Billy Mitchell's Direct Feed Recordings are MAME Generated

[YesAffinity]

^Great, scores stricken, Billy keeps his stature because he "didn't know".  Either way it seems the scores are coming down from TG.

I really can't imagine carrying all that weight of the wrong-doing around for so many years, and then having it surface and piece-by-piece getting fully exposed.  From the point of the initial frame-by-frame comparisons, to the irrefutable proof that the board swap was faked, to the most recent revelation that not only was MAME used but also likely tweaked to run at arcade monitor frequency. 

[Firebrandx]

We've seen what the transition frames in MAME look like if we adjust MAME's timing parameters (adjusting the CPU speed or the refresh rate) in the TG thread and it showed that the transition frames still do not look like frames from DK hardware.

I was curious to see what it might look like if we were to change timing parameters on DK hardware to see if that could possibly make DK frames look like MAME frames.   I reprogrammed DK's own drawing code to mathematically model the way DK hardware renders video and then made the program run through a range of timing adjustments.

Here are the results.  Keep in mind that this is simply a mathematical simulation and I used MAME to do it.  It was only meant to satisfy my own curiosity.  It does reproduce results very similar to that of actual DK hardware and I think it makes a good visual aid to demonstrate how DK hardware rendering is different from the way older versions of MAME render video.   Even if a DK PCB were hacked to run very different timing it still doesn't render frames that look like MAME.  It can't for the same reason that older versions of MAME won't render frames that look like DK hardware even if you adjust MAME's timing parameters.  They both have very different methods of composing the video and give different results

Nice work! So even if you HACK the arcade hardware, you STILL cannot get it to mimic the MAME artifacts in Billy's video tapes. Lance Armstrong... er, I mean Billy Mitchell is going down. Maybe he should book an appearance on Oprah and do a full confession. Might even make some cash out of it. :-P

[hooch66]

All the evidence points to it being MAME.

But even beyond that, if he really truly cared and had the skills he could do a score (above what his disputed "tapes" show) live on a neutral machine. Isn't he the one that said only live scores matter?

[Firebrandx]

I don't share the sentiment that Billy should somehow be forgiven if he does a legit million point run live. That doeasn't excuse the disgusting lengths he went to in order to not only deceive the public, but actively troll his fellow competitors. Case in point is the 1.047 million run where he intentionally kills off his men at the end. Now that we know the tape is MAME gameplay, it makes perfect sense why he would pull such an absolutely insane stunt: Banking scores on tape using MAME completely takes the element of fear out of the picture. Had he been playing on legit hardware live, there's no way in hell he kills off that world record there.

At any rate, Jace posted the MP4 rips in the dispute thread. I spent 3 hours screen-capping every girder transition that appears in his 1.047 million recording. Of the 58 girder transitions (some were killing off his men and one attract mode at the end)...  31 out of 58 screens showed the girder finger. That's over 50%! Now when you factor in the recording is 30 FPS from a 60.6 output with tape wobble, it perfectly falls into the law of overages that about half the time the girder finger frame is skipped in the video.

My zip package of all 58 screen caps can be downloaded from the dispute thread:

https://www.twingalaxies.com/showthread.php/176004-Dispute-Jeremy-Young-Arcade-Donkey-Kong-Points-Hammer-Allowed-Player-Billy-L-Mitchell-Score-1-062-800?p=959163&viewfull=1#post959163

[smf]

Has anyone else noticed that the direct recording of a pcb on:

https://donkeykongforum.net/index.php?topic=1315.msg24208#msg24208
https://donkeykongforum.net/index.php?topic=1413.msg25141#msg25141

Has the score on the right of the screen, but the tapes on the twin galaxies thread have the image rotated 180 degrees and so the score is on the left?

MAME -norotate is supposed to match a direct recording of a pcb, but it has always displayed it like the video tape.

Is there an explanation for the 180 degree flip from the direct recording, or is MAME (and the videos uploaded to twin galaxies) just wrong?

[SHLONKY]

Has anyone else noticed that the direct recording of a pcb on:

https://donkeykongforum.net/index.php?topic=1315.msg24208#msg24208
https://donkeykongforum.net/index.php?topic=1413.msg25141#msg25141

Has the score on the right of the screen, but the tapes on the twin galaxies thread have the image rotated 180 degrees and so the score is on the left?

MAME -norotate is supposed to match a direct recording of a pcb, but it has always displayed it like the video tape.

Is there an explanation for the 180 degree flip from the direct recording, or is MAME (and the videos uploaded to twin galaxies) just wrong?

the explanation is whoever recorded the first 2 tapes screwed up and had the image rotated the wrong way! Or as you say, the way mame shows it! ...theyd noticed by tape 3 and rotated it the right way.

[smf]

the explanation is whoever recorded the first 2 tapes screwed up and had the image rotated the wrong way! Or as you say, the way mame shows it! ...theyd noticed by tape 3 and rotated it the right way.

If that is the case, then I'd love to hear their explanation for how it accidentally got rotated 180 degrees.

Getting it wrong in MAME is easy, getting it wrong when directly recording from the PCB is hard.

[YesAffinity]

This is one of the points of contention, which I believe Jeremy addressed in the original POST, but yes, folks are aware and I don't believe camp Billy has in any way addressed this.  They are focused on proving that the girder finger is just a figment of everyone's imagination.

[bensweeneyonbass]

 

[YesAffinity]

“Today, he is co-founder of the Kong Off Donkey World Champoionships”

Uhhh….what?

[Sqrlmonger]

Thanks for the feedback, and I do agree just a single outlier is not even close to sufficient to make any sort of claims re: Billy's scores.  Having said that, if the notion is that MAME was used to construct some of these scores, I do wonder how the random nature of a legitimate game would not be altered in some way. Though again, it may be very difficult to find how it is altered in a way that is statistically significant.

More specifically, the approach I had envisioned was to avoid the need to quantify individual decisions by focusing on the degree to which some board scores were outliers.  Analyzing these outliers in a vacuum is certainly doomed to failure for lack of context, but in comparison to the rate of outliers of peer games of the era we should be able to address that need for context. 

With all of that said, I do lack domain knowledge specific to DK so I may not be aware of specific challenges that you have already encountered in your previous efforts (like those from your link).

edit: I do want to add the specific concern you raise about WR games being outliers by definition, is, I believe, addressed to a large degree by focusing on 1m+ point games as so-called "peer games" all of which, I would think, could be considered outliers.  I concede the pool may need more refinement, but I do think that refinement is possible.

There is one way to do a statistical analysis which would help identify MAME abuse scores -

The way the 300/500/800 point system breaks down means that a player gets an average 525 points for a blue barrel or pie, and an average 450 points for a fireball.  The scores are determined at 'random' within the game and are not at all controllable by the player.  If you counted the number of blue barrel/pie smashes and fireball smashes in a game played on an arcade machine, you should get close to the expected averages for each smash category.  For example, a typical arcade machine player might get slightly higher than average on pies but slightly lower than average on fireballs over the course of one game, but in general still close to the averages.  These averages are independent of gameplay, it doesn't matter if a player likes to smash lots of pies for example, their average score per pie should still be roughly the expected average.  If a player was abusing MAME and replaying for favourable scoring, I would expect that their smash scores would be generally higher than the expected average for each of the 3 smash categories.  If for example they consistently scored more than the average in each category, maybe over multiple fake MAME video-tape submissions, it would look like the fingerprint of a cheater.  You would have to get stats on each smash in a game though and whether or not 300, 500 or 800 was scored.  You could then calculate how many standard deviations away from the mean someone was and how likely such a feat was.

There was a chart posted on Twitter showing that Billy Mitchell achieved a high proportion of his score from hammer smashes.  I note that he had a good proportion of smashes on rivets, where you can only smash fireballs which overall score less than pies and blue barrels, yet his overall smash average (his "smaverage"?) was still pretty high.  Unfortunately that chart didn't have an exact breakdown for number of actual blue barrels / pies / fireballs smashed so that average scores could not be determined for each.  Maybe Billy Mitchell is so awesome that his method of playing means that the fireballs line up for the slaughter (averaging over 3 fireballs smashed per rivet hammer), but he cannot be so awesome that, for example, he tosses a coin and gets heads 60% of the time.

Apologies for not responding sooner, I had been dealing with some health issues over the past week that had me unable to do much of anything. Thankfully things are much better now.

On the issue of the stats analysis I think at this point I am planning to hold off as I just finished catching up with the TG dispute thread (wow it was like 40 pages of updates filled with a lot of new info! A hearty "Great Work!" to those involved BTW) and I'm not sure at this point how much the stats analysis will add.  I think there is probably something there but the important thing has already been demonstrated, which is to say that it was shown that the RNG was favorable, digging in to quantify the 1 in X chance at this point won't add very much above and beyond what we already know (e.g. that it was in fact favorable RNG and thus Mr Mitchell cannot claim that it was low RNG and thus served no benefit to him).

In essence, I think in light of the overwhelming MAME evidence with the new understanding of the way finger girder transitions can occur in older versions of MAME and the identification of the girder collapse transition the fact of MAME use is conclusively proven barring remarkable and reproducible new evidence.  As such the stats question is only relevant to show that yes in fact the runs are on the high RNG side speaking to the motive for using MAME.  The degree to which they are beneficial seems, to me at least, irrelevant given the rest of the facts in evidence.

edit: edited for grammar and to remove excessive quote nesting

[dwayne]

“Today, he is co-founder of the Kong Off Donkey World Champoionships”

Uhhh….what?

the lying no no bounds till we put an end to it by getting the truth exposed!

[smf]

This is one of the points of contention, which I believe Jeremy addressed in the original POST

It seems I misunderstood the original post. I've changed the orientation you get with mame -norotate in 0.196.

[Sqrlmonger]

Jace recently posted a bunch of information regarding Mr Mitchell's recording methods and how they seem to be capturing a mono-chromatic recording.

I don't have a great deal of expertise on this but I believe this is most likely due to a mismatch in subcarrier frequency that is actually carrying the color.

Essentially the signal is broken up into luminance and chrominance (and some other bits and bobs as I recall) and luminance is enough to get a B/W image and chrominance gives you your color.  But in order to get the color right the recording device has to have a key reference to the base or zero-level frequency is to make sense of it.  I think a lot of hardware will throw it away if it figures out the signal is garbage which is why you get a B/W recording.

I'm speculating a bit (more) on this next bit but I believe the monitors handle it to some degree because they are actually hooked up to the source during playback, but don't hold me to that bit.

Based on NTSC standards a normal domestic (US) VCR is probably expecting a ~3.58 Mhz chrominance subcarrier frequency but I have no idea what the setup is actually generating.  Taking a look at some of the PAL standards it's possible there are some VCRs out there that just might get some color out of the setup, but no telling if it would be accurate.

I'm not sure how much of this information would be news to anyone at TG, but it's what I could come up with scraping my brain plus a few minutes of googling.  Take it with a huge grain of salt and verify with others before taking it as gospel, etc.. etc..

[YesAffinity]

^Good info.

The crystal on the TBS encoder is 3.57mhz.  The instructions say to convert it to PAL, replace the crystal with one of frequency 4.48 mhz and move a jumper.  I tried moving the jumper alone, and got a black/white signal out of it that was black/white on all devices.