A modest proposal

greenspun.com : LUSENET : MAME Action Replay : One Thread

Ok, here's the deal.

The fundamental problem with MAME inp files are rerecordings, and play on a slow machine.

If MAME was changed with a good encryption scheme which also included frame rate information, cheating would be almost impossible. It really shouldn't be too difficult to add this to TG MAME.

It is all for naught if none of the players use this new version.

So I am proposing setting up an anonymous ftp site with the tournament version. Anyone who wants to can download this new version for personal/tournament use only. This is allowed under the MAME license aggreement.

To put it bluntly, if you are going to have a tournament, why not making cheating as DIFFICULT AS POSSIBLE?

The new version will have the following features:

1) No rerecording ability 2) No pause 3) New and tougher encryption 4) Time stamp for the frame rate 5) No distribution of source

Anyone who wants to can download the new version anonymously. This is probably the last time I'm going to open this particular can of worms. It's crystal clear to me that some players are using some sort of slowdown scheme for T2. This is a viable way to correct that problem. If there is no interest in this from the players then I'm going to stop posting about this on MARP.

Please let me know what you all think.


-- Mark Longridge (cubeman@iname.com), November 12, 1999


I'll second Mark's motion. Maybe instead of pausing - we could count the number of pauses and how long each pause lasted - but I'll leave that up to Mark to decide. Regardless - I will try to insert whatever version it is into the tournament for January.

-- Gameboy9 (goldengameboy@yahoo.com), November 12, 1999.

While I love to cover against illegial recordings as much as the next person, here is some notes. The slowdown is done by a rather little program known as mo-slo. It's intentions were honest enough (to allow play of old MS-DOS games that with the new machines ran way too fast to play), but has become the cheaters weapon of choice. Mark, you program would suffer the same fate in that regard. Gameboy, if we require it for the tournament, then we are back to the problem Nicola mentioned. Anything for public use, (aka tourney) would, according to MAMEDev, require distribution of source code. (I would simply make it next to impossible to find it, though!)

-- Chris Parsley (cparsley1@hotmail.com), November 12, 1999.

This is the best we can do (at the moment) to prevent slowdowns, i'm all for it. But, mark, I strongly suggest to advertise for some hackers to help us. Make sure you give it to some smart but honest hackers and if they can't hack a cheated inp out of it then force everyone use it for the next (or current) tourney, however you will also need to learn how to compile win32 tgmame to allow the "butt kicking" users to play.

-- Chad (churritz@cts.com), November 12, 1999.

As a member of the 'Howling Fraternity' whose motto is 'standardise' - I'm all for it. (1)-(4) should be no problem. As Chris says there may be an issue with (5).

In reference to the source distribution. I've mentioned this before but noone made any comment. Would it be acceptable to MAMEDEV licence to 'delay' distribution of the source until after the tournament. In this way we aren't denying anyone the source but are only making them wait a couple of months for it. 'We are commenting the code to make it fit for distribution' is always a good excuse :-).

I had another read of the MAME licence agreement at http://mame.retrogames.com/. The below extraction is the relevant part:

Source Code Distribution

If you distribute the binary, you should also distribute the source code. If you can't do that, you must provide a pointer to a place where the source can be obtained.

How about providing a pointer to a currently nonexistent ftp site that we are setting up. The setup for this ftp site will be 3 months just after the tournament ceases (surprise).

Another way around it may be to issue the source code 'on request to a particular nonrobotic email address', then only check that email address and respond once every 3 months. I even have the cover email in mind

"Apologies for the late response to your request for source. I've been particularly busy lately. Here's the relevant (out of date) source you were after 3 months ago. Enjoy!"

Now the only problem is that some legitimate competitors with particularly slow machines are going to be excluded from the competitions (can't run games fast enough). This is regrettable but there's nothing else we can do. The words 'upgrade' seem to leap to mind for those people.


-- Tim Morrow (tjmorrow@bigpond.com), November 13, 1999.

Mark, I think you hit all of the major shortcomings of MAME. However, I see #4 as two separate problems. The start/end time stamp for a recording may be correct, but if someone is playing with a high frame skip rate, that's a different animal. Looking at a time stamp alone wouldn't catch that. Also, there may be utilities out there that can pause a MAME thread, where someone doesn't have to use the PAUSE key. The time stamp code would have to look at the bios clock or could count the CPU cycles since the initial INP file creation time.

-- Pat (laffaye@ibm.net), November 13, 1999.

As a co-founder of the 'Howling Fraternity', I fully agree with Mark.
Anything that could prevent, or at least make very difficult to cheat, is most welcome.
Issues #1,#3 and #4 are ok. #2 is ok too, but I have to agree with someone who said the pause key , at home with a wife, kids, and so one, sometimes is necessary. So, even I always asked to disallow pausing, I think that pausing (but blanking the screen, so not to give the advantage to 'stop-and-study- the-situation') could be allowed....it's up to you, Mark.
As for #5, I already proposed to consider TG Mame as a private release, only distributed in the Marp community, so not to be obliged to release sources.


-- Cicca (cicca@writeme.com), November 15, 1999.

Looking at the timestamp code would catch people playing at a high frame rate, since the differences between timestamps would be smaller than expected. I think most games, if played at a higher frame rate than stated, are going to be more difficult to play in such a manner. The only exceptions being if you could slow down the hard sections and then speed up the easy/wait sections, this would be difficult to coordinate however - to ensure you get a timestamp difference at or under expectations. Perhaps the F10 key should also be disabled to prevent people from modifying the duration of play? That would suck since you can't speed up confirmation playbacks... Although i think the playback functionality should be taken out of the distributed TGMAME distribution all together and only mark or an authority would be able to confirm recordings.

-- Chad (churritz@cts.com), November 15, 1999.

I held another tournament at http://get.to/hcc and I suffer the same problems as you. I think the good way is to use on-line gaming. It means that the game is played on-line and data are sent to the server at runtime. The server then knows if the speed is correct etc. This also requires Mame without sources to avoid procedures like making .inp file at lower speed, and then playing it back at normal speed and recording it to another file, etc. BIOS time stamps are obvious, because anything else is very easy to hack. Also, tournament Mame can be only plugin for Mame32. Then you can distribute sources of Mame32, and your DLL will do the work. This way you will get rid of stupid license of Mame. Of course it requires a BIG portion of security and encryption. I am also very interested in making only limited number of attempts per game. This can be also checked easily when performing on-line gaming only.

-- Aley Keprt (AleyKeprt@bigfoot.com), April 19, 2001.

Sorry dude, online gaming can be faked as well, you could make an inp generator that passed what ever security questions you asked the player to make, then it would send the already cheated recorded inp (or input commands) downstream in real time to your online server to play the game and get what ever high score necesary. Plus for the honest players if you get network congestion, your game could be ruined. Online security has potential, with several authentication techniques, but it lacks robust playability.

The real key in security is to check if the recording was made at the time the game was played and at the speed at which the game was played. The only problem with this technique is encrypting the inp and timestamp/framerates so the "fakement" can't generate the inp with cheating means. Something that will be pretty secure in Ben Jos's new mame, MAMETE.

-- Chad (churritz@cts.com), April 20, 2001.

Moderation questions? read the FAQ