Billy Mitchell's Direct Feed Recordings are MAME Generated

[bensweeneyonbass]

Rastafarianization and screen artichokes are the only things those snakes can't fool us with. This whole magic of "ArrInnGee" is something I don't believe in...

[Bounty Bob]

looks like emu + ridiculous rng + no reliable witnesses = cheating

^ looks like it.

Why so little focus on the rng by the believers?

I know Jojo and bh and possibly others have been noting down the numbers.

Does the ridiculous rng look consistent across all three 3 of the games under discussion?

And, are there any bold, clear statistics that should be obvious red flags to anybody who's not an expert at the game? The fact that only 4 of the first 50 blue smashes (in the game under review at TG) registered as 300s certainly should raise an eyebrow for anyone who's only played DK half a dozen times.

Are there any other startling and easy to digest stats like that knocking about? Or are we just going to be bogged down in rastafarianization and screen artichokes (I think those are the technical terms) and such for ever?

The trouble with looking to the rng for proof is that it is literally random. As unlikely and improbable as the outcomes are, they aren't impossible and so it can't be used as proof of cheating on it's own. When combined with everything else though, things start to look more and more suspicious. But for absolute proof we need something which is irrefutable.

[Sock Master]

lots and lots of words

It's funny how the narrative trying to explain away how Billy's games look like MAME keeps changing every time the previous one gets debunked.

Billy cheated.  Billy's defense is cheating too as the current narrative contains my own words in plagiarized form explaining how that board works.  Changing stories and using whatever bits of text they can find online to splice together new stories shows their fundamental lack of understanding about how the hardware actually works.

Their current explanation is that by connecting this device it will somehow magically cause a DK PCB to change the way that it renders frames in such a way that it looks like MAME?

I'll save everyone some time:

No.

I'll be eagerly awaiting the next new story.  I hope it's more inventive than the last one.

[gstrain]

Mame is an Emulator that is designed to Mimic the processors on the Arcade board. It will follow the commands and rules set in the instruction codes programmed in the Rom. If The Rom states to draw letters first, Floor Second, and Characters last; The Mame is follow those commands just like the Arcade is designed to follow the commands of the program ROM.

The issue at hand is that due to an emulation "defect" in MAME, it does not draw quite the same as Arcade.  We can see those same drawing differences in Billy's tapes. 

He never expected to be questioned over 15+ years later about some recording method he did in the days of VCR and VHS tapes. But we want to get the answers that everyone are now asking out as quick as possible.

The Boomers 1,062,800 score was from July 31st, 2010.  That's less than 8 years ago.

[LMDAVE]

looks like emu + ridiculous rng + no reliable witnesses = cheating

Why so little focus on the rng by the believers?

I know Jojo and bh and possibly others have been noting down the numbers.

Does the ridiculous rng look consistent across all three 3 of the games under discussion?

And, are there any bold, clear statistics that should be obvious red flags to anybody who's not an expert at the game? The fact that only 4 of the first 50 blue smashes (in the game under review at TG) registered as 300s certainly should raise an eyebrow for anyone who's only played DK half a dozen times.

Are there any other startling and easy to digest stats like that knocking about? Or are we just going to be bogged down in rastafarianization and screen artichokes (I think those are the technical terms) and such for ever?

Let me start this by saying I'm not flip-flopping on my stance of Billy tapes are suspect, but I do want to give my own fair analysis of the RNG from what I've read.

I would have liked to have seen the RNG issue broken down a little more. I kept getting confused about the 4 out of 50 blue smashes being 300, but then seen it was clarified as being  the first 50 "blue barrels", so that even discount the fireballs and only focuses on blue barrels. Still...4 is a lot less than the statistical 12.5 out 50 one would expect from even 25/50/25 distribution.

I looked at Wes' chart about the 1,050,200 game. I would have like to have seen the average value of the blue smashes per board type instead of lumped in the end. Because Billy had much high blue hits on conveyors and rivets than average, and that could be contributed to never taking the "Free Ride" on conveyors, and averaging 3 hit per hammer on rivets (a nice outcome of course).

I've been playing this game pretty regular over the last 10 or so years and consider myself one of the more advanced players. So, I can give basic analysis without even breaking down the statistical averages, I just kind of know what to expect at this point of playing Donkey Kong. And that is, blue hits on barrels tend to be pretty normal 25/50/25, conveyors always tend to yield more 500/800's (that's why getting the hammer on conveyors every time is not a bad move), and on rivets you tend to get way more 300's. And that's just my observations from playing the game and not dissecting code.

One big thing from Wes' chart was the percentage of final score from blue smashes, 17.1% for Billy is rather high, but that could be contributed to lack of barrel grouping. The scores we are seeing now over the last 3 or 4 years are being pushed up due to "grinding it out" with barrel grouping, so it would make sense that other player's percentage of total game from blue smashes wouldn't be as high. (LOL...I wouldn't be surprised if Wes or Robbie could pull off a 1.1M game with single hammer if either of your are looking for a new challenge)

Overall, the average blue smash for Billy of 491 is only 5 pts more than Wes' 486, and could be contributed to lining up more hits on conveyors and not taking a free ride. If there's another thread breaking down the numbers more I would like to see that one.

[xelnia]

looks like emu + ridiculous rng + no reliable witnesses = cheating

Why so little focus on the rng by the believers?

I know Jojo and bh and possibly others have been noting down the numbers.

Does the ridiculous rng look consistent across all three 3 of the games under discussion?

And, are there any bold, clear statistics that should be obvious red flags to anybody who's not an expert at the game? The fact that only 4 of the first 50 blue smashes (in the game under review at TG) registered as 300s certainly should raise an eyebrow for anyone who's only played DK half a dozen times.

Are there any other startling and easy to digest stats like that knocking about? Or are we just going to be bogged down in rastafarianization and screen artichokes (I think those are the technical terms) and such for ever?

Let me start this by saying I'm not flip-flopping on my stance of Billy tapes are suspect, but I do want to give my own fair analysis of the RNG from what I've read.

I would have liked to have seen the RNG issue broken down a little more. I kept getting confused about the 4 out of 50 blue smashes being 300, but then seen it was clarified as being  the first 50 "blue barrels", so that even discount the fireballs and only focuses on blue barrels. Still...4 is a lot less than the statistical 12.5 out 50 one would expect from even 25/50/25 distribution.

I looked at Wes' chart about the 1,050,200 game. I would have like to have seen the average value of the blue smashes per board type instead of lumped in the end. Because Billy had much high blue hits on conveyors and rivets than average, and that could be contributed to never taking the "Free Ride" on conveyors, and averaging 3 hit per hammer on rivets (a nice outcome of course).

I've been playing this game pretty regular over the last 10 or so years and consider myself one of the more advanced players. So, I can give basic analysis without even breaking down the statistical averages, I just kind of know what to expect at this point of playing Donkey Kong. And that is, blue hits on barrels tend to be pretty normal 25/50/25, conveyors always tend to yield more 500/800's (that's why getting the hammer on conveyors every time is not a bad move), and on rivets you tend to get way more 300's. And that's just my observations from playing the game and not dissecting code.

One big thing from Wes' chart was the percentage of final score from blue smashes, 17.1% for Billy is rather high, but that could be contributed to lack of barrel grouping. The scores we are seeing now over the last 3 or 4 years are being pushed up due to "grinding it out" with barrel grouping, so it would make sense that other player's percentage of total game from blue smashes wouldn't be as high. (LOL...I wouldn't be surprised if Wes or Robbie could pull off a 1.1M game with single hammer if either of your are looking for a new challenge)

Overall, the average blue smash for Billy of 491 is only 5 pts more than Wes' 486, and could be contributed to lining up more hits on conveyors and not taking a free ride. If there's another thread breaking down the numbers more I would like to see that one.

I haven't delved too deep into the 1.05M game yet, but a further analysis of the 1.047M game shows nothing out of the ordinary in terms of variable smash RNG. Analyzing that tape was difficult and I think there was a bit of a rush to judgment before it could be independently verified by multiple people. You can see my breakdown here: https://goo.gl/4BY7gW

Note that I only transcribed variable smashes (blue barrels, pies, fireballs)...no other considerations are made regarding RNG or scoring. Pies are tricky to analyze outside of MAME because the scoring bug that forces some of them to always be 300.

[Sqrlmonger]

looks like emu + ridiculous rng + no reliable witnesses = cheating

Why so little focus on the rng by the believers?

I know Jojo and bh and possibly others have been noting down the numbers.

Does the ridiculous rng look consistent across all three 3 of the games under discussion?

And, are there any bold, clear statistics that should be obvious red flags to anybody who's not an expert at the game? The fact that only 4 of the first 50 blue smashes (in the game under review at TG) registered as 300s certainly should raise an eyebrow for anyone who's only played DK half a dozen times.

Are there any other startling and easy to digest stats like that knocking about? Or are we just going to be bogged down in rastafarianization and screen artichokes (I think those are the technical terms) and such for ever?

Let me start this by saying I'm not flip-flopping on my stance of Billy tapes are suspect, but I do want to give my own fair analysis of the RNG from what I've read.

I would have liked to have seen the RNG issue broken down a little more. I kept getting confused about the 4 out of 50 blue smashes being 300, but then seen it was clarified as being  the first 50 "blue barrels", so that even discount the fireballs and only focuses on blue barrels. Still...4 is a lot less than the statistical 12.5 out 50 one would expect from even 25/50/25 distribution.

I looked at Wes' chart about the 1,050,200 game. I would have like to have seen the average value of the blue smashes per board type instead of lumped in the end. Because Billy had much high blue hits on conveyors and rivets than average, and that could be contributed to never taking the "Free Ride" on conveyors, and averaging 3 hit per hammer on rivets (a nice outcome of course).

I've been playing this game pretty regular over the last 10 or so years and consider myself one of the more advanced players. So, I can give basic analysis without even breaking down the statistical averages, I just kind of know what to expect at this point of playing Donkey Kong. And that is, blue hits on barrels tend to be pretty normal 25/50/25, conveyors always tend to yield more 500/800's (that's why getting the hammer on conveyors every time is not a bad move), and on rivets you tend to get way more 300's. And that's just my observations from playing the game and not dissecting code.

One big thing from Wes' chart was the percentage of final score from blue smashes, 17.1% for Billy is rather high, but that could be contributed to lack of barrel grouping. The scores we are seeing now over the last 3 or 4 years are being pushed up due to "grinding it out" with barrel grouping, so it would make sense that other player's percentage of total game from blue smashes wouldn't be as high. (LOL...I wouldn't be surprised if Wes or Robbie could pull off a 1.1M game with single hammer if either of your are looking for a new challenge)

Overall, the average blue smash for Billy of 491 is only 5 pts more than Wes' 486, and could be contributed to lining up more hits on conveyors and not taking a free ride. If there's another thread breaking down the numbers more I would like to see that one.

Good info, and I think we do need a very detailed breakdown as well.

The thing is that without a thorough breakdown of this by board/type I'm concerned we are open to even more uncertainty.

If specific categories of RNG are in line with other players while other specific categories of RNG are high then it would indicate cheating to me as true RNG should show us that these categories where high variance was achieved would change from board to board and game to game where as MAME splicing games together would allow for a human who is focused on specific RNG elements to create those high variances within the specific categories they are focused on.

But as you point out the category grouping also allows us ( and by us I mean those who actually understand the strategy to a serious level, which is certainly not me) to assess whether strategic choices made during the performance could explain some of the variance.

I think if we break it down we will either see randomness in the RNG upswings as it moves from category to category across boards and games which would be indicative of non-cheated games or we will see even higher variance (it would have to be higher to be visible when clumped with all categories) for specific boards across multiple games and that, to me, would be strong evidence of cheating.

On that point I do want to state that this idea that RNG is not "proof" is just not a serious position to me.  It has to cross certain thresholds certainly, for instance RNG that is 1 in 1000 runs is not proof, but if you show me 1 in 1000 RNG or better occurring on every barrel board across three runs we aren't talking about 1 in 1000 anymore, we are talking about rock solid evidence because we are talking about 1 in 1x10^189 - we are talking about odds so ridiculous that every single one of the 7 billion humans on earth could have been playing DK from the big bang until now with each game taking 3 hours and we still will only have completed 2.86x10^23 games.  If you even understand the math on this a little bit you know how completely and utterly ridiculous this notion is.  Even a single game with 1 in 1000 on just the barrel boards and no other anomalies would be 1x10^63.  Again, my point here is not that this evidence has already been produced but that if it were then we would be talking about smoking guns in my view.

[colecomeister]

Maximumsteve - as an addendum to Robert Childs' video analysis, and the extensive commentary he offered on the set-up, I wondered if he could provide analysis or comment on something also familiar to him - two of the three videos he uploaded back in August 2010 in order to document Billy's back-to-back high scores at Boomers on July 31, 2010. He has never weighed in on questions posed about them.

In Robert's "Billy Mitchell speaks (july,2010) moments after breaking donkey kong record" video, Billy's machine has DK Junior installed, not Donkey Kong. This poses obvious problems about DK play.

- fast forward to 2:00 mark - DK Junior visible at 2:02 (frame advance)

Is Robert able to re-upload the original version of the "After DK record Billy Mitchell goes for donkey kong junior record" video? - for unknown reasons, Robert dubbed over the copy with classical music, and the viewer can't hear what Robert is describing about the cabinet (direct feed?) set-up for Billy's high score run.

[LMDAVE]

I haven't delved too deep into the 1.05M game yet, but a further analysis of the 1.047M game shows nothing out of the ordinary in terms of variable smash RNG. Analyzing that tape was difficult and I think there was a bit of a rush to judgment before it could be independently verified by multiple people. You can see my breakdown here: https://goo.gl/4BY7gW

Note that I only transcribed variable smashes (blue barrels, pies, fireballs)...no other considerations are made regarding RNG or scoring. Pies are tricky to analyze outside of MAME because the scoring bug that forces some of them to always be 300.

Thanks. I was wondering if you every did a "memorable game" chart on either of those games. But looking at your spread sheet I guess my gut feel that more 500/800 on pies is not correct at all. Just seems that way sometimes.

[bh]

I haven't delved too deep into the 1.05M game yet, but a further analysis of the 1.047M game shows nothing out of the ordinary in terms of variable smash RNG. Analyzing that tape was difficult and I think there was a bit of a rush to judgment before it could be independently verified by multiple people. You can see my breakdown here: https://goo.gl/4BY7gW

Note that I only transcribed variable smashes (blue barrels, pies, fireballs)...no other considerations are made regarding RNG or scoring. Pies are tricky to analyze outside of MAME because the scoring bug that forces some of them to always be 300.

Thanks. I was wondering if you every did a "memorable game" chart on either of those games. But looking at your spread sheet I guess my gut feel that more 500/800 on pies is not correct at all. Just seems that way sometimes.

The 1.04m game is really tough to transcribe. There are serious color errors and tons of blur. Jeremy found four barrels that I missed which make the run good but unexceptional. I have a lot more confidence in the 1.05m game. The tape is clearer and color better.

[Kewydee_17]

After literally acres of word salad, Jry's evidence still hasn't been disproved

I'm sure at some point there was a direct capture from an DK pcb set up.

Thing is, the direct feed from a DK pcb was NOT used to make those scores. MAME was used. The evidence is clearly there for all to see.

There is not the slightest chance of making a direct feed video from an DK pcb look like MAME in a frame by frame analysis.

Which is why we haven't seen such a video. And we won't.

Loyalty to ones friends is all well and good, but it's getting to ridiculous levels with the mental gymnastics used to NOT SEE the evidence for what it is.

Those who have chosen not to see will never be swayed by evidence. "Your hair is on fire mate", "No it is not, it's just a bit warm today"

[dwayne]

the tapes i was asked to put up were recorded off a sanyo tv.  I am not technically inclined.  I dont have the capacity to capture the video in a transfer to a computer. i used a canon camera that i have  vixia hv 30 to record to upload a file to youtube.

Im really a luddite. 

dwayne

[qnz]

Hi,

I tried to post this over on twingalaxies, but my ancient account can't post there for some reason.

(...)

I'm trying to piece your post together here and understand it. 
(...)
I won't delve further into your comment for the time being because of these timing issues, but if you can clean up or clarify your concerns in light of this information perhaps a more substantive discussion could be had.  Alternatively, let me know if I made a mistake in my understanding of what you were saying.

I wrote up a long, detailed reply to this explaining the part where I was quoting Jace.  After looking at it in preview, I think it's probably longer than anyone wants to read.  I have it saved, so if you really want to see it, let me know.

Unfortunately, it's getting too late here   Can someone summarize what the RNG argument is?  I.e., Here's the raw data, here's the processed data, this is what we think it shows and why?


Todd

p.s. I see Dwayne has answered my question about the videos.  Thank you, Dwayne.

[YesAffinity]

Short summary of the RNG arguement:

The number of blue barrels smashes is extremely high, well outside of the mean.  If someone were playing level by level in MAME and using the save state feature to ensure a particular pace, a possible result of that would be a higher frequency of blue barrels assuring that pace, possibly unbeknownst to the player.

In simpler terms, if you are using the save state approach to get to a certain score, and you complete a level but it's not quite at the pace needed at the end of the level, you go back to your save state at the start of the level and re-perform.  Once you get the pace needed at the end of the board, you save state and go to the next board.  Barrel levels offer more potential to "tweak" the pace, either through steering mayhem or through blue barrels (or both).  Steering introduces greater risk.  Getting better RNG resulting in a greater number of blue barrels does not.  The higher blue barrel numbers on Billy's performance (percentage of total points, total number of blue barrels) potentially correlates to MAME/save state usage.

[Sqrlmonger]

https://www.youtube.com/watch?v=234Y76_3YPE

This video has some issues, but there are some things that are new and quite a bit of it is true.

First, I want to point out the opening quote from Billy in that video actually continues on to say something like "but even worse still, would be to deny someone a legitimate score."  - No real bearing, just some context I found amusing.

Second, the video presents a case that the one game that does have audio out of the three being disputed still on TG shows signs of splicing.  I am not qualified to say if the things he points out are, in fact, proof of that.  Can anyone here take a look at the relevant section of the video (starting at 4:09 and going for about 10 seconds) and offer a more informed opinion on this audio evidence?

To be clear I am not suggesting you can render a verdict on the entire tape from what he shows, but does the video's assertion hold any water based on what is presented?