About Phil - What do we do now?

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

The following broadcast from David M. Waters "Aquatarkus", is brought to you by Seldon's personal shielding...

That Phil Lamat is not recording with an acceptable framerate is no longer a theory.

In case anyone is a bit awkward with the term "framerate", it is a measurement of the number of moments, or "frames", that take place in a game during a fixed amount of real time. MAME measures this in frames per second. Most, but not all games, have an expected speed of 60 frames, or 60 moments, per second. The nature of computer hardware makes it difficult to achieve 100% speed at all times. In my experience having the speed at 100% most of the time with short "bumps" down to 95% is the best that can be expected.

It is a publicly admitted fact in the MARP forum that Phil is using a 486 class system to record games. It is unlikely that the majority of his score submissions are above 80% framerate. Many of them show signs of being below 60%. What this means is more real time to think about and act on each problem in a game.

It doesn't matter whether he's doing it intentionally or not, what matters is there's no longer any significant doubt. To the best of my knowledge, this is the first time anyone is known to have cheated on more than a few games. So, MARP has changed. It will change further even if no one any further direct action. It won't stay the same as it was before even if everyone quit talking.

I see three obvious courses:

The first is to treat score entries as slightly more than mere claims. The belief would be that each score is essentially unproven anyway. From this the reasoning is that where there is no proof, there can be no cheating. The action taken against Phil will be none.

The second is to treat score entries as fact unless there is a clear reason to disbelieve the entry itself. The belief would be that irregularities will show themselves in the recording. Therefore game recordings would be routinely examined, and those that look "wrong" would be deleted. Quite a few of Phil's would disappear if this happened, but others would stay in place.

The third is to treat score entries as fact unless there is a clear reason to disbelieve the entrant. The idea would be that if someone cheats more than a few times that they've probably cheat regularly. Phil's scores would be deleted - ALL of them - and he would be barred from future score entries.

It is quite possible that a different course of action will be taken by the current tournament than by the larger MARP scoreboard.

What do I think should happen:

You just think "GrouseGod" is hostile, Mr. Pixel Perfect.

If I was running either of them Phil, the entrance screen would have a picture of me waving at your pike mounted head as a warning to future players. I would have deleted you SIX MONTHS AGO. That I'd settle for making up a cartoon in photoshop rather than saving the money up for a trip to France to come get a photo has only to do with my solid sense of right and wrong - qualities that you yourself lack. To put it in my native Texan, Mister, yer lower than a rattlesnake's belly and this town jes ain't big enuff fer da both of us.

Hopefully this will be clear enough to survive translation to Conehead...

-- Aquatarkus (aquatarkus@digicron.com), December 04, 1999

Answers

Unfortunately, the MAME developers don't care about optimisation at all. Like I've said lots of times before, I think it's very sad that a game that ran fine on a, say, 10 MHz 68000 processor is not emulated at full speed on my 225 MHz Pentium MMX. Emulators like HiVe show that the emulation speed can be an awful lot better than it is in MAME.

As a result, it's sad that people who don't have the means or intention to get a faster computer can't compete with others, but the fact remains that MAME is meant to emulate every game as close to "the real thing" as possible, and that, beyond any discussion, includes the speed at which the game runs. There are lots of games that won't run at full speed on my computer without reverting to using no sound or playing with frameskip. As a rule of thumb, I nowadays won't even TRY to run a game on MAME released after 1986 or thereabouts.

So, it is my belief that since MAME tries to be as close to the actual arcade machine as possible, anything that differs significantly from that should not be allowed. I think that people should have the decency to not upload scores to MARP for which they know they had an unfair advantage because it ran a lot slower on their computer. Unfortunately, for some people, this does not seem to be the case. Just another example: I can't run Galaga '88 on anywhere NEAR true framerate on my computer, so I find it very hard to believe that someone using a Pentium 166 can put up the top score on MARP and claim it ran at true framerate on their computer. After all, true framerate on slow computers means you have to use no sound and/or frameskip, either of which makes most games at least somewhat harder.

Again, I truly believe that if you know you have an unfair advantage because the game runs too slow on your computer (and it's extremely easy to see that by hitting F11), you should NOT upload your score to MARP or TG or anywhere else where you are competing with people who play the game as it was designed to be played. In my opinion, knowingly uploading those recordings is the same as cheating intentionally. If you want to allow it, then create a place where everything is allowed. People made tongue-in-cheek remarks about having separate Olympic Games where steroids were allowed right after (one of) the Ben Johnson incident(s). This no longer seems so ridiculous to me. I myself could put up a lot of incredible scores by slowing down my computer or using other types of cheats (re-recording, autofire, ...), but I am not because it's not fair within the set framework of rules. I could relatively easily skyrocket my way into MARP's top 10 if I cheated, because I have enough knowledge about a lot of things to do so. Cicca has already shown that by slowing down his computer, he was able to score a lot higher on Crystal Castles than he would have been able to do otherwise, and, just to experiment, I myself happened to play Jr. Pacman on 25% speed yesterday, and in my FIRST attempt, I scored more than 33% more than the best I have EVER scored in a few hundred attempts on regular speed.

I can definitely see that some of the people I suspect of cheating (and whether their computer just can't handle the game's "real" speed or whether they intentionally slowed it down doesn't make any difference in my definition of "cheating" as I've indicated) on MARP have clearly studied the games they play, but that does NOT give them the right to then play it at a low framerate and submit it.

What I should really like to see is these people sending in something in the neighbourhood (say 75%) of their scores on MARP using a version like Mark Longridge's upcoming TG3 version of MAME. I bet a lot of them couldn't. Sure, they may be rusty on some games, but it shouldn't take that long to get back to somewhere close to your original level after some practice.

I know this post probably won't have any effect. Cheaters won't be converted by it. But having been attacked once myself because basically one person doubted one of my scores (without even bothering to look at my recording), this matter has become dear to me. So I had to give my opinion.

What good does it do you anyway to have incredible scores on MARP if most of the people doubt you? And you won't win any money or prestige from it. And does it feel good to know you have the high on MARP, but you know you cheated? I don't think so...

Ben Jos.

-- Ben Jos Walbeehm (walbeehm@walbeehm.com), December 04, 1999.


Ok, I'm coming out of the woodwork now, but just to put in my $.02. I don't have any problems with any of the marpers, but my biggest concern back some time ago was one of Phil's Track n Field scores he submitted. I mentioned in on the board, and mentioned is personally to Zwaxy that it looked like it shouldn't belong because when I played it back, there were several occasions where he pinned the power bar WAY beyond the max power level of 1,500. His score also ended up beating out the 13 year old world record of 95K+ and when I asked Billy if this was possible, he said flat out NO. He witnesses the guy who set the world record of 95k+ and never seen the power bar go crazy when it went way beyond the 1,500 max point. My assumption was that some kind of cheat of controller was used to get enormous speed, this is just not possible with the human fingers nor is it possible with a joystick instead of buttons. So, would playing on a 486 have anything to do with super high scores on T&F? not sure. I just wanted other players to watch it and see for them selves, and if I was wrong about it, I'd stand corrected.

PS Reading all these threads over the past few months about cheating and ways to bend the rules or cut corners to push ahead of another player just means that when the next edition of the Book of World Records is ready to go into production, I can tell you that a hugh amount of scores will not make the book if any score is suspect to a player cheating in any way. I personally along with Mark Longridge and perhaps a few other TG judges will "hand pick" those players and their scores and that's it. Bring back the old marp because that's where most of the scores are going to be derived from anyway.

-- stephen krogman (skrogman@bellsouth.net), December 04, 1999.


I don't like the idea of removing a player because they happen to have a slow computer, yes computers are cheap but many people are poor. I do like the idea of removing recordings if they are deteremined to be not played like the origonal aracade played. But i also think keeping the "illegal" recordings around isn't a terribly bad idea if other people can't get to levels "legally" and want to see what the higher levels look like. So here's what i radically suggest.

Create a new leader board that tallies only TGMAME scores. So people who want to complain about recordings need not care if someone climbs to the top of the non TGMAME leaderboard because it means you might have cheated in doing so.

Of course you can interet this like "wait, i'm high in the leader board and i didn't cheat????" which is an uncofortable position. But, you can always start uploading with TGMAME scores and you'll be competeing on a level where everyone starts from 0. You might also think "why the hell should i redo my scores just because some people cheat and i dont?" That I don't have an answer for, but i'd rather alienate people upset about having to make future recordings than people upset about having made past recordings.

I think cleansing the leaderboard pallet with a possibly uncheatable mame might eliminate the bickering and acusations (except for autofire for now (which also may go the way of the dinasour with a detection program.)) I don't think there would be any complaining if a scarily high score made by regular mame gets uploaded, since if they really wanted to show their stuff they can upload it TGMAME as well.

-- Chad Hurwitz (churritz@cts.com), December 04, 1999.


Here we go again, round and round, howling at the full moon! Unfortunately it's seems to be bows and arrows against the lightning. It's all been said before but here I go again ...

Several people want to administer/formulate new policies on MARP. Most of the new rules people want introduced pertain to cheating (either deliberate or unintentional). However none of those proponents for new rules are in a position to implement said policy despite what appears to be overwhelming support. The main reason for this is probably because it's too much work to put into place.

Check out my board entry "Non Standard Recordings - It's Getting Irritating" for my previous harrangue on the subject.

I still believe the best approach is to introduce a standard upload template of questions (for both tournament and MARP leaderboard) that has to accompany any uploaded recording. Below is a sample template that I use with all my recordings and would be a good starting point. It includes all the favourites - autofire, slomo, pausing, rerecording etc. I hope it cut and paste's alright:

STANDARD RECORDING INFORMATION

01. Name/Player 02. Initials 03. Date recorded 04. Game 05. Romname 06. Score 07. Stage 08. Recording Time 09. Dip switch settings 10. MAME version? 11. OS Version 12. Sound

ADDITIONAL RECORDING INFORMATION

13. Settings/command line used in recording 14. Where did you get your version of MAME? 15. Specify system details CPU : Memory: Video : 16. What input hardware did you use? 17. Did you have cheats enabled? If so specify what cheats you used. 18. Did you use any slow motion software? If so please specify. 19. What framerate was reported by MAME at the end of your recording (fps)? 20. Did you use autofire? 21. How often did you pause the game in the recording? 22. Please indicate other hardware/software methods used to make the game easier? 23. Any other pertinent details surrounding the recording?

Recordings will then be rejected (deleted) if

1. Any questions aren't answered (or don't make sense) on the standard upload form. 2. If any answers turn out to be false as demonstrated by other MARP users.

Finally, only certain answers are acceptable to be elgible for leaderboard points. e.g. framerate within 95% of 60 fps etc. Defining what's acceptable and what's not will have to go to a MARP vote.

My view is if your machine can't perform within acceptable limits for a given game then find another game or upgrade. I can't image any serious competitor having any time for people who upload recordings at 60% correct framerate. That is a joke, a very sad joke.

A good way to force the issue is to adopt Mark's tournament MAME.

-- Tim Morrow (tjmorrow@bigpond.com), December 05, 1999.


Creating a new marp,only with tg scores is a great idea. All scores zeroed, then everybody have to record again using TGMame, i vote for that we use tgmame in the future, (of course if it is correct handled according to Mame license (open source,Mark told me this is no problem at all). Removing all old scores,but keep backup of them, then zero all recordings,then start all over with tgmame scores, is a great idea.

Only my 2 $

Stig

-- Stig Remnes (sremnes@samsen.com), December 05, 1999.



If my scores are cleared, I will not resubmit! I worked damn hard for my scores, and will not torture myself to make them again!

-- Chris Parsley (cparsley1@hotmail.com), December 05, 1999.

Adding a new leaderboard function wouldn't clear scores. Just another way to tally scores up. of course in the tgmame leaderboard everyone would start with a zero, but you could still look up your standing in the old leaderboard.

-- Chad Hurwitz (churritz@cts.com), December 05, 1999.

In pure "Howling Style", right besides Tim, I realized that the only thing that we could do now is: nothing !!!
I've been one of the first to raise up the slow down issue, but after all theese discussions about, I now guess no action could be takes against Phil and Mvdv, since we can not prove they slowered down their PC intentionally. I do agree with Ben Jos that posting such scores with a slow PC is cheating as well, but since it's not stated in capital letters in the rules that "A framerate below (i.e.) 50 fps is not allowed!", nothing can actually be done !!!
I'm sorry if someone with a slow PC will be forced to give up, but it's for the same reason that I'm posting SF2 scores, or I'm not complaining with anyone if my slow car does not bring me to the seaside in half time !!!

This T2 is sadly gone, in a hell of cheating, suspects, insults and so on.... rules since T3 hopefully will be clearer and more severe, TG3MAME will prevent any cheating technique, and tournaments will result in real, meaningfull leaderboards of the actual partecipants skill !!!

Cicca

-- Cicca (cicca@writeme.com), December 06, 1999.

Errata corrige: I'm NOT posting SF2 scores, of course.....my PC can't run it !!!

-- Cicca (cicca@writeme.com), December 06, 1999.

Moderation questions? read the FAQ