I don't think at all like adding yet another category/variable, but that's just me.

It is possible to start in a saved GP in story mode.

As this requires having done the beginning of the GP on another profile, and allows to have an arbitrary number of AI karts, this is obviously banned from full-game runs.

Beware though, as if the conditions for it to trigger are met, you can't avoid the bug and it can ruin your run.

Yeah, that's what I meant

How is that route balanced ? The goal is to get the hardest/luckiest part behind you as fast as possible, and reducing kart switching to a minimum. I kept it similar to the any% route at the beginning.

Considering the improvements made here and there, here is what I think is (nearly) the optimal route : 1)Start with a light kart in front of Scotland 2)Scotland 3)Oliver's Math Class 4)Switch to a medium kart 5)Cornfield Crossing 6)Cocoa Temple 7)Shifting Sands 8)Hacienda (with skip if executed reliably enough (>50%)) 9)Northern Resort (with skip if reliable enough) 10)Zen Garden 11)Snow Peak 12)85-points GP

• a)Skip Cocoa Temple
• b)Do Northern Resort (with skip if reliable enough)
• c)Either skip Hacienda, or use the skip (confusing I know !)
• d)If you did Hacienda, skip. If not, do Zen Garden
• e)Do Snow Peak 13)Candela City 14)Switch to light kart 15)Do 40-points GP
• a)Do Shifting Sands (use the skip)
• b)Skip Cornfield Crossing
• c)Do Oliver's Math Class
• d)Skip Abysses
• e)Do Scotland 16)Do the FTL 17)Do Blackhill Mansion 18)Do Green Valley 19)Do the 125 point GP
• a)Skip Gran Paradiso Island
• b)Do Green Valley (with skip)
• c)Do Blackhill Mansion (with skip)
• d)Do STK Enterprise (with skip)
• e)Skip Volcano Island 19)Do Minigolf 20)Switch back to medium kart 21)Do the 165 point GP - this is one of the hardest part as you need to pull-off a double-skip in a 4-races GP
• a)You don't have unlocked Fort Magma, it's not among the track you have to do
• b)Do Minigolf
• c)Skip XR591
• d)Skip The Old Mine
• e)Do Around the Lighthouse. (At this point, all major difficulties are behind you) 22)Volcano Island 23)Abysses 24)Gran Paradiso Island 25)Around the lighthouse 26)Old Mine 27)XR591 28)Fort Magma

Yes, it's seconds losts vs sum of best, it's written in the subtitle.

I'm now waiting for Fouks's run. I'm not sure if he can do sub 31:30 without resetting an unhealthy amount of time, but there is certainly room for him to beat my new WR. I'll see if he has some new tech or improvement, and I'll try to beat it back afterwards.

In theory, as the sum of best shows, sub-31 is doable, but I don't expect it to happen any time soon unless some new time-saving tech is discovered.

IL boards have not been designed to manage multiple versions of a game, it's impossible to remove a given level for a given "category" (we use it to indicate versions).

I've finally done a sub-32 time in Any% ! After a few more frustrating resets, I finally got a good run. The video is uploading, it will be available in a few hours.

In the meantime, here is the breakdown of it went, compared with the previous WRs :

1m20s12 ; no nitro ramp

PS : I'm no student and this is no homework :p

EDIT : 1m18s145 (I got 1m19s232 earlier, but I decided to insist so that the WR will stay for a while).

And you were right about the 27 minutes : I was actually at 27m08s previously, I'm now at 26m58s.

This skip looks quite usable in full-game runs.

As long as you have some consistency, over enough tries, you get one without unluck/big mistake, sure. But making many tries can be exhausting :c

Thankfully, I managed to get more consistency on several skips (green valley, black hill mansion), so hopefully I should be able to do a good run this week-end.

I'm curious to know on what track you beat me. Hacienda using 3 times the skip ? New skip somewhere else ? Or a track where you could execute better than me.

Btw, the sooner you can submit a time for all tracks, the sooner I can keep track of your total time. With the improvements of the last days, I'm at 27m10s total. It is certainly possible to break the 27 minute barrier,

RAHHHHH!!!!!

End of XR591 : I use some nitro because I have 20 and only need 18. Maybe I used a bit too much. I get the sound telling me it's finished, and I'm thrown back to the race selection menu (probably used too much). I try to move in the menu. It hangs. Crash.

I was at 24m18s, so approximately 20s ahead of the 32m11s "WR" !!!!

I'm done for tonight. I will submit my 32m08s time later this evening.

Thanks for your answers, guys.

I have started the upload of that run, in case, and it won't finish before 3 or 4 hours ; but I'll try to beat it in the meantime.

I did a single try of full-game run tonight.

I tried to use some new techniques, and it didn't work out very well (I tried to use the northern resort skip and failed on each of the 3 laps...).

However, by some miracle, despite some usual stupid errors, it was overall rather clean, and I didn't get trolled by basket ball or parachutes.

The new time is of 32m08s599.

So I'm now torn about submitting it or not.

Pros :

• It IS the current WR
• Accurate leaderboard history.
• Allow the people interested here to see how it paned out and maybe get ideas.

Con :

• It takes only 2,5s from the record ; while it is clearly possible to get 15-20s less without having to do a perfect run.
• With some luck, I may be able to beat it again more convincingly rather soon.
• My internet upload speed is poor, and it would take several hours to be uploaded to youtube.

Your take on this, guys ?

I'm firmly against having better items depending on the distance from the first kart. While over many races, skill certainly matter in MK ; the average race is too luck-dependent in my taste, and this only compound the issue.

At the beginning (the only moment when most karts are close), those in the last positions can indeed get a good item making their position not so bad ; but remember that STK do not have items as overpowered as many of MK's item. The star, the bullet bill, and many others, are extremely powerful, much stronger than what you can get in STK. While I think it would be nice for STK to have one item better than the zipper, this is not really a concern.

And when the first kart is far ahead, it means that the player driving it drove much better than the other players/AIs ; if despite the standard item advantages they can't keep up, they don't deserve to win. Giving them many OP items to make them catch up with the first is rubber-banding at its worst. The 1st lap of a 3-lap race barely matter in MK, you can wait 10s before starting and your odds of winning are barely reduced.

In MK, you often have in multiplayer people who deliberately try to not be first in order to get a blue shell to use close to the end. I don't think this is an healthy dynamic at all.

Of course, I acknowledge that for casual players who drive badly, getting a lot of OP items to compensate their ineptitude make them happy, so it is not a bad trick when the goal is to sell to millions of people.

As someone who has been more often on the other side, being at the receiving end of things like a blue shell - thunder - red shell - red shell combo ; I don't want STK to clone this. For friendly multiplayer between players of different skill level, using the handicap option (which ought to be improved) is imho a fairer way to keep it entertaining.

This change list is not up-to-date.

Notable changes since my last update :

• Many networking improvements I can't keep track of
• Many android bugfixes I can't keep track of
• New capture-the-flag and free-for-all modes in networked multiplayer
• Improvement to the unlock cutscene so it display things correctly when 5 or more things are unlocked at once
• hiker fixed the cushioning bug : sometimes, when jumping over a downward a slope in 0.9.3, it is possible to never land on the ground, because of the way the compensation meant to smooth the landing work. This makes the kart uncontrollable as it remains in the air much longer than it should. This is no longer an issue now.
• hiker also tried to improve the collisions issues : head-on collisions now reduce greatly the kart's speed instead of sending it flying around. However, when it is closer to using a ramp to jump than a full head-on collision, the behavior remains similar to 0.9.3.
• Huge revamp of keyboard navigation. Previously, up/down functioned like next/previous, cycling across elements like the "tab" button typically does. Left/right only had effect inside elements (spinners, ribbon grids). Now, the navigation is fully multi-directional : it selects the closest element in the requested direction (and can wrap around)
• Addition of vertical tabs. The main advantage is that more of them can fit in the screen and they can all keep displaying a title text.
• Revamp of the options menu using vertical tabs (separating the language selection to its own tab)
• Revamp of the help menu using vertical tabs (I improved many descriptions, and added some new sections)
• Sorting issues with several lists have been fixed. The sorting behaviour now always matches the triangle icon in the column, and sorting ghost replay by track name sorts according to the translated track name, not to the internal short name. The default replay sort is also much more useful, and the sort used is stable so sorting according to one thing won't make a mess in the order of replays identical for this thing.
• PNG image sizes have been significantly optimized, and redundant data between skins has been moved to a common folder instead of being duplicated. In the current git, it is possible to get an android APK with all official tracks fitting.
• Display nitro efficiency among kart stats in the kart selection screen, improve the values for the others bars
• Display times up to the ms in modes available in replay (time-trial and egg hunt), be it on the finish screen or when showing highscores in the track info screen.
• And - I guess you waited for it - a full revamp of the probability to get a given item. Sadly, some of my changes to fix inaccuracies have not been merged, and I have my doubts on the random number used to determine an item (I'll have to try again in the future to get this fixed). You can see my explanation of the changes here : https://github.com/supertuxkart/stk-code/pull/3216 In summary :

• Instead of having probabilities of item depending on if the kart first, in the 1st/2nd/3rd 1/3rd or last, the probabilities are different for every kart. There are reference points at 0 (1st kart), 0.25, 0.5, 0.75 and 1 (last kart), and a kart's probability to get an item will be a weighted average between two reference point. E.g. ; you have 4 karts. The 2nd kart will be at 0,33 ; so its probabilities will be 1/3 those at 0,5 and 2/3 those at 0,25 (1/30,5+2/30,25 = 0,33).
• Instead of having a single probability for all kart numbers, the probabilities change depending on the number of karts. Instead of using 20 probability list, 5 are used, and number of karts in between use a weighted average.
• A full rebalance of item probabilities using the new system. I wanted to fix all outstanding balance issues in the previous distributions. See the PR explanation for more details. And yes, parachutes are significantly less common now.
• Specific weights for FTL (no cake at all)

You can change this yourself, you have the mod power to do so and it has been agreed upon here :)

We're maybe 0,1% of the players, but we have an outsized importance in discovering bugs, tricks, etc. And we have a higher playtime. So overall, I'm adamant that we matter.

Also, exposing that timer in the UI would make more people try and do speedruns of STK, and this would be good.

The achievements field is not a great solution, in that it would still need to have an internal timer running to get the data, and this timer management is the major part of the code compared to displaying the current value. In particular, such a field would have to contain valid data when closing the game and reopening it, or switching profiles ; which is the most annoying part to manage.

I agree with you that current achievements are mostly worthless. In the latest git, you can even complete the 5-laps one by simply doing the Oliver Math's challenge (it has been bumped to 5 laps).

The issue is that, currently, achievements are hardcoded. This is not a great solution, and especially not a scalable solution (for 100 achievements, you'll have to add achievement-related code in many place, cluttering it). There have been talks of making achievements scriptable, but as this is low priority, no one has bothered to do so yet.

I look forward to your contributions, maybe you could help fix or improve some issues.

No, I'm not from the STK team. Je ne suis qu'un joueur qui a proposé du code et a réussi à en faire passer pas mal.

At the moment, completing the full game doesn't require to find hidden things like some stars in your SM64 example, nor is it very long (as all challenges runs show).

However, if additional challenges (like challenges in reverse) are added, we would probably reach the point where having access to the fort magma challenge without completing everything would be beneficial, certainly.

Also, in the next version, it will be possible to complete a GP challenge while finishing 2nd/3rd (finish 3rd in expert and you'll get the novice cup), so triple-skip will be possible for a standard any%.

Now, using any character is imho not really a speedrun-like thing. The speedrun-like thing is to choose the best possible tool for the job, or here the most efficient kart. However, the next version should ship with a kart characteristics rebalance making light and heavies much more competitive with mediums (in pure TT with top skill, heavies will be best I think ; but in normal races where you can get hit, this may not be the case ; overall I doubt the differences will be significant enough so that the time to change kart model in the middle of a run will be worth it).

It is currently stalled because maintainers are not very happy that I mixed the intertwinned acceleration, friction and parachute in the PR. You can look here if you want to test it : https://github.com/supertuxkart/stk-code/pull/3331

I think a first stage boost for skidding would be overkill, but you have a point about yellow boost not overriding the red one. I'm unsure about the best way to do it, however. Say you do the yellow shortly before the red is due to expire, when the red do expire, you'd still like to have your yellow boost active. This would require changing fundamentally how these boosts are currently coded.

For the basket ball, did you know there was some code which forbade having more than 2 basket balls obtained in, iirc, 25s ? The issue is that basket ball takes extremely long to reach the player, because the speed difference is small, and when one explode the player, the second one catch up quickly (and if you're unlucky, explodes you again). I had an attempt at a fix for this, but it has gone nowhere yet.

Also, I do absolutely agree that not hearing the basket ball arriving is absolutely infuriating. I myself lost significant time in many run attempts due to this issue. I actually opened an issue in STK's bug tracker here : https://github.com/supertuxkart/stk-code/issues/3246 ; you can make it known that you're affected too so it may get more attention.

The AI do not cheat. And it is way too dumb to use the item changers when you're just about to hit an item box (source : I greatly improved the code used by the AI to decide when to use that item, and even now, it isn't smart enough for this). However, obviously, the AI has the strengths of a computer : lightning reflexes for some actions (activate bubble shield against cake, etc) ; which can be sometimes frustrating.

For the speedometer, I expect you mean the one with the numbers ? In a racing simulation game, those numbers can be useful because a)in a straight line, you keep accelerating as long as you can and gain top speed ; b)you have a maximum speed you can take curves at without accident, so a number helps to gain greater precision.

Neither of this is really applicable to STK, and the numbers previously displayed were quite meaningless (no real unit like km/h). The visual speedometer has been significantly improved for accuracy (by myself), but I don't expect the devs to accept the return of the number indicator, as they have explicitly rejected it before.

For the speedrun timer, I have been trying to get it integrated, but I got some resistance. As you can see, there is a working patch for the 0.9.3 version ; I have submitted a significantly improved and expanded it, but the devs are reluctant to accept it, finding it too niche, as you can see in the comments here : https://github.com/supertuxkart/stk-code/pull/3258

Note that the current version of the PR has something broken in the loading between profiles. I haven't tracked down when this regression has been introduced yet. Here too, I suggest communicating directly your interest for this feature, be it through GitHub or through other channels like the IRC. If they see that there are more people than me being interested, there is a higher likelihood of them accepting it.

Btw, trivia : the whole options menu redesign, which required the creation of the vertical tabs widget, was in great part motivated by my desire to move the language selection to its own tab (done in the current git version), so as to remove the "the UI options menu is too cluttered to add new options !" argument against the timer's integration.

As far as I know, you can't select misc status per IL (the IL leaderboard lack options in general to match the fact that different versions have different available levels), it's a global thing. I guess it wouldn't hurt to remove the misc categorization for the time being, but there are only 3 IL runs for this version, so long term I expect it to move back to misc.

[quote]No custom user made tracks, as alayan says, this would be too many levels to track, potentially infinite.[/quote]

Do you mean, adding only addons tracks which are former official tracks removed in a more recent version of the game ?

This seems to be already more-or-less the case, with Bovine Barnyard, Shiny Suburbs, and some others which are in the IL list because they belong to one of the tracked versions, and so show up for all versions due to how the leaderboard works.

And there are already some runs for several of them.

While I wonder what the result may be in a few versions, I'm ok with this, as several of my aforementioned concerns don't apply to that kind of add-on track.

No need to worry, I'm not always sticked to STK or SRC. I can take a break. :)

My new dark theme has been integrated :

There are other improvements to STK's theme, some under the hood (reduce duplication) and some more visible, like much much more clear checked status for checkboxes.

Meanwhile, the code for online multiplayer has seen significant progress, but that feature is very complex for a small team, so while time-trials are already in a decent shape, there is still a lot of work to do.

At this point, it is very likely that the next version will not be called 0.9.4. But it may be 0.10.0 instead of 1.0.