Todd's 5.51 Dragster score

[Nukey Shay]

Got to admit though Dragster was one of the few games I never attempted to fry.

 

Doubt you'd get anything useful. Frying simply begins a cold start powerup at a random position in the code, disrupting the intended place to begin. In most games (including this one), this is a small loop to clear all ram and registers. Starting a new race also clears some ram (at least the important ones), so it would affect only a current glitched race. This would work to the players DISadvantage, since the running timers could begin at a point other than zeros. Other effects might be randomly-placed missile sprites or incorrect sprite reflect enables...both of which would be quite obvious in a photo.

[Tempest]

Some people over on the TG thread are questioning some of Todd's other 'questionable' scores like his Skiing one. Maybe that will lead to someone analyzing the Skiing code as well. Regardless of what happens to Todd's scores, the analysis of the game code is a valuable byproduct.

[ZackAttack]

Regardless of what happens to Todd's scores, the analysis of the game code is a valuable byproduct.

 

Assuming that work is made publicly available.

[Thomas Jentzsch]

I have no problem posting my code and sheets.

[AlyoshaTAS]

 

I'm just being devils advocate here. We are all still learning about the quirks of the system, and a year from now will probably still be learning new things. Maybe this has nothing to do with this ROM and highscore, but it can't just be discounted just because we haven't discovered it so far.

 

I think one of the interesting things about Dragster in this regard is how little of the A2600 it needs to work. If it didn't need to read the controller, it could be run exclusively from the CPU in the exact same way (minus the time spent in !RDY.) Many other games at least use the collision detection in the TIA, but nope Dragster can do without it. I think it's this simplicity that lends a lot of weight to Omnigamer's analysis, the game cares very little about the world outside of the 6507. If it were more complicated I might be one of the people calling for a more complete model, but there really doesn't seem to be any need in this particular case.

 

Aside from that, my money is on prototype cart. Another record given to Todd Rogers is for Grand Prix, and that time certainly appears impossible, but would be a proper time if some of the car mechanics that slowed it down weren't in place at the time, i.e. a prototype, like the BarnStorming one was.

[Tempest]

 

i.e. a prototype, like the BarnStorming one was.

 

Has that actually been confirmed or is that just a theory?

[ZackAttack]

If it were more complicated I might be one of the people calling for a more complete model, but there really doesn't seem to be any need in this particular case.

 

Wouldn't a more complete model be required in order to know that it's not more complicated? Sounds like a self fulfilling prophecy to me.

[Shawn]

 

Has that actually been confirmed or is that just a theory?

 

It's just more BS cause Todd, while an amazing gamer, has multiple impossible records tarnishing his otherwise great gaming feats. It's like the uber baseball player who uses steroids even though they are already of the best in the world.

[D.Yancey]

Some people over on the TG thread are questioning some of Todd's other 'questionable' scores like his Skiing one. Maybe that will lead to someone analyzing the Skiing code as well. Regardless of what happens to Todd's scores, the analysis of the game code is a valuable byproduct.

Todd's Skiing game 3b score of 27.51 is not possible (without using the cheat) in my opinion. I am in 5th place on this one with a score of 27.97.

Todd's Starmaster game 1b (Ensign) score of 3,985 is not possible in my opinion. I have the 2nd place score at 3,972 which would be extremely difficult to beat by more than a couple of points.

Someone mentioned Grand Prix game 1b score of 29.47. I have a 2nd place score submission on that one of 29.61. Several have tied me since that time. I'm not sure if I contest that score or not. I never tried to improve my score any further, so this one may be OK.

I think that may be the only Activision scores by Todd that I disagree with.

[Tempest]

 

It's just more BS cause Todd, while an amazing gamer, has multiple impossible records tarnishing his otherwise great gaming feats.

That's kind of how I'm feeling about the whole thing right now.

[andrewg]

I've been interested in the Dragster 5.51 for many years at this point. I posted in the TG dispute thread, but wanted to lay out my thoughts here as well. What's most plausible? Well, that's what I've laid out:

 

http://www.twingalaxies.com/showthread.php/175364-Dispute-Dick-Moreland-Atari-2600-VCS-Dragster-NTSC-Game-1-Difficulty-B-Fastest-Time-Player-Todd-Rogers-Score-05-51?p=914703&viewfull=1#post914703

 

http://www.twingalaxies.com/showthread.php/175364-Dispute-Dick-Moreland-Atari-2600-VCS-Dragster-NTSC-Game-1-Difficulty-B-Fastest-Time-Player-Todd-Rogers-Score-05-51?p=914718&viewfull=1#post914718

 

 

I don't doubt Todd's abilities whatsoever. However, I do believe he has the ability to stretch the truth or not recalling facts 100% accurately.

Edited September 1, 2017 by andrewg

[jnice]

I don't know if anyone else caught it on Twitch, but someone was playing this game over and over and over and over the other night. I think this was his best attempt:

 

[andrewg]

I don't know if anyone else caught it on Twitch, but someone was playing this game over and over and over and over the other night. I think this was his best attempt:

 

 

Yeah, who was that guy........

[AlyoshaTAS]

 

Wouldn't a more complete model be required in order to know that it's not more complicated? Sounds like a self fulfilling prophecy to me.

 

I don't think so. the game dissassembly should privde enough guidence here. Ex. does it read from the collision registers or not.

 

It's just more BS cause Todd, while an amazing gamer, has multiple impossible records tarnishing his otherwise great gaming feats. It's like the uber baseball player who uses steroids even though they are already of the best in the world.

 

The reason I'm skeptical of outright lies is that it would very difficult in some of these cases to know what time to pick if it were just a lie. Maybe not so much for Dragster, but certainly for Barn Storming or Grand Prix.

Edited September 1, 2017 by AlyoshaTAS

[D.Yancey]

Here is a Youtube video of the official story of Todd's Dragster record as told by Todd himself. This is a NEW video today. Todd released the video on September 1, 2017 for the 35th anniversary of the world record.

Edited September 3, 2017 by D.Yancey

[Thomas Jentzsch]

Meanwhile the thread at TG goes on and on and on and... Mostly because the real problem is a pretty common one. It is usually waaay easier to prove something exists than to prove it does not.

 

In this case the definite proof would either be 5.51 on real hardware proven in public OR to provide the exact values for Omnigamer's and my spread sheets. The latter could then be very easily verified by a hacked autoplay ROM.

 

To definitely prove the opposite seems almost impossible. Yes, you can analyse the game code and built models based on that. But how can you prove that the analysis and models are 100% exact? Even for a simple game like Dragster there will always remain room for doubt.

 

So I am pretty sure that the score can only be confirmed, but not refuted so that everyone is convinced. And the thread at TG may go on and on and on and...

[eshu]

Meanwhile the thread at TG goes on and on and on and... Mostly because the real problem is a pretty common one. It is usually waaay easier to prove something exists than to prove it does not.

 

In this case the definite proof would either be 5.51 on real hardware proven in public OR to provide the exact values for Omnigamer's and my spread sheets. The latter could then be very easily verified by a hacked autoplay ROM.

 

To definitely prove the opposite seems almost impossible. Yes, you can analyse the game code and built models based on that. But how can you prove that the analysis and models are 100% exact? Even for a simple game like Dragster there will always remain room for doubt.

 

So I am pretty sure that the score can only be confirmed, but not refuted so that everyone is convinced. And the thread at TG may go on and on and on and...

Random idea - would it be possible to bruteforce every single possible input state for every frame in an accelerated emulator - seems highly parallelizable etc...

[andrewg]

Random idea - would it be possible to bruteforce every single possible input state for every frame in an accelerated emulator - seems highly parallelizable etc...

 

I came to the same conclusion.

 

My last post mirror what you two have just said here, and my suggestion to add the other 5.51 players to TG.

 

http://www.twingalaxies.com/showthread.php/175364-Dispute-Dick-Moreland-Atari-2600-VCS-Dragster-NTSC-Game-1-Difficulty-B-Fastest-Time-Player-Todd-Rogers-Score-05-51?p=917593&viewfull=1#post917593

[Thomas Jentzsch]

Random idea - would it be possible to bruteforce every single possible input state for every frame in an accelerated emulator - seems highly parallelizable etc...

Nope. We have 4 options each frame (shift or not, gas or not) and ~330 frames. So that would be 4^330 = ~4.8 * 10^198. Which is "a bit" more than the number of atoms of the universe (~10^85).

Edited September 2, 2017 by Thomas Jentzsch

[Omnigamer]

That type of brute-forcing also doesn't consider any effects that may come from dirtied state after extended use, only power-on. We pretty well understand those effects by now (according to the code), but if the point is to understand the realm of all possibilities, you would have to cast some of the knowledge aside and extend it out for many more (possibly infinite) frames.

[Keatah]

..or aging electronics. Or different revisions of hardware.

[+stephena]

..or aging electronics. Or different revisions of hardware.

 

I'm leaning towards the latter, either in the cart or the console itself. If it's the former, then I suspect it will be really hard to track down.

[+5-11under]

I don't know if I've every played Dragster (if I have, my record is probably somewhere in the 10-20 second range ), but anyway...

 

I'm wondering if the code does multiple direct input reads within a cycle?

 

For instance, here's some sample "code" for one cycle (start with A=0)

- IF joystick = left then A=A+1

- IF joystick = notleft then A=A+1

- IF fire = pressed then A=A+1

- IF fire = unpressed then A=A+1

If your timing is impeccable (and if the joystick and fire buttons can be read directly at any moment in time), then you could conceivably end up with A=4.

/just a thought

[Omnigamer]

5-11under that was one of my initial thoughts as well, but alas, the game only checks specific inputs once in a given frame. When the input is needed for multiple sets of calculations, like the clutch, it is latched into memory and utilized from there.

[Thomas Jentzsch]

Actually it's different between gas and clutch. The joystick (clutch, restart) is read every frame, but the fire button (gas) only every 2nd one. Since the calculations for each player alternate between frames, the latter is sufficient. And the clutch is stored in a variable, which is evaluated also only every 2nd frame. So effectively both, gas and clutch, are checked only once every 2nd frame.

 

For a brute force simulation you would have to accept that as a given parameter. Then you would "only" need 4^165 = ~2.2 * 10^99 runs.