ZSR Leaderboards

[mzxrules]

http://www.zeldaspeedruns.com/leaderboards

If any times seem fishy, please post here about it.

[mzxrules]

The issue we had with the leaderboards earlier today should be resolved.

[benstephens1000]

I've got something that seems a little fishing... maybe I'm just overreacting here but a guy just got the "WR" for OoT3D any% NG+ and I can't say I think this is legit. This guy can be found nowhere. He's not on the forums, not on twitch, not in the irc, and miraculously he is second to only to EnNopp in MM 10 mask by a mere 6 seconds. What's even weirder: Both runs were made on the same day. What's even weirder than that: Neither run was submitted until after TSSB. That's right I'm calling shenanigans here. Also it can be noted that there is no video evidence of either of these runs.

[Pedalpowertoast]

Fox, me and a few others also deducted the exact same things as you We are master detectives.

[Yashichidsf]

I think he should be sentenced to death without parole.

[Cosmo]

proposing a policy change for OoT/MM leaderboards:

1. only times with video proof may be verified

even if a verifier saw a run live, without video proof to look back upon, that verification means nothing. we want all runs that are verified to be archived in a way that lets anybody look back and rewatch them, double check them, retime them, etc. therefore a time without a video shouldn't be a valid run to be verified.

2. only times on official releases may be verified

we at ZSR have (arbitrarily) chosen PJ64 1.6/1.7 as a baseline to use for emu runs. while this has been successful in increasing the number of submissions, it was chosen under arbitrary conditions: "it isn't faster than the fastest version." this is only somewhat true. for example, emu has N64 style resets yet a full fps file select screen, with (approx) VC lag during gameplay, yet slightly slower loads. it is faster/slower depending on the situation in the game. due to this we've had to ban the OoT 1.0 rom from being used on PJ64 in order to 'balance' emu against categories in which N64 has an advantage, as emu would obsolete N64 if using the OoT 1.0 rom. my point is that PJ64 1.6/1.7 is an arbitrary platform. continuing from this, due to the large amount of uncertainty in the exact emulator build used (there are several builds of PJ64 1.7 for example), the pc hardware that is used (graphics card, operating system, etc), the plugins that are used (there are many that give different results), and the settings of those plugins (which can be easily tweaked and change the speed of the game slightly), there are often many emu runs that operate at different speeds in various circumstances. finally, there is a concern about using memory watching in a live run to gain more information than normally possible (for example, have all the blue warp timers on your screen when you do a run so you never miss a wrong warp). due to these unknown factors, emu runs are not properly verifiable, and therefore shouldn't be verified. this isn't about banning emu, it is about not verifying runs that do not have any proper method of verification, especially on a platform that was arbitrarily allowed in the first place, and has concerns about easy potential cheating.

if we can agree on these proposals, we should then unverify all OoT/MM runs without videos, and all OoT/MM runs without an official release, to improve the quality of the zsr leaderboards verification system. this seems like the best compromise: we keep running OoT/MM accessible, while also keeping it legitimate.

thoughts?

[Zas]

I completely agree. This will also hopefully encourage more people to move from emulator to official releases.

[Jbop]

I'm fine with both of those points. While we're at it though, how about making it so that people can't verify their own runs?

[Cosmo]

I'm fine with both of those points. While we're at it though, how about making it so that people can't verify their own runs?

this is a good idea

[Tr621]

I'm fine with both of those points. While we're at it though, how about making it so that people can't verify their own runs?

That's what I've always thought should be the case for all games, but what do I know?

[Apasher]

Could we implement notifications for ZSR editors for when a run is submitted to the leaderboards? This makes it so that we won't be far behind in verifying runs.

[Zefie]

I fully agree with what Apasher has said and will show my support through this post.

[Apasher]

With that being said, I would like to come up with another proposal for the leaderboard itself:

Hide unverified runs by default, and implement the option to show them. I feel like having unverified runs shown by default defeats the purpose of the verification system in the first place since potential fake times could still be seen.

Thoughts?

(Also, seconded Jbop's proposal of not letting editors verify their own runs)

[doktor_m]

I also agree with apasher. The default leaderboards should only show verified runs.

[Irregular programming]

I also agree with apasher. The default leaderboards should only show verified runs.

I'm all for that but then there has to be more people who verifies runs.

I've been speedrunning on official consoles since summer 2014 and I've had one run in total verified. I get that people don't really have time to (or want to) verify runs that are bad, and I totally understand that (I wouldn't have time to verify 30 minute any% runs). However if the leader boards are gonna default to verified runs off unless you are really good there is not much reason as it is now to actually post your run as it's most likely not going to be verified anyway.

I am not at all critical against those who verifies runs, but I would actually like to see unverified runs on the ZSR boards per default as in 99% of the times those runs are probably OK and are just as valid as any other run on the board.