Snowflake's Feed

Your argument is doomed to failure, from the outset, reason being that no one is actually analysing INPs. All I've seen, of late, is various members' putting up screenshots of basic data displays, which come from MAME itself.

Barthax liked this post

thats fair. to be clear it realy is more a question than an argument is i acknowledge i know not much on mame. but your point is still valid. screenshots should not be confused for analysis, i hope noone is treating them as such

Originally Posted by Snowflake

screenshots should not be confused for analysis, i hope noone is treating them as such

.
For better or worse, your hope is probably in vain. As more disputes of unchecked INPs are raised, a lot of members are going to lose a lot of credibility.

I'm not sure we can declaratively state that the recent slew of "INP analyses" were not the end result of actual analyses, notwithstanding the fact that the data presented can be generated without actually analyzing. Has anyone thought to ask the analyzers if they actually performed an INP analysis, or it just assumed that they didn't because of the possibility that they didn't?

I don't see any reason why someone who performed an INP analysis would post screenshots instead of the analytical results, so it's safe to assume that the screenshot posters have not performed any analyses, and likely don't know how to. The ultimate proof is that unacceptable submissions have been accepted.

We may live in hope that voters will stop relying on screenshots, and learn to perform analyses, so that all MAME submissions can be properly adjudicated.

if i'm misreadin anyone i apologize, but i think you're both saying different things

garret i think you're talking about the results of analysis printed off, and taking A.Ds comments to mean the person running the analysis didnt truly look at the results

I think A.D. though is refering to truly no analysis being done, not even running the analysis program. just taking screenshots of settings. i've certainly seen some such subs were people wait for the analysis, but then approve once just screenshots of settings are shown

In no particular order, the most important aspects of INP analysis are:-

1. Playing the INP, in full, and matching the score shown by that playback to the score being claimed
   
2. Checking the hexadecimal code of the INP for relevant information, such as MAME version used, and settings
   
3. Checking the file properties and hexadecimal code of the INP for anomalies
   
4. Researching ROM history, to check for content alterations, title revisions, etc.
   

swaggers, Barthax, Garrett Holland liked this post

Originally Posted by Almighty Dreadlock

Researching ROM history, to check for content alterations, title revisions, etc.

A reminder to the general audience, there's a pair of wiki pages to guide on this subject: https://www.twingalaxies.com/wiki_index.php?title=Game:Exploring-MAME-history-for-TG-rules and a fictitious example (to highlight the various points) at https://www.twingalaxies.com/wiki_index.php?title=Game:Exploring-MAME-history-for-TG-rules-Elevator-Action-example [it "converts" an arcade set of rules which obviously will be missing various MAME-specific info into MAME-worthy rules].

Snowflake thanked this post

Almighty Dreadlock, nads liked this post

I normally don't adjudicate MAME since I don't have the skill-set to analyze INP files, but I've heard this topic come up here and there over the last year - a build-up of MAME submissions due to (as I understand) a process that traditionally relied on volunteers to break out/break down INP analyses that were posted to the thread for other members to scrutinize. Maybe it was wrong, but I got the sense that other members would treat these analyses as an "all clear" signal that everything checked out.

Are there specific reasons why MAME submissions - at least as far as video recording goes - don't follow some of the guidelines for submissions on emulator for computer/consoles? eg. start recording from desktop, launch program, load ROM, highlight setting through menu highlights, etc? Is the issue that it's easier to still "cheat" at MAME even if one walks through the steps as is typical with EMU submissions for console/computer?

Kelly Kerr liked this post

As far as I am aware, unlike all other platforms, MAME provides a method to scrutinize the state of the performance every frame. The things that could otherwise be hidden by not including bootup, etc., would be exposed within the data in the INP.

There are a bunch of circumstances which are traditionally caught through MAME INP analysis because MAME (until ~0.145, IIRC) didn't display any of the recorded DIPs. Once the recoreded DIPs were displayed by the playback there were still exceptional circumstances. In particular with multiple older games, if booting up with one set of DIPs they would dictate the entire original recording even if the player innocently wasn't aware they had to quit MAME after changing DIPs. The older MAME analysing peeps (mainly me, really) would draw on the knowledge of previous scans until @terencew and I started pooling that knowledge into https://github.com/Barthax/TGDips - you'll see version sub-folders because the combination of different versions & games would require different positional indicators in the hexadecimal view of the INP.

Yes, it also became a source of "erroneous green light" for the epitimous blind voter.