MAME - Which version for the new era of TG?

Some different points:

WolfMAME (any version) is the best option available. The fact the source code is available means that cheaters have the option of cheating - it isn't simple but it is possible.

If TG produces a new MAME, the source code must legally be released as this is a requirement of the MAME licence. This will likely put the TG MAME into the same category as WolfMAME and ergo the same possibility of cheaters. Additionally, WolfMAME has many years of testing at MARP for flaws.

Video feed (web cam or other) of MAME play is more open to cheating than WolfMAME.

If the version of WolfMAME is changed to a single new version, the ROM sets will need to be reviewed as many sets have altered over the course of time.

If any version of WolfMAME is made available, many different computers can be used. However, the rules will need to be changed to accomodate the changes of ROM sets used as many have changed over time.

Personally, I'm in favour of any version of WolfMAME. Personally, I probably have less than a tiny influence on the decision. Something will happen, something will get evaluated and something will be the new way of things.

why cant i see my post on this thread ?

It's on the previous page.

NOT entirely true!

The license that is easiest to use does that. However it is possible to negotiate a new license with the major developers of Mame to make a new closed source. Also TG has a potential legal loop hole in Mame, the can Go back to their closed source version and develop off of it and NOT worry about the Open license at all.

NOT entirely true!

The license that is easiest to use does that. However it is possible to negotiate a new license with the major developers of Mame to make a new closed source.

A thought that hadn't occurred to me, true. ;) One that did, however, would be to publish source code under the public licence which loads a closed-source library (obviously the interface would be public). It'll still be possible to subvert the inp output even this scenario, however.

Also TG has a potential legal loop hole in Mame, the can Go back to their closed source version and develop off of it and NOT worry about the Open license at all.

Yeah, but who wants to return to version 0.35. ;)

I'm agree with Kernzy, a lot of new games a variations in lastest wolfmame. I personally think that the staff should tryit a few month and see what happends.

I ve heard that in 106, its difficult to check the Dip Switches option, but in 155 its easy.

Doesn't the mame version become a moot detail with the new submission process which is based on voting on credibility of a scoring session?

like the old Pepsi vs Coke tests- take a player who can score several hundred thousand points on Donkey Kong.
Record the person playing new mame and record them playing old mame.
Side by side which is different?
Let's say they get 325k on each session, does playing on one version label them as an inferior player to their session in the other version?

Using that as a thought experiment, I'm having trouble understanding the pedantic nature of a technical detail.

Let's say you alter the CPU processing speed of a particular game platform in mame, it'll be quickly evident in the video evidence submitted. and face it, no one is going to repeatedly vote positive for a person's score if they claim to be "best" at a bunch of titles and they only show a screenshot over and over. somewhere in that mix of timeframe a video or other observers are going to be involved in watching their ability. the truth always susses out, it's just whether or not the crowd wishes to embrace the facts, in the old emotional TGIS mentality many "facts" were NOT embraced.
I see Jace's new method as fairly inspiring from a community interaction mentality.

and I'm still not seeing where mame version has much of anything to do with the community factor.
maybe a particular brand of MAME wished to pay for a TG endorsement of their version/product, that would skew the impartiality of the TG scoreboard wouldn't it? So what would be the motivation in this new plan to paint yourself intentionally into a limited corner?

I ve heard that in 106, its difficult to check the Dip Switches option, but in 155 its easy.

0.106 needed a hex editor & source code. Haven't tried with more recent versions. Many "necessary" DIP switches can be understood from playback anyway: number of lives and when they are awarded, for example.

Doesn't the mame version become a moot detail with the new submission process which is based on voting on credibility of a scoring session?

[...]it'll be quickly evident in the video evidence submitted.

You seem to be misunderstanding just how easy it is to cheat with a non-WolfMAME version of MAME. Emulation brings with it a whole bunch of other factors that are not present in the arcade. Any video footage of MAME is not evidence. The INP recorded by MAME holds various details about the gameplay including how fast the emulation was at the time of recording. You could record the gameplay on one (slow) machine or with tools to slow your machine and then playback the footage on a faster machine (or sans tools) and the video evidence would suggest you played at full speed. Also the issue of autofire can be easily seen in the playback of the INP but not in the video footage. Other technical details such as unnatural pausing (hibernate machine) can also be seen in the technical detail of the (modern) INP but could be cut out of video evidence.

0.106 needed a hex editor & source code. Haven't tried with more recent versions. Many "necessary" DIP switches can be understood from playback anyway: number of lives and when they are awarded, for example.



You seem to be misunderstanding just how easy it is to cheat with a non-WolfMAME version of MAME. Emulation brings with it a whole bunch of other factors that are not present in the arcade. Any video footage of MAME is not evidence. The INP recorded by MAME holds various details about the gameplay including how fast the emulation was at the time of recording. You could record the gameplay on one (slow) machine or with tools to slow your machine and then playback the footage on a faster machine (or sans tools) and the video evidence would suggest you played at full speed. Also the issue of autofire can be easily seen in the playback of the INP but not in the video footage. Other technical details such as unnatural pausing (hibernate machine) can also be seen in the technical detail of the (modern) INP but could be cut out of video evidence.

That really sounds forensic with a microscopically focused intent on mischief. Could it instead by the pitfall of "Analysis Paralysis" which can cripple the intentions of a project? If someone truly was going to go to that level to cheat the system, wouldn't they be so engrossed in the mindset that they could "cheat" with much less effort expended?

And let's say someone does this photo-editing method to become the greatest DK player on the planet, what is the end result? They sure aren't getting million dollar endorsements, the DK player community won't herald them a master until they've placed well at a Kong Off....yada yada.

There has to be some psychology of the payout for cheating which drives cyclists to dope, nascar to modify engines/suspensions, fighters to use steroids, etc. what is the payout for a video gamer, getting their name in a printed 1982 magazine, can't be that because we are in a digital media age where everyone can post their own website.....

When Walter describes the top 3 of why the old TGIS scoreboard has some fake scores, one of those on his list is simply- the submitters saw the scoreboard as a joke so the fake submission was the extension of said joke. In the mindset of community run environment, eventually that much "negative energy" or "joker energy" is going to suss out.

0.106 needed a hex editor & source code. Haven't tried with more recent versions. Many "necessary" DIP switches can be understood from playback anyway: number of lives and when they are awarded, for example.

Dip Switches are very easy to check in WolfMAME from 0.149 and later. While the input viewer (-iv ) is active, you press Delete. Recorded dip switch settings are displayed during playback. Default dip switch settings are also shown in braces.

http://forums.marpirc.net/viewtopic....=15224&p=58045

-George

That really sounds forensic with a microscopically focused intent on mischief. Could it instead by the pitfall of "Analysis Paralysis" which can cripple the intentions of a project? If someone truly was going to go to that level to cheat the system, wouldn't they be so engrossed in the mindset that they could "cheat" with much less effort expended?

And let's say someone [...]

No requirement to "let's say" - any MAME referees and MARP editor knows lots of players cheat. No question: it happens even with all the stringent methods of old TG & the tools available at MARP. Still is happening at MARP. When not dissuaded (caught and persuaded not to cheat), the methods get around and the cheating happens more - which in turn cheapens the plays that make the scoreboard. One incredibly easy cheat that has been used several times on TG's own scoreboard is downloading a[ny] MARP submission and uploading it to TG as their own.

For me, the question is who will bother to check the INPs. Cheating methods are easily understood and some are well known (auto fire, Magic Speed, hibernate for pause). Still others are probably not heard of publicly and still filtering through.

The answer to this question is, whatever version you want. You will then have to lobby to get through verification, but I am definitely going to be submitting with the latest version available.

Amusing to see Barthax weighing in. Next we'll have Flewin here saying we should use such-and-such version. We waited years for permission to use a new, stable version, and the answer was always 106.

I think it would be worth insisting on Wolfmame, but that would have to come from the top, and all I'm hearing is, "Let the community decide". Ha, we're not going to reach consensus in the verification threads. It's a free-for-all. Good luck to everyone.

Any non wolfmame version of mame should not be accepted. I think the community for the most part agrees on that.

Any non wolfmame version of mame should not be accepted. I think the community for the most part agrees on that.

I only see Wolfmame posted as source. Are windows and linux binaries available? I doubt many verifiers are going to be willing to compile their own.

Maybe it will be videos then, with the hands visible to confirm it's not a playback. Just do something reasonable. Verifiers will be hesitant to be on the wrong side of the adjudication, but they'll want to vote to earn more submissions. We'll see some passes with new versions.

I only see Wolfmame posted as source. Are windows and linux binaries available? I doubt many verifiers are going to be willing to compile their own.

Windows binaries for all versions of WolfMAME are posted at wolfmame.marpirc.net. Linux binaries for some versions are also posted. INPs created using a Linux binary should generally playback ok under Windows.

-George

Just curious...is there documented evidence, either from TG or MARP, that one or more gamers over the years attempted to submit a manipulated INP file or an INP file based on a manipulated ROMset ? And if so, how were these instances caught ?

yes there is.
on MARP, one guy was banned because he managed to slow down his gameplay, and then get his inp to play back at real time. the mods did not know how he done it.

Wow, did not know !! I encountered a variety of cheating with respect to vintage and modern (thru 2006) console performances, but never anything on MAME.

The only thing we were wondering about was what would happen if years down the line PC's got blazing fast...we wondered if it might be in a player's best interests to keep an old Pentium III or some other system minimally capable of being used for MAME submissions, would that give a player an advantage in some way. I never really knew the answer to that issue as I was not in the technical arm of TG.

Let me throw this scenario at you...seeing how you need the proper ROMset to play back the performance, does that make it automatically impossible to submit something via a doctored ROM ? I would imagine no because the resultant INP "should" require the commensurate ROM to handle any nuances encountered in a performance which is the result of a doctored ROm, but I could easily be mistaken. For example, suppose someone were to reduce the hit-box of certain enemy sprites just a little, enough to give a competitive edge but such that it is impossible to spot visually ? Or increase their jump height/distance nominally ?

Or (even tougher) change the percentage chances of encountering high-point random fruits in "Ms Pacman" ? How would you know since the INP playback SHOULD be the same ?

Those were some of the concerns I had back when I was the referee for MAME and thankfully no one blew one past me attempting something like that...at least as far as I know !!

Or (even tougher) change the percentage chances of encountering high-point random fruits in "Ms Pacman" ? How would you know since the INP playback SHOULD be the same ?

I'm far from a technical expert when it comes to MAME, but this question got me thinking. It's reminiscent of something that Tim Sczerby has long complained about when it comes to Donkey Kong: "Who's to say some nerdlingers with their pocket protectors aren't giving Billy and Steve juiced ROMs so they can score higher?" I'm not comparing you to Tim, but the idea of "juiced ROMs" is sort of a running joke in the DK community, so I thought I'd do a test.

I took the DK ROMs and made it so every barrel smash yields the maximum amount of points (800). This was just a quick and dirty test, so anybody who knows DK would know something is wrong right away when a brown barrel gives up 800 points when smashed. I'm sure an experienced programmer would be able to alter the scoring in subtle ways. However, this is more about INP playback and that's really what I was testing.

The results were that a DK INP recorded with an altered ROM set will eventually desync when played back with the proper ROM set. The "juiced" INP of 800 smashes desynced rather quickly. There are other Donkey Kong hacks out there ("No Killscreen", "Wild Barrel") that desync the same way when those INPs are played with the original ROMs. Again, I'm not an expert, but my guess is that the random number generator in the game is affected by every little thing that happens, including scoring. Once the games take different branches, the outcomes will never be the same. This doesn't answer the question of Ms. Pac fruits, but I imagine the results would be similar.

So, all of this is to say that visually inspecting a bogus MAME performance may or may not result in questions, depending on the skill and subtlety that goes into the deception...but an INP created on an altered ROM set will almost certainly cause issues when other people attempt to play it back.

Rob, maybe a doctored rom CAN be played back on the standard rom?
just like a recording made on widows MAME can be viewed via wolfmame?
as long as mame recognizes it is the correct rom, it may allow the doctored rom to work, as long as it is not too doctored?
as for the other thing I mentioned, I thought the cheater could slow gameplay by having many heavy programs running while recording, that's all I could think of.
Apart from this one instance, I have never heard of anyone else cheating on wolfmame.
But, after watching the Q and A's with Jace yesterday, he reckons he can mess with an INP...

After reviewing this entire thread we wanted to offer some thoughts and some guidance to the matter or MAME submissions for the time being.

First and foremost:

For the near immediate future, Twin Galaxies is just trying to get everything back up and running as it was before the scoreboard went down. We have not added any new tracks, we have not changed any posted TGTS rules for any particular game or platform.

Under this circumstance, it would be logical to assume that the MAME submission rules that stood before, in terms of what versions of MAME players should use and etc., are still the recommended guidelines.

Please understand that the current scores in the TG MAME database were accepted according to those rules/standards and it would seem to be a logical approach for everyone to stick to those rules/standards if they would like to legitimately compete alongside those scores. Generally, you want things to be as "apples-to-apples" as possible if you really want to know if you are the better player.

With that being said:

We feel that MAME (and emulation in general) is important to the future of score chasing on legacy titles.

As time marches on, it becomes increasingly more difficult (and expensive) for the general population to find the original platforms of titles that have not been in active production for many years. While certainly it is fantastic to compete on original platforms whenever possible, the reality is that wide accessibility to these game titles and platforms is necessary to support growth and to increase player participation in joining the art of score competition.

The only solution to this of course, is emulation.

Score tracking with today's current forms of emulation is filled with challenges because the fundamental computing platform that is performing the emulation is not standardized. People use different devices to run emulators on (and even different versions of the emulators themselves) and that undesirable variation adds an unknown "X" factor when trying to compare score performances against each other.

We have given much thought to this general problem and decided a while ago to move forward with a solution that will help alleviate some (but not all) issues revolving around emulation score tracking.

We have been creating a custom TG COMPETITION EMU CLIENT. We are starting with MAME as the first emulation "module" that it will support and step by step add more emus from there.

We can not go into full details here and will discuss further at a later date however here are some highlights:

1.) Built-in direct world record form submission / access to your VERIFIED TG account
2.) Anti-cheating, secure verification protocols / loaded ROM version check against verified TG database ROM set.
3.) Auto INP and screen capture recording (video) for submissions.
4.) Ability to identify and register secure controller / input devices (allows 'official controllers" to be set if desired)
5.) CPU power profiler/rating mechanism (similar to 3DMark) so that all submissions will include a profile index number that players can reference.
6.) A "high end" version for powerful computers.
7.) A "low end" version for older computers.
8.) Versions for PC/Mac/Linux
9.) Etc.

Obviously this is still being worked on but we feel that this will help make score chasing more accessible to larger groups of people and increase competition.

We also feel that it will help provide an additional level of standardization to score competition on emulated platforms.

Here is how it will be handled in the TG score database:

All current MAME scores/tracks will REMAIN. People will still be able to submit to those tracks according to the previous rules and complete on those tracks. There will be no changes and it will continue as is.

The new TGCEC will be considered its own platform and will be added to the database with its own tracks. There will be the "high end" version of the platform with tracks for submissions from people with computers that have been CPU index rated a certain value and above - and there will be a duplicate set of tracks for the "low end" version of the platform for people with computers that have been CPU index rated underneath a certain value. (When the term "CPU" is used here, it does not just mean a computer's main microprocessor, it means total system performance.) We may change this and just group everything together with CPU index ratings as an indicator, but we are not sure yet.

We are aware of the complexities of MAME licensing (and EMU licensing in general) and we have been in contact with the MAME organization. Due to the way that we are engineering the application, are able to fully comply with the MAME licensing without having to release the parts of our source code that must remain secure for all of this to work. We are excited and think that this will create great opportunities for players.

So, until the TGCEC is released, we recommend that the community adjudicates MAME accomplishments according to the previous established rules, and behaves reasonably in its judgements. The whole point is to have fun, and not turn the process into an excessive exercise of scrutiny. It is important that people just use common sense and keep the spirit of what we are all trying to do - which is have fun and appreciate each other!

More about this later, but in the meanwhile we hope this helps.

-TG