GAR rules and support

[dreadnaut]

I'm a bit wary about adding GAR-only features to the website, since the original idea was an experiment, open to variations and dependant on the pipsqueaks' interest.

GAR feels to me very much like its own thing, which makes me question its place next to the main competition: car bonuses don't apply correctly, tracks can be awful under restrictive rules, public/quiet replays periods don't apply. To coexist properly with ZakStunts would require a number of code changes and a fair amount of duplication.

Does it make sense to run it in parallel with the main competition, or should it be a separate one? Can it run partially from the forum, as StuntsLOL and USC have done in the past?

Opinions, suggestions, support? pipsqueaks interested in the alternative scoreboard?

[Friker]

I strongly disagree with adding additional scoreboards in ZakStunts. Especially in case of GAR which is a stand-alone ruleset and a competition on its own. As far as I like GAR I cannot see any reason to have GAR in ZakStunts. Lack of Marco's coding skill is not an excuse - there are many ways to have a competition without a website.

[Duplode]

I believe any parallel races sharing the ZakStunts tracks and schedule should remain unofficial, that is, without a season scoreboard. Having a separate competition so tightly coupled to the main one would be suboptimal for everyone involved. Also note that a proper title to fight kind of presupposes that rules like IRC and NoRH compliance to be strictly enforced and verified, beyond a gentleman's agreement.

[zaqrack]

I'd say I agree

[Alain il professore]

WHY A GAR SIDE-SEASON SCOREBOARD HELPS EVERYONE
(an invitation to discuss a light-touch, win-win setup)

I've seen the concerns about adding GAR-only features to the main site: different rules, car bonuses not mapping 1:1, track design that can feel "awful" under restrictions, and the quiet/public replays cadence not fitting. Fair points.

But there's a middle path that preserves ZakStunts simplicity and gives pipsqueaks who enjoy NoRH/OWOOT a real thing to race for: a GAR side-season scoreboard. Separate lane, same monthly schedule, minimal coupling.

WHAT "GAR SIDE-SEASON" MEANS (IN ONE PARAGRAPH)
GAR = Gentlemen's Agreement Rules: NoRH + stay on/over the road with clear, documented checks. It's a stricter, "race-craft first" mode. We keep it parallel to the main comp (same ZCT track/month), but results are tallied on a separate page/thread. The main site can remain unchanged or add a single "GAR results" link per race—nothing more.

WHY IT'S WORTH DOING

Keeps the main comp clean
No rules entanglement, no code forks. Main scoreboard stays "anything-goes with RH." GAR points live elsewhere. Simple.

Low lift, proven format
We're already running GAR as an unofficial parallel scoreboard and it works, even attracting new (and returning veteran) pipsqueaks. Season numbers were compiled with minimal tooling and posted alongside the regular news. This is not theoretical; it's established practice.

Onboarding + inclusiveness
New and returning players who bounce off RH meta get a welcoming lane. GAR rewards classic skills (lines, braking, consistency) and lowers the "secret-route" barrier to entry. That's healthy for the community.

Better track feedback
Running a GAR tally in parallel reveals which monthly tracks translate well to fair racing and which ones crumble under restrictions—useful data for future design without forcing changes on the main comp.

Two kinds of excellence, both celebrated
Let the main comp keep rewarding sandbox ingenuity. Let GAR spotlight pure driving and consistency. More reasons to race; more replays to watch.

Minimal engineering
"GAR Side-Season — ZCTxxx" with a maintained season table.

HOW WE'D KEEP IT SIMPLE

• Schedule: identical to ZakStunts (same track, same deadlines) -done.
• Submissions: "post your RPL + short compliance note" -done.
• Ruleset: link to the standing GAR rules reference -done.
• Car bonuses: either freeze to 1.00 for GAR or publish a simplified GAR-specific table once per season (there's some work I'm not wishing to do, so I'd expect imposing a single car for GAR...)
• Points: reuse ZakStunts scoring or a simple top-12 scheme -this is also already done.
• Verification: community review on request (newcomers, suspicious gaps) with the standard check; a team of three independent moderators arbitrate edge cases.

WHAT THIS IS NOT

• Not a takeover of the main site.
• Not a rules war.
• Not extra workload for track authors. Tracks stay as is; just think about GAR pipsqueaks unplayable traps.

PILOT PROPOSAL

• Run a side-season (same ZCTs), starting January 2026.
• Volunteers: 2–3 people to rotate verification at the end of the month. (I can help if needed.)

CALL FOR INTEREST

If you'd race or help maintain a GAR side-season, reply below with:
• "RACE" (you plan to submit laps), and/or
• "MOD" (you can help verify rpls), and
• any preference on car bonuses (1.00 for all vs. fixed table).

Two scoreboards, one community. Keep the main comp wild and creative; give GAR fans a fair, friendly season to chase. Low effort, high upside—let's try it.

Thank you.

— end —

[Victor Narl]

I fully support the introduction of a final scoreboard for GAR, including at the bottom of the main page of the site https://zak.stunts.hu/seasons/2025 with the caption "Alternative competition" It is unclear why(!) the competition is held at all if there is no final scoreboard? What is the point of a monthly scoreboard then?

[Argammon]

I believe it would be a good idea to introduce a GAR season scoreboard. The cleanest way to implement this would be to start in 2026, since GAR was only added as a "test" partway through the current season.

However, there is one point I find puzzling:

On the one hand, members of the Hyperion team have expressed concerns about cheating in RH competitions. Examples they've mentioned include:

-Track authors sharing tracks in advance with certain pipsqueaks

-pipsqueaks downloading non-public replays

-pipsqueaks modifying replays via hex editing

-Admins viewing the times of hidden replays

To their credit, they have also proposed measures to address some of these issues.

On the other hand, they actively support NORH competitions (such as GAR and Dosreloaded), arguing that the experience is much closer to "real driving." Yet here lies the elephant in the room: cheating in NORH competitions is arguably much easier, since using RH cannot be detected. Why, then, is there no apparent concern about this type of cheating? I would appreciate some clarification.

To help address this, I propose that all NORH racing be conducted on a (ZakStunts?) server using @Daniel3D's NORH version of Stunts. Hopefully, it will not be too difficult to ensure that the server version runs smoothly on most of our systems.

[Matei]

To help address this, I propose that all NORH racing be conducted on a (ZakStunts?) server using @Daniel3D's NORH version of Stunts. Hopefully, it will not be too difficult to ensure that the server version runs smoothly on most of our systems.

There are ~2 problems:

1. Daniel3D's NORH version is outdated.

https://github.com/AlbertoMarnetto/restunts/tree/supersight

Daniel3D first has to take these modifications and add them to his version.

2. Do people have to play on the internet through the browser like at archive.org? I can't even play it there, Firefox too slow and I have to turn JavaScript on.

[Victor Narl]

On the other hand, they actively support NORH competitions (such as GAR and Dosreloaded), arguing that the experience is much closer to "real driving." Yet here lies the elephant in the room: cheating in NORH competitions is arguably much easier, since using RH cannot be detected. Why, then, is there no apparent concern about this type of cheating? I would appreciate some clarification.

Usually, RH can always be distinguished from NoRH with careful viewing and it is not really possible to cheat here.

for example:
The RH replay clearly shows that it has all the hallmarks of a replay-processed race, such as overly smooth corners and more near-runs than you might expect.
these are not my words, but the words of Overdrijf https://dosreloaded.de/forum/thread/7264-4d-sports-driving-stunts-jahrescontest-2024/?pageNo=4

[Argammon]

Usually, RH can always be distinguished from NoRH with careful viewing and it is not really possible to cheat here.

for example:
The RH replay clearly shows that it has all the hallmarks of a replay-processed race, such as overly smooth corners and more near-runs than you might expect.
these are not my words, but the words of Overdrijf https://dosreloaded.de/forum/thread/7264-4d-sports-driving-stunts-jahrescontest-2024/?pageNo=4

Sorry, but this is not correct. Even if a pipsqueak only uses RH for "crash recovery" they get a massive advantage in NORH racing. Using it only for crash recovery is undetectable.

For example, when I am trying to get a top NORH time i am driving so risky that I only finish 5%-10% of all runs. (The 90-95% of non-finished runs also include runs in which I violated the ruleset, for example OWOOT).

If I were to continue racing after a crash, I would be able to finish many more strong runs than I would be able to do otherwise. And of course, finishing 40 strong runs likely gives a better time than finishing 2-4.

[Matei]

I don't understand what the problem is. RH is the most important feature of the game.

https://en.wikipedia.org/wiki/Next_(2007_film)

Cris Johnson can see into his future. He can only see two minutes ahead,

That's about as long as a track takes.

Cris is able to walk right up to the terrorist leader while using his power to dodge bullets.

That's like playing Quake, actually.

https://speeddemosarchive.com/quake/

Is there any game that is not a racing game?

[Victor Narl]

Usually, RH can always be distinguished from NoRH with careful viewing and it is not really possible to cheat here.

for example:
The RH replay clearly shows that it has all the hallmarks of a replay-processed race, such as overly smooth corners and more near-runs than you might expect.
these are not my words, but the words of Overdrijf https://dosreloaded.de/forum/thread/7264-4d-sports-driving-stunts-jahrescontest-2024/?pageNo=4

Sorry, but this is not correct. Even if a pipsqueak only uses RH for "crash recovery" they get a massive advantage in NORH racing. Using it only for crash recovery is undetectable.

For example, when I am trying to get a top NORH time i am driving so risky that I only finish 5%-10% of all runs. (The 90-95% of non-finished runs also include runs in which I violated the ruleset, for example OWOOT).

If I were to continue racing after a crash, I would be able to finish many more strong runs than I would be able to do otherwise. And of course, finishing 40 strong runs likely gives a better time than finishing 2-4.

there is a wonderful program Replay Analyser (RA) by Robert Riebisch http://www.kalpen.de/luke/4dreplays.html it can be used to determine the presence of RH for each pipsqueak based on the statistics of the keys pressed, it can also be further developed so that even with single recoveries after accidents it would be possible to determine the presence of RH

[Argammon]

Sorry, but I don't think that conclusion holds.

Replay Analyser (RA) is excellent for spotting sustained RH patterns, but it can't reliably prove or disprove a single, targeted rewind used only for crash recovery. Consider this concrete case:

Start: Driver X begins a NORH attempt.

Crash: In the final corner, they lose control.

Rewind: They rewind to the previous straight where the car runs steady and only "accelerate" is pressed.

Finish: On the second try, they take the corner cleanly and finish.

What RA signal would reveal RH here? On the straight, key‑press statistics are indistinguishable from a normal clean run. Key-count distributions alone won't expose this kind of targeted rewind. If you believe it can, please outline a step‑by‑step method.

This isn't a defense of using RH in NORH—quite the opposite. It's about aligning on what is and isn't detectable so we focus on real problems, rather than speculative ones (for example drivers downloading hidden replays). If there's a robust, reproducible procedure, please share it so we can test and adopt it.

[Victor Narl]

a deeper analysis of the replay is needed, for example, recovery after an accident usually occurs in certain places, and a normal press of the gas pedal in this place would be strange, On a straight line it's also strange to press the gas several times - then there was RH

also, together with the replay, you need to analyze the HIG file, if the HIG was tampered with, then this can also be determined by examining the last seconds of the replay for evidence of pressing the gas pedal just before the finish, also the time of saving the replay and HIG should be identified to avoid tampering

in NoRH competitions need to demand both replay and HIG

[Duplode]

While I'm not opposed in principle to having a side season scoreboard for GAR, it's important to have clarity on what it would (or wouldn't) mean for the competition.

It is unclear why(!) the competition is held at all if there is no final scoreboard? What is the point of a monthly scoreboard then?

The alternative rules side scoreboards (there were other ones at the beginning) were added in 2016 as a way to make it possible to race for fun on the ZakStunts tracks under these rules, and to give visibility and recognition to the informal racing that was already happening spontaneously. The GAR scoreboard was never meant to become a full-blown parallel competition, or to have parity with the main scoreboard. We already have a proper OWOOT competition, with OWOOT-first track design, next door at Race For Kicks, and I don't see the point in trying to compete with R4K on that turf.

there is a wonderful program Replay Analyser (RA) by Robert Riebisch http://www.kalpen.de/luke/4dreplays.html it can be used to determine the presence of RH for each pipsqueak based on the statistics of the keys pressed, it can also be further developed so that even with single recoveries after accidents it would be possible to determine the presence of RH

Sorry, but that's a hard no. It's not possible to reliably prove or disprove RH from keystroke patterns. Going down that rabbit hole will only fuel conspiratorial thinking and witch-hunts, two things we absolutely do not want here.

To help address this, I propose that all NORH racing be conducted on a (ZakStunts?) server using @Daniel3D's NORH version of Stunts. Hopefully, it will not be too difficult to ensure that the server version runs smoothly on most of our systems.

I don't think that's a good fit for ZakStunts either. An important part of the target public of the GAR side scoreboard are new pipsqueaks who are still getting used to the idiosyncrasies of free rules RH racing. Making the game less accessible by imposing a game version separate from the one used in the main scoreboard and/or use of a server platform runs counter to that.

I don't understand what the problem is. RH is the most important feature of the game.

There might be a touch of hyperbole in "most important feature", but there's definitely a grain of truth in it (the speedrunning comparison is also apt).

One analogy I like for the RH/NoRH, free/OWOOT and async/live dualities is with the apparatuses in Artistic Gymnastics: modalities of the same sport, each one drawing from different but overlapping skill sets, and each of them just as valid as the other.