Game Time vs Real Time

User avatar
MeGotsThis
Schoolkid
Posts: 25
Joined: Thu Apr 24, 2014 2:55 am
Location: Middle of No Where

Re: [TALK] Game Time vs Real Time

Post by MeGotsThis » Fri May 09, 2014 3:29 pm

Thomaz wrote:I have one question that's important for me.

I do segmented runs of Blue and Red. Right now, as far as I know, I have the segmented WR for glitched Red any% (using Item Underflow). How would you guys handle timing such a run? Especially interesting because there are no real segmented runs on the site at all (a few S+Q's here and there doesn't count as a segmented run imo).
Right now there are no multisession runs for NSC, No Save Corruption. For NSC, Timing is by RTA, where timer starts at New Game and ends when screen fades to white after Hall of Fame in game time is shown. Source. The reason it is done as RTA it is because of the save and quit needed for brock skip. Right now the only way to use Game Time is only single segment runs (no save and quit). Don't forget to check the leaderboards here.

Thomaz
Preschooler
Posts: 5
Joined: Fri May 09, 2014 9:15 am

Re: [TALK] Game Time vs Real Time

Post by Thomaz » Fri May 09, 2014 9:27 pm

I realise that. It doesn't take away that I currently have the fastest ingame time. Just because there's no category here doesn't mean it doesn't exist or should not be allowed. I was just wondering what people thought of it.

ExtraTricky
Bug Catcher
Posts: 21
Joined: Thu Apr 24, 2014 12:39 am

Re: [TALK] Game Time vs Real Time

Post by ExtraTricky » Fri May 09, 2014 9:48 pm

I've decided that I don't care about world records except as a thing to compare my PB to. If I do more runs (I probably will eventually, but who knows when), I'll be using real time for my PBs, though I'll list game time as well as long as the community continues to use it. If my PB becomes competitive again, then I will go with the community's conventions for what is considered WR.
werster wrote:The game timer is accurate, even if it has a slight disparity to what we call a minute that is 100% irrelevant, the accuracy within itself and it's comparisons to other versions of itself are all that matter
I agree with this, but in my opinion game time's usefulness in comparisons to itself is greatly reduced by the fact that it's only displayed to the minute. I don't think the fact that the frame-to-second conversion is off is a problem.
shenanagans wrote:A point to add to this. RTA is really only needed in very fast runs (like 45 minutes to an hour in length max). In glitchless runs there is no reason to use RTA, because any 1:51 is an amazing run and deserves to be called a "WR", however in very fast categories somebody could say fight a whole extra trainer and still achieve the same IGT as another runner.
This happened in glitchless at 1:57. The run was actually faster real time as well, despite losing 1min 12s to fighting an extra trainer. If you said back then that any 1:57 is an amazing run and deserves to be called "WR", I would strongly disagree with that, having had four runs at 1:57 of significantly varying quality.
shenanagans wrote:Limiting ties does not matter to me, if two people have a nearly identical run but somebody loses 2 seconds to a crit or whatever who's to say that run is better. But when Ties become ridiculous there may be a problem, as the runs begin to crowd the WR page (see the old SS glitched 5 way tie) I think RTA can be a good option.
Did the 4-way glitchless tie at 1:54 count as ridiculous? What about the tie where werster's PB and 4 of my runs were tied at 1:57?
jolteontrainer wrote:Based on this, there would be more competition to break that tie for more godly runs.
I believe that only considering times to the minute discourages competition for two reasons. First, if the time is 40+ seconds from improving the minute, then people will be less inclined to go for that minute threshold than they would be if they were just trying to get a better time by a few seconds. Second, from what I've seen nobody is motivated by breaking a tie (except maybe extrinsically since it gives other people more reason to give encouragement). The motivation is generally improving one's PB.
vulajin wrote:This literally does not matter. Is game time consistent with itself? If you run the game twice in a row the *exact* same way, and complete the game in the exact same amount of real time/frames/whatever on the same console twice, do you get the same in-game time? If so, then the fact that it doesn't accurately reflect real time is completely irrelevant. No one in any other game community cares about this as long as the game time permits accurate comparison of runs (i.e. if game time is lower, the run was faster in a way that matters).
The pikachu cry issue means that two runs can have the same number of frames but have different ingame times, or even ingame times in the wrong order. The chance that it crosses a minute boundary is quite small, but there.

entrpntr
Pokémon Trainer
Posts: 141
Joined: Wed Feb 04, 2015 11:22 pm

Re: Game Time vs Real Time

Post by entrpntr » Sat Jun 20, 2015 1:04 am

Since there are still regularly fusses over this and it doesn’t look like there was resolution the last time around, I figured I would throw my two cents in and try to get some closure on this (it’s clearly not going away on its own).

I am working from the premise that when the game provides you an in-game timer, there needs to be a compelling reason not to use it. Switching to real-time introduces a whole slew of issues that can be avoided by using the in-game timer.


Considerations (and my personal opinions thereof)


(1) What is the purpose of the leaderboards?

To foster healthy competition and to provide a maintainable reference of the best completed runs.


(2) What attributes should be weighed when choosing between IGT / real time?

Must-haves: Unexploitability, intuitive alignment with what ‘going fast’ means, consistency/accuracy/verifiability

Nice-to-haves: Ease of reasoning about, granularity of time units, positive influence on the aesthetics of the run

Other: Interesting strategic/routing differences in IGT vs RTA (to warrant 2 categories, for instance)?


(3) What things would be affected materially by a switch from IGT to RTA (wouldn’t necessarily change strats)
  • PC strats: Gen 1 (HOF counts towards IGT), Gen 2 (HOF doesn’t count towards IGT)
  • Pikachu sound effects in Yellow
  • Waiting through E4 credits in GSC
  • Lag/loading times in Gen 3
  • 59.7275 frames per real-time second, 60 frames per game-time second
  • Others?
Commentary: Pokémon has a lot going for it with its in-game timer compared to other games. Only a few minor changes would be made for RTA runs, and the things that do exist are easily explainable under the IGT paradigm. We also can do RAM watches on the in-game timer and have the benefit of the disassembly for Gen 1/2, so things are far less murky when it comes to measuring how long things take in game time.

So basically the only legitimate reason I see to switch would be that the game time is only displayed to the minute. However, I take the opposite stance of ExtraTricky; I think the granularity of real-time doesn’t provide enough benefits to warrant the switch. Game time is achieving the purpose of the leaderboards and doesn’t present glaring issues in the quality of speedruns produced.


Problems with real-time

I think the cons with real time run far deeper than most people think.

The two things you get from a typical Twitch VOD is the video of gameplay and splits. I don’t believe Twitch VOD lengths can be trusted, but maybe things have changed (doubtful). Also, splits aren’t consistent between runners nor are they completely accurate (humans don’t have pinpoint precision with timing nor perfect reaction times). I would think this is too obvious to even mention, but after Gunner’s 3:20 in Crystal, there was an uproar from multiple runners over the fact that he was ahead of Keizaron’s 3:19 splits by 0.02 seconds (never mind the fact that Keizaron starts his timer about a second before Gunner).

For verification, it would certainly be possible to take the timing between a standardized start time and end time. However …


Technical Issues introduced by RTA

I don’t know how well split programs can be trusted. A LiveSplit developer gave a detailed answer about how time is kept here: http://www.reddit.com/r/speedrun/commen ... er/cfi92tz. It’s clearly not a science, though if we only care about granularity to the second, it might be “good enough”.

However, there has been one issue where this presented a major accuracy problem. Jadin (Link to the Past speedrunner) noticed a major discrepancy between RaceBot on SRL and his timer and ran some tests. Some pictures with context that he posted to Twitter: (1) http://i.snag.gy/SZ3jr.jpg, (2) http://snag.gy/ykUn9.jpg. When he got the tip to the reddit comment above, he was able to resolve his issue by disabling the HPET in his computer’s BIOS and enabling the fallback timer in LiveSplit (Llanfair and Wsplit remained broken). He retimed his Link to the Past PB, and it was changed from a 1:24:40 (9 seconds from WR at the time) to 1:24:51.

Tweets for posterity:
(1) https://twitter.com/apparentlypants/sta ... 9517097984
(2) https://twitter.com/apparentlypants/sta ... 1902478337
(3) https://twitter.com/apparentlypants/sta ... 5562365952
(4) https://twitter.com/apparentlypants/sta ... 4006067200

This is only one instance, but I don’t think many know about the issue, so I wouldn’t be surprised if there were other timing inaccuracies that have been overlooked that relied solely on splits. In any event, using real time introduces these technical issues that runners and verifiers need to know about in order to maintain an accurate leaderboard.

Of course other communities can and do get around these issues by doing a framecount on their local recordings from their capture devices. But there is another host of issues that I don’t know have ever been resolved; see cyghfer’s post here: https://forum.speeddemosarchive.com/pos ... conds.html.
cyghfer wrote: Now, there's still one big problem with all this - Game Boy runs. Neither the Super Game Boy 1 (well-known to run at ~61.17 FPS), the Super Game Boy 2 (seems to run at the SNES' frame rate, not the GB's frame rate, from some hasty tests of mine), NOR the Game Boy Player (runs slightly faster than an actual Game Boy from a quick test of mine, not sure of the exact frame rate though, help wanted here!) run at the exact speed the Game Boy, Game Boy Color, and Game Boy Advance run at, which is:

262144 / 4389 ~= 59.72750056960583 [bsnes source code] (I've read in multiple locations that the GBA frame rate is the same as the GB/C frame rate, not 100% sure though)

In order to get accurate-length recordings, probably the only solution is to record in 59.7275 anyway (regardless of the device used) and accept the periodic frame drops. I recommend that official timing conversions be determined for all 3 devices, though, for people who may not want dropped frames in their recording for whatever reason (or who record without knowing all this information).
Final words

If there is a decision to go down the RTA road, there is a lot that needs to be figured out and console runners / verifiers have even more things that they would need to know in order to get accurate recordings and timings. It adds a lot of administrative/maintainability issues that you don’t get with game time. Most other communities have worked around this, but the vast majority of such communities lack a reasonable in-game timer. If the community thinks the minute granularity of IGT is that intolerable and that the issues with RTA are manageable, then I suppose a switch to RTA would make sense. I personally think it’s a needless burden.

The first game that jumps to mind that has a game-timer with minute granularity is Super Metroid. Zoast’s SDA submission explains why Super Metroid switched from game-time to real-time:
zoast wrote:There are quite a few reasons why real time is the standard now. It's a more fluid run, completing the game without a single pause and generally avoiding lag. This involves different routing such as delaying the collection of the Ice Beam until later in the run and completely skipping a number of items such as Spazer, Ridley's E-tank, and the crab supers in Maridia. Many, many strats such as fast pillars (avoiding two door transitions by short charging to get through the pillars in Lower Norfair) have added excitement and new tech to the run. Also, the fact that the in-game timer only displays minutes means a considerably worse run could be considered a tie with another run. A lot of us on SpeedRunsLive (SRL) like to race this (there is a big weekly Friday night any% race), which obviously makes more sense in real time as well.
That the minute granularity was apparently a tertiary reason in a hyper-competitive/execution-heavy game as Super Metroid is telling. The concerns ExtraTricky mentioned never really came to life (to my knowledge) in Super Metroid when they were using game time, and I think it’s awhile before the state of Pokemon speedrunning approaches this (it’s been 13 months since ExtraTricky's post and times continue being pushed down). However, I can see how some people would be turned off because of the minute granularity, so I wouldn’t be opposed to unofficially tracking RTA times somewhere (no idea exactly how this would be done, but it’s a thought).

Last thought: anyone who resets before the E4 credits finish in the GSC games should be ridiculed mercilessly, regardless of the circumstances.

User avatar
Amoeba
Cooltrainer
Posts: 280
Joined: Thu Apr 24, 2014 12:26 am
Contact:

Re: Game Time vs Real Time

Post by Amoeba » Sat Jun 20, 2015 12:47 pm

I'm going to be devil's advocate/annoying/a general arsehole and pick a point with one sentence in your post.
entrpntr wrote:I am working from the premise that when the game provides you an in-game timer, there needs to be a compelling reason not to use it.
Why?
Why shouldn't it be the other way around, why shouldn't we require a compelling reason to use an in game timer as apposed to RTA?
~

entrpntr
Pokémon Trainer
Posts: 141
Joined: Wed Feb 04, 2015 11:22 pm

Re: Game Time vs Real Time

Post by entrpntr » Sat Jun 20, 2015 3:15 pm

Amoeba wrote:I'm going to be devil's advocate/annoying/a general arsehole and pick a point with one sentence in your post.
entrpntr wrote:I am working from the premise that when the game provides you an in-game timer, there needs to be a compelling reason not to use it.
Why?
Why shouldn't it be the other way around, why shouldn't we require a compelling reason to use an in game timer as apposed to RTA?
I kind of went into why I think the burden falls on RTA advocates—you get a lot of annoying issues to deal with, instead of being able to just use the timer the game provides you with. Unless the in-game timer is inaccurate, abusable, or otherwise corrupts the quality of the speedrun, you're taking on a lot of administrative issues when you could just be focused on playing the video game.

Lemonian
Preschooler
Posts: 6
Joined: Fri Aug 01, 2014 2:32 am

Re: Game Time vs Real Time

Post by Lemonian » Wed Jul 01, 2015 5:01 pm

For all the hassle that I read about this I am going to have to with IGT.

I don't see a reason to use RTA if IGT is accurate between all players.

Patastrophe
Preschooler
Posts: 7
Joined: Mon Feb 22, 2016 5:23 pm
Location: Cleveland, OH

Re: Game Time vs Real Time

Post by Patastrophe » Tue Jun 28, 2016 7:06 pm

Image

If we want to get more specific - just actually display the values that are in memory already. 0 HR : 1 MIN : 29 SEC : 47 FRAME

User avatar
Luckless
Site Admin
Posts: 113
Joined: Fri Apr 11, 2014 3:28 am
Location: UK
Contact:

Re: Game Time vs Real Time

Post by Luckless » Fri Jul 22, 2016 8:37 pm

It's been brought to my attention that the existence of this thread is causing some confusion since timing is being revisited. The recent discussions are in another castle: viewtopic.php?f=117&t=607. Also this thread was just for Generation 1, while that one is for all generations.

Before I lock this, I'd just like to address what happened here: While looking specifically just at Generation 1, the discussion was inconclusive. Although there were many positives to switching to RTA, they didn't have enough weight to actually make the switch at that time. This doesn't mean that the switch won't happen now that timing is being revisited for all generations. We want the timing method to be what the runners want and it’s clear to me that more people are in favor of RTA now than they were before. Now this is probably because it includes other games which have issues regarding in-game time too, (for example; Gen 3) but also; moving the entire Pokemon genre to RTA feels like a more wholehearted decision than just Gen 1 on it’s own, so I feel like more Gen 1 runners have also had a change of heart too. Anyway, that's all folks! Locked.
"I see now that the circumstances of one's birth are irrelevant. It is what you do with the gift of life that determines who you are" - Mewtwo

Image Image Image Image Image

Locked

Return to “Generation 1”

Who is online

Users browsing this forum: No registered users and 7 guests