IL player ? You could think so by looking at the leaderboard, you wouldn't see there is no competition at all for those ILs

And if those are crappy times, should I take a day to record 75 videos and to submit them to knock him off the first place ? And should then other skilled runners try to knock me off on them ? That's not a reasonable expectation. We can't seriously compete on 200 ILs per version.

I'd just have have the levels without the difficulty specified : all allowed. If someone does beat the SuperTux WR on Expert, then certainly this is a superior performance which would deserve first place on the leaderboard. Of course, in practice, you'd have SuperTux score pretty much anywhere (is there any situation where it can be slower ?), but the point is that you'd have 4 times less categories.

If you look at other kart's racing game leaderboard, you'll see that they didn't bother to divide by difficulty, as they had way enough categories without that. For example : https://www.speedrun.com/mkw Even MK8, which does divide between 150cc and 200cc due to the Deluxe version, didn't bother to add 50cc and 100cc leaderboard as they'd be completely irrelevant, see : https://www.speedrun.com/mk8

That's not very practical if there is no way to do so unless contacting SRC's staff.

I guess that's what we'd have to do, unless there is any objection to my proposal ?

(Off-topic, or not : would be nice to have some more runs even if they're not that great, you're not required to post an excellent time)

What made the FTL race close to the end painful is in great part that it was very random and that some bad luck/minor error could cost you up to 2 minutes.

But the egg hunt challenge would be another story : if there is no time limit, doing some maneuver wrong and losing 3s in the maneuver means just losing these 3s ; and as there is no item/no AI, the random part would be zero, only a player with his skill. So sure, if there is one closer to the end, making an optimal run will be harder, but if there is less reset possibilities, it will reward on average the more consistent (and so skillfull player). As it's possible to train for the egg hunt separately, this wouldn't make training itself uselessly painful.

What's cool with egg hunts too is that there is a bigger routing part. What's the optimal path between the eggs ? That's an interesting question, and there will be some new strategies discovered over time.

Egg hunt could probably be adapted for story mode with some thematically rethought goal (finding some kart-related items or whatever can make sense as a story-mode goal, if you have ideas...)