The Long-Standing TG Policy on "Free Play"

Hello fellow gamers:

Now that TG is under new management, I wanted to air one long-standing policy that I think warrants serious discussion.

Arcade games (and MAME) have always been required to not be set to "Free Play". What I was told years back by some of my fellow staff members is that they were not 100% sure whether for some titles being on "Free Play" would in some capacity change the behavior of the AI, so the unilateral policy was that a coin must be inserted.

Another reason, which applied to far fewer titles, was that in some cases the presence of the credit in the game, whether due from a quarter/token or free play, prevented the "splash screen" from showing up which, depending on the title, was a valuable part of the adjudication process. For example, it was Chris Ayra who pointed out to me years back that when "Donkey Kong" already has a credit in you will not see that "Nintendo of America" logo show on the bottom of a certain splash screen (it's been awhile so I do not remember the specific details).

Anyway, seeing how so many scores were validated by TG from events such as "Mall of America" (2000-2001), Philly Classic (2004) and other venues over the years, where all titles were typically on "free play" for the duration of these events, it seems incongruous to have this provision part of the standard rules when so many already-validated scores over the years were achieved while in "free play" mode.

I am guessing that this policy still exists seeing how I still see it as part of the rules/settings as is appended to several variations that I have recently been looking at.

Thoughts ?

I think that "Free Play" should be optional unless there's proof for a particular game that it actually affects game play. It can be dealt with on a case by case basis.

Do you know why the free play rule was instituted?

I would agree. I don't see how it would affect gameplay, maybe in a certain game or two but that would be one of the "cases" where it would be flagged and changed for that title. I'd guess that in 99% of cases it doesn't change the game. Same for language options and other things that are not difficulty settings.

I'm no arcade expert, but I think that if it affects gameplay, only then should free play be disabled. Like, if it changed the HP of enemies, it should be off.

I have a computer programming background, so please bear with me.

It is possible that the "state change" of the software being in "free play" alters the outcome of the seeded psuedo-random number generator that the game uses.

What this means is that it is possible for games like Pac Man or Donkey Kong which probably utilize this method of creating randomness will produce a different experience depending on the state of the software when started. This is because it is possible that being in "free play" versus "1 Credit mode" changes the "seed" number that gets produced when the code that generates random numbers makes a call to that function.

From my point of view, the policy must have been put in place to try to make sure that everyone was playing from the exact same starting point with the same starting conditions. The smallest differential, no matter how slight, can in theory make a difference - and when you are dealing with top players who are THAT good, that small differential can mean the difference between a world record and not.

So, I would encourage the community to try to adhere to the well established TG policies whenever they can while adjudicating. Most of them evolved for a reason and were not just made up for the heck of it.

I hope that made sense,

Jace

With Berzerk it has nothing to do with freeplay. It has to do with starting your game before the splash screen has painted the first maze.
When the game is turned on the value fed into the RNG is based on zero. Since there is no true randomization back then, it was based on a series of addition/subtraction/multiplication, so the seed number was massaged then bits picked out. Those bits represented N, S, E, W for the pillars being created in the maze.

Thus if you want to play the exact same series of mazes to start your game, hit the start button before the first maze is drawn, before the pseudo-RNG has a chance to vary the seed value.

If a true Berzerk champion wants to show finesse, they play after the game has been on awhile ensuring there is no way to guess the starting set of mazes.
if interested a deeper analysis is here- http://www.robotron2084guidebook.com...mazegenerator/


But for more randomized games like Robotron, Joust, Defender, the concept of a coin drop is nothing more than superstition or adhering to long-term policy as Jace states above. It is founded in no programming logic on those titles.

I sent a message over to Matt Osborne on KLOV about this topic. He builds free play and high score save kits for golden era arcade hardware across various manufacturers and thus various hardware setups. I would say he is the top man in this space and has sold many of his kits. I figured his take on this would be relevant to the discussion and will post his reply.

With both Ms. Pac-Man and Pac-Man Plus, I believe that patterning mazes becomes possible if the game is set to free play. One simply plugs in the game and holds down the player 1 button, and the game will begin with the RNGs seeded the same way every time. In this scenario, the spirit of each game is compromised.

I suppose one could still allow free play on these titles, provided that the player be required to wait for the splash screen to occur before starting.

Dave

Dave, patterning on Pac-Man Plus is possible regardless of whether the game is set to free play. Threads giving details can be found both in the historical TG forums and on the classicarcadegaming.com forums. Here's a starting link: http://www.classicarcadegaming.com/f....html#msg60808

I don't think patterning on Ms Pac-Man is possible for a human even on free play. My understanding based on what I've read David Widel (who added Ms Pac to MAME) write about the Random Number Generator (RNG), at least for the random fruit generation, is that it effectively modifies itself continually based on user input each frame which in practice makes it impossible to pattern. In Pac-Man Plus if you've got the same starting RNG seed and you run a single continuous forward motion pattern for the whole game, it can be fully patterned like Pac-Man. For Ms Pac-Man you'd need to have the same starting RNG seed and also run a pattern that has identical input every single 1/60th of a second for the whole game. So you would have to not only do the same pattern, but you would have to push the joystick the same 1/60th second frame of the game AND release the joystick after the exact same number of 1/60th of a second frames. Which just isn't physically possible for a human. You can also do an experiment in MAME. Create a save state in Ms Pac-Man after starting a game and just before you take control on the first cherry screen. Try to run just the start of a simple pattern that makes a loop around the screen. Reload the save state and try to use the same pattern 10 times. You won't be able to, even though the save state started with the exact same RNG. As a human you just can't duplicate even the beginnings of a pattern with the frame perfect precision Ms Pac-Man would require. Here's a relevant thread: https://www.twingalaxies.com/showthre...S-PACMAN/page8

That said, it's certainly possible from a programming perspective that having the free-play dip switch, or ANY of the seemingly irrelevant dip switches, set to a non-standard value could affect gameplay on a game. However I'm very skeptical that all of those extra non-difficulty/lives related settings have always been strictly enforced in the past. As such I could probably be persuaded to allow them to be "players choice".

-George

The mindset behind the original policy on "free play" had to do with, admittedly, no one knowing for sure whether there was any impact on gameplay/behavior. Thus the policy was precautionary and not due to any particular insight of what was actually happening but rather what "could" be happening.

The same mindset existed for having multiple credits in the game. For example, take a game with a continue feature...imagine if a programmer created a sub-routine that checked for the presence of additional credits inserted and somehow that ramped up the difficulty so as to increase the burn rate for continuation that much faster. Again, it was conceivably programmable so "could" have happened in one or more titles regardless of how prominent or obscure that they were.

I agree with Jace that there was an original thought process behind it, but considering that the policy was created more out of suspicion based on what was conceivably possible rather than what was known for a fact, in that spirit I believe that the policy should be open for discussion now.

Let's wait for what Matt Osborn might have to say on the matter. Surely there is no rush-reason to change this policy...I just think it should be opened for discussion and see where it goes based on more insight and understanding that what the original team that made the decision had between them.