I'm new to this site. I've had a good look at several of the hi score threads on this message board but haven't found any reference to the use of slowmo in game recordings. Clearly it is a banned technique, but how is it enforced?

I've downloaded several hi-score recordings. After playing several of them back certain thoughts came to mind:

1. Some players are using turbo/auto fire joysticks 2. Some games may have been recorded on slow machines or have used slow motion software 3. Can games be paused midgame?

1 has been discussed in another thread so ignore it here.

As for 2, I got out an old copy of moslo.com, an innocuous 986 byte piece of software that you can use to slow down old dos games. I then used it to record starforc for a few minutes just to see what would happen when I played it back without moslo. After finishing the recording, MAME indicated that the game ran at an average of 25fps - less than half the 60fps intended speed. I then played it back. Boy did I look good! The game played back at 60fps and the ship was wickedly weaving through enemy bullets and blowing away enemies with lightning reflexes. I found this particularly disturbing. I have a fear that the framerate is not being recorded in the recording *.inp format, nor are timestamps being used. This raises questions of validity on any recorded game. This is an example where you are at an advantage playing on a old 486 :-). Can somebody please confirm whether this is the case or is there a way to determine at what refresh rate the game was being played at? And if so, is there a rule banning recording on say less than 90% of intended speed? I hope so.

As for 3, I did a recording of starforc again and paused a couple of times and then played back the recording - there was no indication that I had paused the game! I believe this is banned by the current rules. Is it possible to tell whether a recording has been paused midgame?

I'd like to start submitting some recordings but it seems to be a bit pointless if the above ploys are being used on existing recordings. I'm not sure what the MAME DEV's views are on the game recording feature (I figure it's pretty low compared to emulating new games) but they might be prepared to put some more details into game recordings if we asked nicely. I don't think it would be much programming effort since the information is already being generated by MAME (e.g. refresh rate, whether pause has pressed etc). It's incredible how easy a game becomes when you play it at half speed, or you can take a rest halfway through and come back refreshed.

Anybody have any comments / thoughts on the above.

Cheers, Tim

-- Tim Morrow (tjmorrow@bigpond.com), September 18, 1999

Answers

While I'll be the first to agree with you on Moslo (since I have never been able to get it to work), I disagree with pausing (except for games that you can use to an advantage. Examples of this being Tetris and MPE. For any other game no, people I know who have more than just to sit in front of a computer have other things to do than to sit at a game for hours on end. It is a simply, if you need to do something, put MAME on pause, and when you are done, come back and finish the job, so to speak.

-- Chris Parsley (cparsley1@hotmail.com), September 18, 1999.

There will always be people who cheat, and the more people know of ways to cheat, the higher the chance that someone cheats, and so the higher the number of submitted recordings in which the player somehow cheated. Did you consider that before you posted your question? Because while you are concerned about people cheating, you have just given detailed descriptions of how to cheat, and lots of people did not know the ways you just described...

I did not know about the slowing down trick, and I have only known about using pause for a short time. And the only reason I know about pause now is because I was actually more or less accused of using it by someone who hadn't even bothered to look at my recording.

Like I said, there will always be people who cheat, and I am not too concerned about that, because there is no way you are going to stop it. I think (but of course there is no way to be sure) that the majority of the recordings on MARP are legitimate. And the less public talk about cheating, the more likely it is that it stays that way.

My opinion on things that are possible on MAME, but not on the actual arcade machine is this: If you are recording a game that you plan to submit to some repository, everything should be as close to the situation on the actual arcade machine as possible. So if the arcade machine does not allow you to pause, then you shouldn't do it while recording on MAME. And if the arcade machine does not have a keyboard or trackball or gamepad or whatever, then you shouldn't use any of those while recording on MAME. I don't care what you do when you are not recording and planning on submitting it.

Having said all that, there is a very good argument to use whatever MAME enables you to, even if that is different from the arcade machine: It is very clear that Twin Galaxies will always make a distinction between arcade machines and emulators, so, for instance, my MAME world record on Donkey Kong will never be considered the world record for the arcade machine. In short: People used whatever the arcade machines allowed them to (and whatever was allowed by judges watching them), so why not use whatever MAME allows you to, too?

And, yes, I think you are right: The MAME Dev team probably does not care at all about concerns like the ones you raised. Their main concern is to emulate, as realistically as possible, the arcade machines, and as such, recording a game does not even fit in to start with.

If setting world records on emulators ever becomes a big thing, the judging will probably become more strict too, and merely submitting an inp may not be sufficient anymore then. A recording is nothing more than a nice way to preserve your accomplishments forever then, but it may not be considered anything more than that unless approved by some official judge of some kind then.

And there is one more thing that a lot of people seem to forget: MARP is simply a repository for people to submit their recordings to, and banning recordings has so far not been part of that policy. Whether or not to allow certain recordings to earn points for MARP's leaderboard is a different matter, and maybe the way to solve this is to not automatically assign points to every submission, but to have the uploader explicitly indicate whether he wants his submission to count for the leaderboard. So this would effectively mean that there were several categories. The same goes for autofire; there is no MARP policy to ban autofire, but there may be some decision soon on that with regards to accepting it for the leaderboard. Pending that, I have clearly indicated which of my recordings made use of autofire, and regardless of the final decision, I am planning on replacing those recordings with ones in which I don't use autofire.

But if you really want to compete and are concerned about the validity of some of the recordings of some of your competitors, then maybe you shouldn't upload to MARP but to the TG MAME site instead, because there, at least, every submission is submitted to scrutiny by a TG judge.

Just some thoughts...

Cheers, Ben Jos.

-- Ben Jos Walbeehm (walbeehm@walbeehm.com), September 18, 1999.

wow finally some long winded posts that didn't include much flaming :)

Unfortunatley Ben's right in that there will always be cheating and i think even the cheaters know that marp stands for fair play regaurdless of what actually gets uploaded, so the only thing stoping it right now is guilt. The bad thing is if there isn't a full proof method to determine if someone's hands are really that fast and staminatious (i knew many non-cheaters that were incredible with their hands!)

There is a slightly possible although severly big-brother way to theoretically prevent a majority of pausing+slomos. What would happen would be for each inp file there would be an encrypted header and tailer that would store the begin and end time of recording the game. To verify a recoring as not cheating you'd simply decrypt the time difference and divide by the frames in the inp file to get the recorded fps. However i haven't figured out a way to keep the encryption algorithm secret so savy cheaters couldn't just run the encryption algorithm with bogus times to match the game's frame rate. and even if you have an encryption algorithm hiden behind a link, anyone would be able to run it because mame has to. So that idea proly will never pan out.

cheating will always be around, maybe it could be banned but there would be a lot of subjective judging to ban a cheating inp. At least with playbackability, if it doesn't playback there's your indisputable proof that it doesn't belong on an "Action" Replay Page.

-- Chad (churritz@cts.com), September 18, 1999.

Chad,

While you are right that putting some encrypted verification inside every inp would not be fool proof, especially not since MAME's source code is public, it WOULD drastically cut down the number of undetectable cheats. Because without it, everybody could use pause and slowdown cheats, but with a time stamping of some sort in place, only people who know how it works could successfully modify it, which, at least initially (until somebody makes some program available), would mean that only programmers could successfully modify the inp to make it seem valid.

So your idea really isn't that bad. The MAME Dev team probably would not go for it, but just about any programmer could make a special version of MAME, and a site like MARP could require every uploader to use that special version. I guess the best thing would be to make that special version so that it puts the encrypted verification all the way at the end of the recording, because that way, the recording would still play back correctly on the regular version of MAME, and a simple program could be written that would verify if a given inp created by that special version of MAME is valid.

Personally, I am not too worried about cheaters. If someone cheats and beats one of my scores, fine, I have nothing important to gain or lose by having or not having a top score. It's all a game and there are no real prizes won or lost. If somebody really wants to go through the trouble of making sure that recordings are less easy to be rigged, I may even offer some help, but I really don't think it's worth it.

Ben Jos.

-- Ben Jos Walbeehm (walbeehm@walbeehm.com), September 18, 1999.

Every sporting event, even pee wee soccer (football for the UK guys) has some sort of judging.

The one fundamental thing I tried to accomplish with MAME was that everyone was competing on a level playing field. My thinking is that this is the single most important thing to do, if we wish to meaningfully compare scores. The only way to do that was to have standardized .cfg files for every game, something I am working slowly towards.

You can't stop cheating on MAME with 100% certainty.

However.... the best players in the world will video tape their games on the real thing, and the cheaters will not be able to do that. Armed with INPs and video tape, that player would fully deserve the World Champion designation. All the scores I've accomplished on MARP I've done before on the real thing, and I think there are players here who are, in fact, the best in the world.

If some elaborate cheating scheme bypasses all detection all I can say is: Ok, now play on the real thing, and send me the video tape. If the player in question refuses, the score is rescinded.

Simple.

I would think that such is so rare, that it has never happened. I'm happy with the current situation.

I saw Rick Fothergill play Ms Pacman in New Hampshire and get 901,000+ which back in 1985 I thought "no way is that humanly possible". I thought "no one will break 5 million on Galaga". Steve Krogman got 7 million at Funspot in NH, and I saw it with my own eyes.

The remote possibility of cheating does not cancel out the incredible skill of these players. I can say that the best players I've had the good fortune to see in action were 100% legit. Once again, I maintain that faking an INP file would be very VERY difficult.

My bigger concern is that everyone use the same difficulty settings. A Joust score of over 100 million is possible on the factory settings (and it was done in 67.5 hours!) but is this a true test of SKILL? No way! It's a true test of how much coffee you can drink or worse. That is why I'm always advocating the use of TG settings for Williams games, and other games. Joust with 5 men no extras is exciting, and factory settings are a colossal bore (that goes double for Gauntlet and Q*bert)

I've said my piece.

Mark

-- Mark Longridge (cubeman@iname.com), September 18, 1999.

>There will always be people who cheat, and the more people know of >ways to cheat, the higher the chance that someone cheats, and so the >higher the number of submitted recordings in which the player somehow >cheated. Did you consider that before you posted your question? >Because while you are concerned about people cheating, you have just >given detailed descriptions of how to cheat, and lots of people did >not know the ways you just described... Funny you should mention this because it was the first thing I thought of before submitting. However, after some thought I realised that "I ain't telling them nothing cheaters don't already know". Like you say most people are basically honest and play by the rules. My thoughts were mainly aimed at those who might not be aware of what might be going on so and what to look out for. I'm mainly concerned at just how easy it is to cheat with slow motion software - I hope a way can be found to prevent it.

>And there is one more thing that a lot of people seem to forget: MARP >is simply a repository for people to submit their recordings to, and >banning recordings has so far not been part of that policy. Good point, I overlooked this. I guess I'd just like to see a larger list of what information be included with each recording submitted, as well as MAME version, dip switch settings etc, specify 1. What was the final frame rate reported at the end of the recording? 2. How often did you pause? 3. Did you use autofire? 4. Any other relevant comments on the recording.

Cheers, Tim

-- Tim Morrow (tjmorrow@bigpond.com), September 19, 1999.

I'll just quickly jot down a thought or two and get back to business because I think this subject is an important one.

I am all for allowing pauses in mid-game. Were we in arcade situations, I'd be all against it. :) Yes, we should duplicate the arcade experience as much as possible. Sadly, though, some of us have lives which makes that difficult. The best example I can think of is that I own a dog (black lab). She's getting on in years and doesn't have the bowel control she used to. Now if I go playing a marathon game like Galaga, or perhaps Gyruss (if I ~could~ play a marathon game on that one, which I cannot), and she needs to go out, it's not a question of asking someone else to take care of it or telling my dog to hold it... Though that thought is kinda amusing. :) I either pause the game and let her out or the carpet gets ruined. Should I be disqualified for that? My feelings are that I shouldn't, but the community does not necessarily seem to agree.

Personally, I don't mind either way. One can always make time for the really long games, shut off the phone and such. I'm not interested in changing policy, just saying what happens to be on my mind as I read. Now as for autofire, I can't help but chuckle a little. I will take manual fire over autofire ~anyday~ when I'm playing a game. Yes, even Track & Field. And I suck horribly at that game. But I'm talking about the shooters primarily, because you need precision in addition to firing speed, and autofire doesn't offer that. I imagine some of you have played Quake 2, yes? Who gets more kills with the railgun: Someone who leaves the fire button held down, or someone who lines up each shot before firing? Autofire in that case also leaves you very vulnerable because you're always weaponless except for that one moment you get to shoot, but you're not as likely to be on target with each of those shots. Well, you get the idea. :)

Anyway, the reason I'm amused is because so many people are so, well, passionate about the autofire/manual fire debate. I see all kinds of posts about how someone is cheating (possible, depends on your view), how someone's score is less impressive because they used autofire (that one ~really~ made me laugh :). I'm not laughing to make fun, I just think that many MARPers have lost sight of the fact that this place, while it may offer competition, is not the Superbowl of video games. I can understand that rules need to be made now and then, but when I think back to that person going on about how his score was more impressive for having used manual fire, well, that's when I think egos are getting in the way of enjoying oneself here.

When I discovered this place, I was bordering on ecstatic because to me, seeing how someone got a score is everything. We see farther by standing on each other's shoulders, as it were, and if I can use someone's recording to better my own game and someday down the road, someone else uses my work to do even better, I'm all for it. One of these days we'll all kick butt on almost every game made unless we lose that sense of community first. I guess what I'm trying to say is that while competition is well and good, and I even enjoy a friendly round of trash talking before a good matchup, I don't see it as a good thing for MARP that people are making some of these situations to be a life-or-death thing. This shouldn't be stressful! Geez, go play a video game or something. :)

Taz (OK, that wasn't exactly 'a quick thought or two')

-- Brian McLean (bmclean84@hotmail.com), September 19, 1999.

Taz made me realise something I hadn't thought of before, viz. that there are two kinds of pausing, really. One is the one Taz described, where you are interrupted by something; the other is where you use the pause key to gain some kind of advantage in the game you are playing. Examples of advantages: Take a break to come back refreshed (which doesn't work for me; I lose my rhythm and concentration then and usually lose one or more lives shortly after resuming), and pausing so you can observe the situation on the entire screen and decide what to do after resuming the game. I am not sure which of the two kinds of pausing Tim was referring to, but I had the impression it was the one where the player was trying to gain some advantage by using the pause key.

Anyway, it's not that important to me. I won't use the pause key while recording, but if someone wants to, they're welcome to. And if that person uses it to gain some advantage, then I guess they're not good enough at the game to be able to do it without. And, like Mark said, they would probably not be able to perform well on a real machine.

Ben Jos.

-- Ben Jos Walbeehm (walbeehm@walbeehm.com), September 19, 1999.

As Brian says there are several reasons someone may pause the games - many of them legitimate. Unfortunately there is no way to tell for sure whey someone paused. Actually the pausing issue is not a large one to me - I not against it in a big way. Although it could easily be recorded along with other keystrokes so it is clear whether the recording was paused or not.

>Now as for autofire, I can't help but chuckle a little. I will take >manual fire over autofire ~anyday~ when I'm playing a game. Yes, even >Track & Field. And I suck horribly at that game. But I'm talking >about the shooters primarily, because you need precision in addition >to firing speed, and autofire doesn't offer that. I can't agree with this. It depends on the game. You can find games where autofire is a disadvantage because of the indiscriminate shooting but in most instances autofire is superior. I know this because I've used it in many games and almost always do better than without it. In particular in starforc autofire is diabolical. Here are 4 reasons for this in starforc:

1. In the game by default you have 3 lives with 3 from score if you reach 500000. Periodically, at certain points a 2X4 array of question marks appears and one of them randomly holds a free life. Using autofire guarantees you the free life because you almost always shoot them all out unless you are also unluckily competing with a wicked pattern of flying enemies at the same time. If you are using manual fire, even when no enemies are around it is a challenge to knock over more than 5 of them. Usually you can get about 4 with a probability of a free life of about 1/2. Good players will have at least 4 opportunities to do this so in my estimation the autofire players on average will get 2 more lives then the manual player.

2. Periodically a large block appears. When it turns white, if you can shoot it very, very quickly you will get a 50000 bonus. I've never done it with manual and it's trivial with autofire. This occurs in at least 4 places --> 200000 points.

3. In some places if you shoot 16 small blocks in a line you get a bonus of 80000. Again, very hard to do with manual shoot and pathetically easy with autofire. You can do this at least three times --> 240000 points.

4. The patterns of enemies is fairly simple and not random. By firing quickly you can clear most of them leaving fewer enemies with fewer bullets to avoid at any time.

Now a good players score using manual fire will be between 1.0-1.5 million (ignoring a 1 million bonus that you can get late in the game if you are lucky). Now 440000 pts + 2 extra lives + less bullets is a significant improvement to that score.

>Anyway, the reason I'm amused is because so many people are so, well, >passionate about the autofire/manual fire debate. I see all kinds of >posts about how someone is cheating (possible, depends on your view), >how someone's score is less impressive because they used autofire >(that one ~really~ made me laugh :). If a person is advantaged from the use of autofire it DOES make their recording less impressive than if they had played manual. My line is that I want to know whether they used autofire or not in the first instance and in the second I'd like their effort categorised separately. I have no objection to people using autofire, only to their efforts being compared against someone who didn't. Autofire is a great way to play, more relaxing (your hands don't tire out) and you get better scores.

>but when I think back to that person going on about how his score was >more impressive for having used manual fire, well, that's when I >think egos are getting in the way of enjoying oneself here. Ego has nothing to do with it. I enjoy and am equally impressed with both autofire and manual games (not my own). It depends on how skillfully they played that game. Needless to say if the player stayed in one corner leaning on the autofire button for 2 hours and got a high score, I wouldn't keep the recording. I also maintain that in starforc there are many players who can get to stage 20 but damn few who can make it to stage 24 - those 4 levels are ball breakers. Any person who gets to stage 24 with whatever score has more skill than someone who can only get to stage 20 with autofire. The only thing taking away my enjoyment here is the apparant failure admit that there's an advantage of using autofire over manual shoot in starforc in particular and most games in general.

In finishing I'd like to clarify a few things that I perhaps haven't made clear earlier. I love video games with a passion in all its forms and guises. I also believe that people should be allowed to enjoy them in whatever way that gives them the biggest buzz. Now that may be using autotfire, pausing the game at times, playing the game in slow motion, playing half a game one night and continuing the next, using the joystick with their teeth (my neighbours dog plays it this way), whatever.

I also have a strong sense of honesty. I feel that if you've used any particular method in your recording then you should say so and most people probably do. But I also know that there is a minority that will either unknowingly or deliberately omit from mentioning pertinent information surrounding their recordings. Now I had hoped that the playback binary would store all of the relevant information that at least I would want to know on framerates etc. Unfortunately I found that this is not the case. My first question was whether this could be easily rectified - and it looks like the answer is that it can, but not easily. In the absence of that we have to rely on the honour system. In my own case if I submit any recordings, they will be accompanied with the following information

01. Game (romname)? 02. Date recording? 03. Dip switch settings? 04. MAME version? 05. Where did you get that version of MAME? Downloaded from Got it off a friend Built it from the source code Other - please specify 06. Specify system details CPU Memory Video Card 07. What input hardware did you use? Keyboard Mouse Standard 2,4,6,8 button joystick? NES Advantage etc 08. Did you have cheats enabled? If so specify what cheats you used. 09. Did you use any slow motion software? If so please specify. 10. What framerate was reported by MAME at the end of your recording (fps)? 11. Did you use autofire? 12. How often did you pause the game in the recording? 13. Did you use any hardware/software methods to make the game easier? If so, please specify? 14. Any other pertinent details surrounding the recording? >to be a life-or-death thing. This shouldn't be stressful! Geez, go >play a video game or something. :) Great idea! Let's see, hmm!, how about starforc WITH autofire - bwahaaahaaahaaah!

Game On, Tim

-- Tim Morrow (tjmorrow@bigpond.com), September 20, 1999.