"Preload" uploads

[LN2]

Or simply run a program specifically designed for this purpose.
That was the point of my question.

Well yes, that's different...and certainly if done would be reflected within the recorded speed numbers using wolfmame or alphamame.

There are a few players from the past that were known for doing exactly this...where they actually a special program running that would hog 50% of the CPU or 75% of the CPU so then mame would only have the remainder of the CPU for the games so they would run really slow.

That kind of thing doesn't fool wolfmame as it would show the game is only running at 25% speed...or whatever speed.

The other aspect mentioned above is there are some games where the actual game hardware itself wasn't powerful enough to totally handle the game so the game slows down in spots. That's all within the emulation layer though so doesn't count toward mame speed performance.

Mario Brothers is a good example of this...when you get to later phases where you have 2 fireballs at the start of the phase..and all the animals coming out and icicles etc. things get a little slow...but if you show the mame stats they are steady at 100% even though the game itself is choking some until you can flip and kick off a couple animals.
if you even turn the speed throttle off so the game is running at 200-250% for example, you will still see that slowdown and choking in those spots in the game even though it's actually emulating far beyond 100% speed.

If it were intentional, most of us feel that this would not be enough

Yes, we are all aware of this. If a multi-repeated intentional thing, then yes I agree just to keep zeroing out their scores isn't perhaps enough.

For the most part the player gets a stigma for being a cheater anyway....so editors and confirmers are very quick to notice new uploads from them and are all set to zero out another score from them.

Perhaps that stigma is enough punishment. hehe
There is also the thread that is a hall of shame. Check that out.

Given MARP is open to the public though, it's hard to do anything else. If you deleted their account, then they sign up under a new alias and start all over again...same crap.

It's not really a "solution". This is sort of what alphamame and alphamame blocking was all about...an attempt at a real solution..so you have true confirmation the games were played at speed without pausing etc.

However, as you likely saw(think you had just joined MARP so maybe you missed that) a hacker had hacked alphamame and was going to write a converter so any regular mame inp could be converted to an alphamame inp...thus bypassing the protection and insurances alphamame was providing.

In short, if someone deliberately wants to cheat, nothing is really stopping them even using special checks in MAME like alphamame or wolfmame. You can only make it a tad more tedious for them to cheat....but cheaters are persistent and will cheat anyway.

...as has been proven with many PC games like the quakes etc. with the game patches to allow cheating online.

Half the new stuff in Q3A the past few years was to counteract these cheats...yet almost quicker than id software could fix one, a totally new method of patching to allow cheats was complete. It's the type of war that can't be won.

[Weehawk]

The other aspect mentioned above is there are some games where the actual game hardware itself wasn't powerful enough to totally handle the game so the game slows down in spots. That's all within the emulation layer though so doesn't count toward mame speed performance.

Mario Brothers is a good example of this...when you get to later phases where you have 2 fireballs at the start of the phase..and all the animals coming out and icicles etc. things get a little slow...but if you show the mame stats they are steady at 100% even though the game itself is choking some until you can flip and kick off a couple animals.
if you even turn the speed throttle off so the game is running at 200-250% for example, you will still see that slowdown and choking in those spots in the game even though it's actually emulating far beyond 100% speed.

Please, please, please tell me that this explanation was for the benefit of someone else, because I'm really tired of repeating that I understand this and that it is completely separate from the question at hand.

For the most part the player gets a stigma for being a cheater anyway....so editors and confirmers are very quick to notice new uploads from them and are all set to zero out another score from them.

Perhaps that stigma is enough punishment. hehe
There is also the thread that is a hall of shame. Check that out.

Given MARP is open to the public though, it's hard to do anything else. If you deleted their account, then they sign up under a new alias and start all over again...same crap.

It's not really a "solution". This is sort of what alphamame and alphamame blocking was all about...an attempt at a real solution..so you have true confirmation the games were played at speed without pausing etc.

However, as you likely saw(think you had just joined MARP so maybe you missed that) a hacker had hacked alphamame and was going to write a converter so any regular mame inp could be converted to an alphamame inp...thus bypassing the protection and insurances alphamame was providing.

In short, if someone deliberately wants to cheat, nothing is really stopping them even using special checks in MAME like alphamame or wolfmame. You can only make it a tad more tedious for them to cheat....but cheaters are persistent and will cheat anyway.

I have pointed out why I think all of these arguments are pretty meritless in previous threads. No need to repeat myself here.

[LN2]

rephrase the question so we can figure out what you really want answered then.

I thought it was in regards to if wolfmame would show lack of speed in a game for someone using another program(whether deliberately or not) that results in mame playing the game at slower than 90%.

In a nutshell, as stated above, wolfmame would show this. It would show as MAME shows in the stats the game running at only 75% or whatever speed it's running at. If you launch then launched a background process to hog half the CPU, then the speed of the game in mame would drop also and be reflected in the recording speed using wolfmame.

I think that answers your question...but you seem to be saying stop answering that...so what is your question if not this?

[Weehawk]

SJK reported that this was only on some significant places in the game. What I am asking is if it is reasonable to believe from the places in which it occurs, that this happened naturally as a result of the processing necessary, or does it seem likely that it was orchestrated deliberately to make those parts of the game easier?

I thought it was in regards to if wolfmame would show...

I'm sorry, but I cannot fathom what in that question makes you think that.

rephrase the question

What the hell. Why not?

Starting this subthread, SJK reported that wolfmame showed that, in the recording in question, at certain parts of the game the recorded speed dropped to 80%, but that the average for the whole recording was acceptable.

That is given.

So all discussion since then regarding things that wolfmame would not show (or whether or not it would show) have been irrelevant.

Given that the recorded speed drops to 80% in certain places, two basic possibilities occur to me:

Possibility 1) At these points in the game the processing requirements of MAME are greater than the parts which do not show the slowdown, and therefore in those places only the player's computer is unable to maintain the specified framerate.

Possibility 2) The processing requirements that MAME places on the player's computer are no different in these places, but in order to gain an advantage in the game the player has employed some method (whether playing his favorite classic rock mp3's or running a slowmotion program) to force to game to run more slowly.

My question for any who may have viewed the recording (I was away from home and could not) was whether Possibility 1 or Possibility 2 seemed more likely based on their observations.

If I understand Buttermaker correctly he finally indicated that it was Possibility 1.

[LN2]

Possibility #1 is by far the likely scenario there. I thought we had explained that..but again you weren't happy with it and reasked your question.

When it's stated the speed drops to 80% in spots they are talking only brief periods of time(split-seconds type of thing). If it drops to 80% for a sustained period of gameplay, that isn't acceptable.

To do possibility #2 was where someone potentially would have the CPU hogged up by something else so the typical speed was only 80%, or 50%, or 25%. We aren't talking spikes there but the average.

I seriously doubt anyone would have a toggle setup as your #2 suggests where only for tough spots in the game they would activate it so instead of running at 100% they run at 80% or 85%. If you see a drop to 80% or 85% during gameplay that last beyond a few seconds of gameplay, that isn't acceptable...especially if in other parts it always seems to be at or near 100%. That's a sure sign something started running in the background...whether intentional or not.

The rule isn't where you could play for 20 minutes at 100% then 20 minutes at 85% where the average would still be above 90%. The rule is your gameplay must be above 90% throughout the gameplay. Tiny split-second spikes below 90% are tolerated...but several seconds or minutes below 90% would violate the rule IMHO and should be zeroed out.

In the past gamers have been caught using a CPU hogger so the game in MAME runs at 25%...and remains around 25% the entire game. It's not where you have it running 100% most of the game then slow it down to 25 or 50% for 3 minutes where the overall average still ends up beyond 90%. That isn't acceptable and the score would be zeroed out cuz of speed.

The reported overall average recording speed quickly gets rid of these or those whose CPU just isn't fast enough to have the game run at the frame rate required at whatever frameskip they are using.

It depends on the game of course also IMHO. Any game where time is a factor should have a requirement where it always must be not just 90% but even 95+% or even 98+%.

Take a game like a tile matching game. If you are running at 90% you have 10% more time to do the board than players playing it at 100% have. IMHO that's totally unfair...especially since fast controls aren't an issue here so using a frameskip of 6 or 8 or even higher to have it play at 100% should be enforced.

I personally think the only time speed below 98% should be tolerated is when the player is already using a frameskip of 8 or 10.

If the player played the game with sound and with frameskip 0 and it shows as even 92% overall, that shouldn't be allowed...but that's me. MARP decided otherwise.

MARP decided long before I joined it to just use 90% which is decent as an overall rule. I still think for certain games they could have special rules for that game forcing 95+%. TG has a stricter 95% rule where I think the recording is no good if at ANY time it drops below 95%...but perhaps that's average also.

For many games I'm using frameskip 6 or 8 to play the game at 100% speed.

it would be nice if the frameskip setting could be saved to the inp file if it already isn't....and displayed within inp playback. That way if you see it's running at even 90% yet frameskip 0, I personally would want to zero it and instruct the player to use frameskip 4 or 6.

There is no excuse in having it even run at 90% with frameskip 0.
these games are still plenty playable even at frameskip 6. Beyond fs6 it depends on the game if control is compromised too much to make it as playable anymore. There are many games I can't or won't even play because of this.

[Chad]

The Rule is setup to compare the 90% rule against the averages since it's easy to calculate. A generic INTEGRAL MINIMUM speed rule could be setup (where all recordings must never go below a certain speed) but for a lot of games that's not possible, especially when the slowest parts of the game are mame initializations and cut-scenes that dont involve game play. The 90% average works with these quirks so that recorded game play is probably at an integral minimum.

When a game drops below 90% it could be unintential and inintential background processing , but for the most part it's mame itself trying to compute more processing powere for different parts of the game. The deduction of that will certainly vary from game to game, i don't see centepede using more or less relative cpu power as more bugs crawl on the screen, but other newer games have this side effect all the time.

I'd like to see a new minimum INTEGRAL speed rule setup. But proving recordings that use below speed of minimum integral speed will require the acuser to attach a start and end frame count (which should proly be more than a few seconds) that is DURING game play where an analysis program would calculate the speed between those frames and thus prove slow game play was taken advantage of. Still a complex issue none the less.

[LN2]

oh, I thought that rule applied integrally also.

We need this ASAP then...cuz I just realized you could adjust the frameskip during play quite easily where you maybe were using frameskip 6 to have it play at 100% speed but then in a tough area reduce it to fs0 really quickly and play through that section at like 75%...then toggle it back up to fs6.

As long as it's a game where the playback isn't dependent on the frameskip then it seems you could do this type of thing and still average 90+%....where you maybe were 100% most of the time but then dropped to 50-75% just for the bosses for example.

That can't be allowed. I figured the rule already covered that. From your response I guess not.

For a tile matching game...ok, play the first couple easy levels at 100% speed..then drop to slower speed when necessary to give yourself more time...but keeping your average at 90+%. No way that's right either....yet you say the current rules allow this.

[Chad]

perhaps, and the frameskip trick can be done, and the current 90% rule would "allow" this. But i'll bet 99.999% of speed issues are because of mame or background processing not because of frameskip dethrottling. If a rule is setup it must have a minimum integral speed value and a time frame period 1/3/5/10 seconds (a poll choice?) where the average speed of that period must not be below the minimum. To verify the rule has been broken two frame counts must be supplied (who's distance exceeds the 1/3/5/10 time frame distance), the time frame must be during game play, if any frame between the time is not of game play (level changes, cut-scenes, game initialization) the rule will not have been broken.

[Weehawk]

Possibility #1 is by far the likely scenario there. I thought we had explained

Butter answered my question before you got involved. You spent pages and pages explaining unrelated things I already knew. And if you haven't viewed the recording you couldn't answer the question anyway. Any assumptions made without seeing it I could have made just as easily myself.

you weren't happy with it and reasked your question

After you chimed in, I only reasked the question when you requested me to.

The rule is your gameplay must be above 90% throughout the gameplay

I quoted the rule earlier and that's not exactly what it says. Perhaps it should.

i don't see centepede using more or less relative cpu power as more bugs crawl on the screen

Centipede has a maximum of 14 "bugs" on the screen at a time, and handles them without any loss of speed.

[LN2]

perhaps, and the frameskip trick can be done, and the current 90% rule would "allow" this. But i'll bet 99.999% of speed issues are because of mame or background processing not because of frameskip dethrottling.

perhaps...but I thought that was a big part of why alphamame and it's decendant wolfmame were created...to make sure people aren't pausing or using CPU hogs etc. to play at slow speed.

if you allow adjustment of frameskip to adjust the speed during gameplay...and only care about the overall average speed, that defeats half the point doesn't it?

...as someone that wants to bend or ignore the rules will take advantage of this quite easily for some games.

Perhaps a good feature to add to wolfmame would be that once you start a recording you can NOT change the frameskip! Then that eliminates that convenient possibility to cheat at least.

Are the frameskip F8/F9 presses saved in the replay file? If so then suspect speed replay files could be analyzed for altering the frameskip value also...and zeroed out if it was adjusted. Players should have their frameskip set from a quick test run before playing and recording it.

if those keystrokes are recorded for PC-mames, they aren't for macmame...cuz macmame doesn't even use the F8/F9 but has command keys/menu selections for it instead.

[Chad]

yep, that's not a bad feature to force the frameskip model ONLY when recording of course. But still that would be attacking a less rampant problem (frameskip thottling) other than just mame being slow once in awhile sometimes during gameplay. I think it might not be possible to make a encompasing rule about integral minimum speeds, since you have to pick a period of integration, small periods have problems and large periods have problems, it would be hard to find a good period to choose.

[LN2]

I think it might not be possible to make a encompasing rule about integral minimum speeds, since you have to pick a period of integration, small periods have problems and large periods have problems, it would be hard to find a good period to choose.

Yet another reason why forcing a fixed value of frameskip when recording in wolfmame would be a good thing.

I don't know of many games with a fixed frameskip that would run at 90+% most of the time but just happen to conveniently drop to less than 90+% through all the harder parts of the game.

I think it also depends on the frameskip they used also. if they got 90+% but occasional dips below 90% yet are using only fs0. I don't think that's a gray area at all and should be zeroed cuz they could have easily used fs4 and gotten 100% all the time. Now if the player was alreayd using fs8 or fs10 level of frameskip, then occasional dips under 90% would be more forgiving IMHO....as it would be difficult to tell the player to increase frameskip even more.

If someone has the frameskip or other background stuff running where it's only at 80% in those hard parts, then odds are it's not going to be high enough at the other parts of the game to average out to 90+%.

Also, as mentioned above this becomes really important on puzzle type games. Do you really want to allow someone playing tetris when they get in trouble they can have it setup so going to fs0 makes it run at 75% assisting them to get out of trouble in their game...then set it back up to get 100% when not in trouble?

Again given wolfmame was made to discourage and disallow convenient cheating, the framerate adjustment type cheat is still way too easy to use during gameplay. if you were playing a game where even at fs0 you get 100% speed, you could have some background process running so it's 75% instead...then use fs4 or fs6 for the game except when it gets hard...reducing it back to fs0 at your convenience.

To me that's far worse and deliberate cheating versus someone in the gray area you are talking about where most of it is 90+% and you see an occasionally brief dip below 90% especially if they are already using a fairly high frameskip value. The player has no control when and where in the game those dips occur....unlike adjusting frameskip where they have total control and can take advantage of it.

I think for the integral part it could be fairly easy...like 5+ seconds below 90% during gameplay action isn't acceptable type of thing.

If the speed drops below 90% for only a few seconds, does that really help the player? Odds are it doesn't.

I think in most cases it would be pretty obvious if someone was playing around with the frameskip and speed of the gameplay if you see the recorded speed drop at all the tough spots of the game and only at those spots.

...or is that recorded speed number wolfmame shows in the F11 stats just the overall average and not a running average?

[DRN]

This is all too much for me, I'm off down the pub!

[The TJT]

This is all too much for me, I'm off down the pub!

Thanks a lot Preload, now you made Darren a drunk!

[AL]

Damn....there goes my excuse.....


AL