thread

Robotron CPU performance in relation to marathoning?

User Tag List

Page 1 of 4 1 2 3 ... LastLast
Results 1 to 10 of 39
  1. VERIFIED Senior Member Credibility: 11,761
    Join Date
    Mar 2009
    Posts
    1,892
    Thanks (Received)
    295
    Likes (Received)
    839
    Blog Entries
    3
    Mentioned
    6 Post(s)
    Tagged
    21 Thread(s)
    Follows
    38
    Following
    22

    Robotron CPU performance in relation to marathoning?

    I recently acquired a Robotron which I'm trying to stabilize so Tim Nibbler McVey can start marathoning. I'm looking for some technical insight about these observations.

    Background:
    I know from playing 19in1 Robotron PCB and xbox 360 robotron, that those versions have some inconsistencies from the real game.
    --The player moves a bit slower in relation to the enemies.
    --The speed of the projectiles is a lot faster and the amount of projectiles put on screen is greater.
    --On xbox the speed of the hulks is faster.
    --by wave 28 the game is wickedly hard.
    --personally watching Todd R. and Tim M, who are marathon level players on real robotron, they too struggle by the 20 series of waves.
    --I have attributed this to running the program in a faster/more efficient processor environment.

    In playing 2 real robotrons this year, and watching the Redelf million point game on youtube, I know that
    --the speed of the projectiles isn't as great for the tanks, as 19-1
    --the aggressiveness of the brain's projectiles is much less than 19-1
    --the time it takes the brains and tanks to create projectiles is affected by the cpu performance.
    --on busy screens like brain waves the cpu seriously bogs down to very slow.
    --much like joust, the cpu slowdown gives you time to really pick off the enemy before they can react.
    --the result is that the tank wave in the early 30 wave on real robotron is EASIER than tank wave 7 on 19in1 Robotron.

    Observations on real robotron:
    --I have 2 original CPUs. One came from a joust (has a service sticker from 1994 that indicates it was in a stargate at one point in time) Other is the early stargate rev b board that has jumper wires to make it compatible with later games like robotron.
    --both cpu's have original parts, but some of the ram chips were faulty on one board
    --the same 24 old but working 4116 ram chips in either cpu created same performance as per above. (bogs down during heavy screen enemies)
    --i put in about 10 4164 converted ram chips from Real Bob Roberts. cpu performance the same
    --so then i put in 2 brand new video decoder chips and 24 brand new 4116 ram chips from real bob roberts.
    --hardness setting at default of 5.
    --the cpu performance immediately increased so the game didn't bog down as much. the projectiles became more prolific and fast. I'd say it double the hardness level of the real game so it is more on par with 19-1 robotron on setting 1.


    So how do you explain this. If the age/brand of Rom can greatly affect the performance I've observed it seriously changes the level of play needed to marathon the game.

    Who has a technical explanation for these observations?

    by the way, I wasn't a robotron player in the 80s. only learned it this year after hearing Eugene J. speak at the milwaukee pinball show. i'm about a wave 28 player on 19 in 1 robotron. I've hit wave 58 on real robotron. same skills but greatly different outcomes depending on the game platform.
    Are you inspired to find the joy in gaming?

    TG employee (82-84)
    robotron2084guidebook.com/
  2. Senior Member Credibility: 1,000
    Join Date
    Jun 2008
    Posts
    1,318
    Thanks (Received)
    6
    Likes (Received)
    1
    Mentioned
    0 Post(s)
    Tagged
    0 Thread(s)
    Follows
    4
    Following
    3

    Re: Robotron CPU performance in relation to marathoning?

    The 19-in-1 and xbox are emulations and don't necessarily behave exactly like real hardware.
  3. VERIFIED Senior Member Credibility: 11,761
    Join Date
    Mar 2009
    Posts
    1,892
    Thanks (Received)
    295
    Likes (Received)
    839
    Blog Entries
    3
    Mentioned
    6 Post(s)
    Tagged
    21 Thread(s)
    Follows
    38
    Following
    22

    Re: Robotron CPU performance in relation to marathoning?

    i may have thrown you off with too much detail.
    the 19 in 1 references were to set the contrast/context for baseline of observing real gameplay.

    what i'm curious about is why brand new 4116 ram and 2 new decoder chips made the game play much more robust, than with the original chips. on REAL Williams robotron....
    Are you inspired to find the joy in gaming?

    TG employee (82-84)
    robotron2084guidebook.com/
  4. Senior Member Credibility: 1,000
    Join Date
    Jun 2008
    Posts
    1,318
    Thanks (Received)
    6
    Likes (Received)
    1
    Mentioned
    0 Post(s)
    Tagged
    0 Thread(s)
    Follows
    4
    Following
    3

    Re: Robotron CPU performance in relation to marathoning?

    Quote Originally Posted by 1500points
    i may have thrown you off with too much detail.
    the 19 in 1 references were to set the contrast/context for baseline of observing real gameplay.

    what i'm curious about is why brand new 4116 ram and 2 new decoder chips made the game play much more robust, than with the original chips. on REAL Williams robotron....
    Ah, sorry, I missed that part :oops:
  5. Senior Member Credibility: 1,000
    Join Date
    Oct 2007
    Location
    Seattle, WA
    Posts
    672
    Mentioned
    1 Post(s)
    Tagged
    1 Thread(s)
    Follows
    3
    Following
    3

    Re: Robotron CPU performance in relation to marathoning?

    Quote Originally Posted by 1500points
    i may have thrown you off with too much detail.
    the 19 in 1 references were to set the contrast/context for baseline of observing real gameplay.

    what i'm curious about is why brand new 4116 ram and 2 new decoder chips made the game play much more robust, than with the original chips. on REAL Williams robotron....
    I don't know, but I do wonder if it's possible you're not really comparing apples to apples because no two games of Robotron are ever alike. Reading this...

    Quote Originally Posted by 1500points
    In playing 2 real robotrons this year, and watching the Redelf million point game on youtube, I know that
    --the speed of the projectiles isn't as great for the tanks, as 19-1
    --the aggressiveness of the brain's projectiles is much less than 19-1
    --the time it takes the brains and tanks to create projectiles is affected by the cpu performance.
    --on busy screens like brain waves the cpu seriously bogs down to very slow.
    --much like joust, the cpu slowdown gives you time to really pick off the enemy before they can react.
    --the result is that the tank wave in the early 30 wave on real robotron is EASIER than tank wave 7 on 19in1 Robotron.
    a few things jump out at me. First, regarding "the speed of the projectiles isn't as great for the tanks, as 19-1," how do you know? The speed of the tanks' projectiles is not a constant. They move the slowest when you are in the center of the screen and the speed gradually increases as you move further away from center and closer to the sides of the screen. That's why being close to a side is so dangerous on tank waves, because their projectiles are screaming fast and because you're in greater danger of getting hit when they ricochet off the wall. Since tank projectile speed is variable based on the player's location I don't see how you can make direct comparisons, whether between original hardware game play and newer hardware game play or between arcade Robotron in general and emulated versions such as 19-in-1 and Xbox 360.

    Second, regarding "the aggressiveness of the brain's projectiles is much less than 19-1," on my own Robotron machine the brains' projectiles are always aggressive when there are more of them. When there are only a few (say, one to five) brains left on the wave their projectiles seem to meander more. I have no data to back this up; it's just my own observation.

    "the time it takes the brains and tanks to create projectiles is affected by the cpu performance." That may be true for the tanks but for the brains it is strictly a matter of when they capture the "important" human and has nothing to do with CPU performance or memory speed. If brains start creating zombies very early in the wave their projectiles start streaming out and they home in on you like heetseekers. That's why it's so important to protect Mikey as long as possible. If memory serves, in later brain waves the important human changes and it becomes impossible to know which one it is since there are duplicates, but on the first brain wave you can test this by finding Mikey ASAP (there is only one of him on wave 7) when the wave begins and keeping the brains from touching him. As long as the brains don't capture Mikey they don't create zombies and they don't shoot projectiles. Again, this seems like an apples to oranges comparison.

    "the result is that the tank wave in the early 30 wave on real robotron is EASIER than tank wave 7 on 19in1 Robotron." For what it's worth, I often find the later tank waves a little easier than the first one (wave 7) when playing on my own Robotron machine as well. Not sure why. But I've noticed some tank waves spawn a lot more tanks than do other tank waves. What causes more tanks to be spawned? I'm not sure but I have to believe that shooting more of the bouncing balls early in the wave has something to do with it. The more of those I take out early the easier the wave is because fewer tanks get spawned, because there are fewer bouncing balls around. If this cause is correct (and I believe it is), then comparing one tank wave to the next becomes much harder because it's all a function of how many bouncing balls you take out early. If you're a bad shot early in the wave you will have to contend with tons more tanks before the wave ends. Again, this does not appear to have anything to do with CPU or memory performance. Although I do agree that it is possible that if there are too many tanks on the screen at once the original hardware can slow down, making it easier. I still maintain tank waves are easier far more because you stay near the center of the screen and, to a lesser extent, because you take out more bouncing balls early, reducing the number of tanks to contend with.

    Regarding "on busy screens like brain waves the cpu seriously bogs down to very slow" and "much like joust, the cpu slowdown gives you time to really pick off the enemy before they can react," no argument there. These are both correct observations in my opinion, although on original Robotron hardware the player's movement and shot speed also slow down. It makes it easier not because your enemies are slower than you are but because everything is slower, giving you more time to assess and react.

    I don't have access to a 19-in-1 with Robotron nor do I have an Xbox 360 or easy access to one, so maybe I'm not seeing something that would be obvious if these systems were in front of me. I do believe your general question about CPU performance is valid, though, as Robo players have observed the difference between playing on original hardware and playing on MAME. MAME plays harder because more modern computer CPUs ("more modern" being any Intel or AMD CPU since at least the early '90s) can easily keep up with drawing more on the screen at once, and hence the waves don't slow down as much when there is a lot going on. The MAME devs apparently struggled with this issue for a while (how to emulate the slowdown the original CPU suffers) before giving up.

    All this being said, admittedly I am not a Robotron expert, but I do own my own all original machine and I do play it semi-regularly, and I have been making rapid progress in my skills on it since this summer. I just pushed my personal best from 1,417,xxx (wave 52) to 2,024,700 (wave 72) only yesterday. My observations are based on my own experiences as well as on conversations I've had with other, much more expert Robo players over the past four years since I got back into this hobby. There have been a lot of Robotron play discussions on forums such as this one where I've learned a lot about the game's AI, and several of my friends are world class level players.
  6. VERIFIED Senior Member Credibility: 11,761
    Join Date
    Mar 2009
    Posts
    1,892
    Thanks (Received)
    295
    Likes (Received)
    839
    Blog Entries
    3
    Mentioned
    6 Post(s)
    Tagged
    21 Thread(s)
    Follows
    38
    Following
    22

    Re: Robotron CPU performance in relation to marathoning?

    Today i switched the CPUs from the new ram cpu to the one with old ram, in real robotron.
    the difference is dramatic. I should probably record some of the waves. it's shocking how much different it is.

    on cpu with old ram i can play an unfocused game to wave 58.
    on the cpu with new ram i can play a VERY focused game to wave 40 and i have once gotten to 58.
    then i can switch to 19-1 robotron and squeak out a wave 24 to 28 game with all my focus.
    my play strategies and skills aren't changing during this testing, so i'd call it a constant.

    I just started picking apart robotron this spring and i have been fascinated by Eugene Jarvis' design of this game.
    Quite an addictive adrenaline rush.

    On real robotron with old ram with just 1 brain on screen. you can sit for up to 5 seconds before a shot is fired. on real robotron with new ram, the first brain shot is within a second. on 19-1 robotron the shot is instantaneous and it is on you in a flash.

    so here is the question. watch the redelf 1 million point game. you can see some pretty weak aggression by the enemy at times. I'd like to know what ram was in that CPU. I'm suspecting his game play wouldn't have been a 1 million 5 man game when using all brand new 4116 ram chips.
    Are you inspired to find the joy in gaming?

    TG employee (82-84)
    robotron2084guidebook.com/
  7. VERIFIED Senior Member Credibility: 1,019
    Join Date
    Oct 2007
    Location
    Portland Oregon
    Posts
    233
    Thanks (Received)
    20
    Likes (Received)
    87
    Live Streaming Channel(s)
    View Channel: kencade
    Mentioned
    1 Post(s)
    Tagged
    1 Thread(s)
    Follows
    15
    Following
    3

    Re: Robotron CPU performance in relation to marathoning?

    I guess I'll be the 'expert' here having played on MANY robotron machines over the year and on several different platforms.

    I can tell you to my eye and experience level, the different platforms do seem to play different. It is a chore for me to marathon MAME where its relatively easy for me to marathon any real machine. I have played on the JROK machine at one of our local collectors house, I do not see any difference in gameplay on it vs my cab vs Santes cab. I do see a difference in controllers and that seems to make it harder to marathon and play at the high level I can play at on my machine at home.

    If you really think you see 500% difference, 5 seconds per bullet vs 1 second, you should video tape it and really analyze it. The actually game play with one enemy, one player and one projectile should be virtually identical no matter the speed of the processor and ram. It is governed by the clock crystal and the software as far as I know. You may have a defective RAM in either the new or old batch causing this, if your observations indeed prove accurate.

    We can setup a test scenario on the first brain wave where you trap a brain on the edge and can easily pick off the shots allowing you to time the interval between shots. Another test would be to save all but 1 grunt on wave 1 and outrun him as long as you can. Over many runs you should be able to get a good baseline for how long on average it takes to get overtaken on each platform.

    I would be willing to broadcast my test and someone could time maybe 30 bullets and 30 escape attempts, that would give us a starting place from my machine, which more than one player has commented is 'harder' than other machines. Even John McAllister has said it but scores dont really bear this out as I have scored 835k (my personal best on any machine) 5man and he has done 957k (his third highest score if I recall correctly) on it.

    I look forward to seeing how this little robotron science project works out.

    -Ken
  8. VERIFIED Senior Member Credibility: 11,761
    Join Date
    Mar 2009
    Posts
    1,892
    Thanks (Received)
    295
    Likes (Received)
    839
    Blog Entries
    3
    Mentioned
    6 Post(s)
    Tagged
    21 Thread(s)
    Follows
    38
    Following
    22

    Re: Robotron CPU performance in relation to marathoning?

    Ken. as in Ken House? You don't know me, but we've met at the Twin Galaxies show in Ottumwa, i believe. (I'm an original Twin Galaxies employee, yada yada.....) Tim references you as a great robotron player often.

    anyway, your ideas on quantifying this are interesting.
    I believe Tim McVey has committed to buy my machine so i presume he'll be streaming gameplay this winter.

    Here is a tell tale sign that might also be an indicator the ram is questionable.
    when the brains go down to the bottom of the screen and park. like they get stuck and they don't fire. i don't see that on robotron with all new ram, nor on the 19-1 platform.

    (the 19-1 has wico 8 ways as Tim thought this was the missing variable to the difference...it wasn't)

    I finally hit my goal last night which was rolling the waves (2.7m/wave 105).
    This game is amazing. man, i had joust figured out and mastered way faster than robotron. i wasn't expecting such a challenge in the "reflex" learning curve.

    On a different subject. let me ask you about tanks. i've played dodge ball with them alot just to figure out the design characteristics. it appears to me that they are operating on a double bank shot and boxing algorithm. by shooting a series of shots that vary by trajectory and velocity. this helps them be roughly dead on no matter which way the player tries to move. i see less that it is about your distance from them affecting their velocity, than it is their multi-ball strategy for nailing you. is that a correct observation?
    Are you inspired to find the joy in gaming?

    TG employee (82-84)
    robotron2084guidebook.com/
  9. VERIFIED Senior Member Credibility: 11,761
    Join Date
    Mar 2009
    Posts
    1,892
    Thanks (Received)
    295
    Likes (Received)
    839
    Blog Entries
    3
    Mentioned
    6 Post(s)
    Tagged
    21 Thread(s)
    Follows
    38
    Following
    22

    Re: Robotron CPU performance in relation to marathoning?

    Here is some technical insight from Clay which seems to support the observations.
    The CPU has everything to do with the performance experienced.
    So that would tend to lend itself to new ram having a performance effect.
    Looks like caps and crystals would come into play too.

    So my final view on this topic as I let it drop, is.....
    Robotron as the code is designed, and run from an efficient/fast/new school CPU is too hard for the human player.
    The marathoner is benefiting from the "lag" created on the old school CPU by lots of action on the screen.

    I'm going to say if an old original CPU was updated with fresh components, that the 1 million 5 man score would not happen.

    Fascinating stuff. :)

    http://forums.arcade-museum.com/showthr ... 82&page=71
    Are you inspired to find the joy in gaming?

    TG employee (82-84)
    robotron2084guidebook.com/
  10. VERIFIED Member Credibility: 1,057
    Join Date
    Jan 2011
    Posts
    56
    Thanks (Received)
    3
    Likes (Received)
    9
    Mentioned
    0 Post(s)
    Tagged
    0 Thread(s)
    Follows
    16
    Following
    3

    Re: Robotron CPU performance in relation to marathoning?

    Hi 1500points.

    I took the liberty of forwarding your initial post to the KLOV forums - I hope you don't mind (?). This is just too interesting to be dropped so soon.
    (I've been thinking about wether this could be done for Defender (Williams 1980) - to create a super fast machine, without the lag and disappearing enemies. Played on the Blue/Green roms, and set to maximum difficulty (99-99), this would be a monster!)

    Link to the thread on the KLOV forums: http://forums.arcade-museum.com/showthread.php?t=206368
    DEFENDER: 999.975 pts (Blue Roms / max difficulty: 99-99 / Bonus ship level: 10K)
Page 1 of 4 1 2 3 ... LastLast