20-02-2004 03:50									#1
	FA: Gamecatcher Alliance Diplomacy Thread #2 Increase the size of your Attachments

Link to previous thread.

	20-02-2004 08:11									#2
	Remove this text

All I can say is keep the PMing with Lucky alive and TRY to eventually get around to posting it in the UN.

Lucky is falling for the simplest psychological trick in the book: make him "brag" into confession.

Lucky is the kind of guy who cannot admit to failure. He considers himself too smart and capable to have been outplayed by us. This pretty much makes him capable of cheating in order to save his own personal honor.

How to catch him? Make him brag into "confession". The more he is lured into showing all he knows about the RNG, the more he is actually shooting himself in the foot since it will be more than evident that he is QUITE capable both in will and capacity to do the crime he's being accussed of.

Now, we'll never get a "confession" by him, and frankly all his attempt to convince us that he never did anything wrong is utter horse-crap, but if we can get to post these PMs a the UN maybe we can make the other teams so wary of GCA that they'll either be not invited to future ISDG or everyone will gang up against them.

-MZ

	20-02-2004 08:59									#3
	Full PM-box? Change here!

If we're going to post all of his PMs in PUBLIC on the U.N. forum, we're going to need his clearance for that... since he already agreed that I could pass them to any of you guys however I wanted, that shouldn't be hard. I'll ask him about that in the next PM.

I will say this, however... 2 things:

1. He was rather insistent that we include each of his PMs in their entirety rather than selectively picking out pieces. That's going to make for a monster thread if I have to post each of them IN WHOLE (which I'd like to respect... gives him less to complain about).

2. I think we're actually best off and we maintain the air of trying to honestly discern what he did (which is really what I have/am doing) rather than trying to lynch him in public (which I don't quite agree with) if we simply post the evidence and let people come to their own conclusions. We post all of the pertinent PMs by Lucky and we post statements by wervdon and others who have critiqued them. I'm reasonably sure that others may come to similar conclusions as many here have done.

When and if they post such conclusions, Lucky will feel a need to defend himself and will likely continue talking about the issue. He may at that point provide a great deal EVEN MORE things that can be poured over to determine the validity of his claims.

The next point, however is critically important.

Part of this exercise IS, afterall, giving Lucky a chance to defend himself from these charges. If his defense is clearly inadequate (as seems to be the case), then we handle things from there. Lucky, despite whatever you may think of him, does at least deserve that opportunity and deserves that we should look at the evidence with as much of a skeptical and also reasonable eye as we can.

If we do indeed handle this responsibly and make it very clear that we are not out to LYNCH Lucky, but rather to FAIRLY point out what we believe he may or may not have done AND LEAVE LUCKY TO EXPLAIN HIS WAY OUT OF IT (rather than us just trying to drown him out), far more people will believe the weight of what we have exposed. I refuse to take part in lynching-style justice, even if the culprit is deserving. Afterall, anyone who has such a serious charge leveled against them deserves the chance of a fair hearing to defend themselves from the charge.

Besides... Lucky's attempt to defend himself, assuming he is actually guilty, may be far more damning than ANYTHING we could ever say. Remember that.

	20-02-2004 09:55									#4
	Suffering from ads?

I'm very pleased to see Arnelos post this and I completely concur. In my line of work, I see very few people get the benefit of the doubt, or get a fair chance to defend themselves, and it's a constant frustration.

I am somewhat stunned by his admissions thus far, but I do agree that if this matter is to be handled fairly and responsibly, and I am very pleased to see the progress that has been made thus far.

--Togas

	20-02-2004 10:34									#5
	Support Apolyton buy from Amazon

I think we need to milk this PM exchange for all its worth.

He is just DYING to fix his reputation.

Pride is a weakness, use that to our maximum advantage

	20-02-2004 11:23									#6
	Support Apolyton or Terrorists Win

Yes, definitely continue this the way you have already done Arnelos. He does deserve the chance to prove himself innocent, though what he has done so far has only hurt his cause. If he truly is innocent of wrongdoing then it will actually make me feel better about the game, but if we give him the rope to let him defend himself and he uses it to hang himself instead that won't be too bad either.

	20-02-2004 12:00									#7
	Remove this text

I have a gut feeling he's going to use it to hang himself but we'll just see about that...

considering that the whole disbanding issue came up, I'm starting to think what he really wants is to save face and somehow make up all these allegations of cheating before GCA ends up looking like a farse in this and other demogames.

	20-02-2004 12:02									#8
	Tired of ads?

btw, a HUGE thumbs up to Arnelos, wervdon, BigFree and anyone else who has been heavily involved in this issue. I know that ultimately this is kind of a team fetish to try and prove Lucky did in fact cheat, but I really want the truth to come out, not so much for the sake of this demo game since it won't change anything but so that future demo games do not make the same mistakes that we have.

	20-02-2004 12:03									#9
	Support Apolyton or Terrorists Win

IMHO, this needs to reach a resolution before the C3C ISDG begins. Luckily for us its taking forever to start. But his true nature needs to be revealed and either he or his team be banned from participating.

	20-02-2004 17:48									#10
	Inflate your Upload Space

Just a note on GCA disbanding their units.

Do they get to convert them to shields by disbanding within a city ?

	20-02-2004 18:08									#11
	Support Apolyton, buy Civilization: The Boardgame

I disagree with a public lynching before any proof of guilt. Now I'm convinced, but that doesn't mean I can prove it. Its unfair to him to go public with allegations that we are sure of but can't prove IMO.


Now that doesn't mean we shouldn't show Mongoose and any future admins of the game or future games I just don't feel a need to go on a smear campaign against the guy's reputation at this point.

	20-02-2004 18:45									#12
	Support Apolyton buy from Amazon

We need the "Official Spokesman" in the UN as soon as possible, the Germans are starting to backtrack on the unit disbanding issue.

We came a Loooooooong way towards making all these dam exploits "right" yesterday.

The pressure needs to be kept up.

	20-02-2004 18:53									#13
	Suffering from ads?

The consuls wanted to handle it.

And I was just itch'n to post....

	20-02-2004 19:44									#14
	Remove this text

quote: Originally posted by Hot_Enamel The consuls wanted to handle it. And I was just itch'n to post....

Me Too!!!!!!!!!!!!!



It is getting out of control.

"Official spokesman for Apoly, please report to the UN".

	20-02-2004 21:04									#15
	Avatar Enlargement: We've got the solution

Thank you for your restraint.

We need to keep wraps on people posting when they want to. As I said, it leads to disorganization and may give teams impressions other than what we as a team are trying to convey.

	21-02-2004 00:21									#16
	Increase Your PM Length

From Lucky, section 1. Please note for the upcoming posts that in some places I slightly altered wervdon's posts for the sake of Lucky's disposition.

quote: Originally posted by Lucky quote: Originally posted by wervdon Personal opinion as someone with a BS in Comp Sci and a programmer myself whose put some (read considerable) amount of time with his own game related projects: This is a mixture of BS and truth. Its probable that he is telling the truth about using a program that monitors memory changes to observe the change in values at each call of the random # generator that he describes, and it in fact would work very much like he describes. The fact is someone with a basic knowledge of C programming and a little luck guessing at the actual method used by Firaxis to transform a random 1 to RAND_MAX value to a 1 to whatever value (hint: its very often a modulos operation) could very easily take that initial value and the published working of the combat system and make a quick program in C that predicts the result of the next battle with it only necessary to type in the value read from memory, the attack value of the attacker, the defense value of the defender, and any fortification/terrain mods. Do I think Lucky did this? I don't know, but I am sure that given what he has already described doing that it would be *possible* to do it. In any case the rest of what he says is total BS in my opinion. Feel for the random results? give me a break. Its better to attack with horses than swords against mercs on plains? He expects people to either believe: Firaxis took the time to code weird quirks that go against the known values in the combat system involving specific units and streaks......or that he's both supersistous and ......well lucky. Well, I said it before, but I´ll do it again. 1. To my knowledge there is not a single RNG at work for combat calculation. I´m almost entirely sure that things like retreat and upgrading (gaining a level) need to be additionally triggered by another RNG function. The same may hold true for the actual battle as well. 2. The following RN is independent on the last RN. So no, you cannot predict the next random number by knowing the current one. So such a program as described below won´t work. Possibly if you know ALL RNG function, however as said before, I don´t. 3. Concerning the Num Merc example. Well, you can test it yourself. Create a scenario with 1 horseman and 1 swordmen on one side and 1 Numidian Mercenary on plains on the other side. Preserve Random seed needs to be on naturally. Load it, attack with the swordsman, note the exact results (e.g. won with 1 HP left and gained a level), reload the scenario, attack with a horsemen and witness the very same outcome. I have such a scenario if you want. This is only true for a single battle of course, meaning whenever you have the choice to attack with a horseman or a swordmen against a Merc on plains, the outcome will be the same, except that the horsemen might retreat. 4. Well, about the experience, not sure but if you ask the people that did one of those 1000 attacks scenarios, I´m pretty sure they will tell that they often have the correct inclination about what will happen next after seeing a few battle results in a turn. If you think that´s bullshit, not my problem. Maybe it´s just my imagination, but I´m perfectly sure that I can better predict and "read" civ battles after I did the majority of those tests.

	21-02-2004 00:25									#17
	Enter the AD-FREE zone

From Lucky, Section #2.

quote: Originally posted by Lucky quote: Originally posted by wervdon Take it one step further: Assume that the value he is seeing changes from 1 to 20 mil + is the random seed (ie the last random value generated by the rand() function), and I'm almost sure it is. Then assume he (or someone on their team or that he knows) is at least as familiar with C as I am. All he'd need to do is this (literally): #include #include int main(void) { long value; int cnt; printf("Enter seed value: "); scanf("%d", &value); srand(value); printf("Next 100 results: \n"); for (cnt = 0; cnt < 100; cnt++) { printf("%ld\n", rand()); } return 0; } Piece of cake, and I'm almost sure it'd work if what he is seeing is the last random value generated. Edit: Just to be clear, its that easy to get the next set of results from the rand() function. It's a bit trickier to figure out how those 1 to RAND_MAX values are going to be used to generate a winner in a unit battle. But if you're motivated and determined, it shouldn't be too hard to play around with various algorithms until you can reliably predict results with a fairly simple program. No, as said above this is not possible. The RNG is a little more complicated than just using the number generated by a rand fuction. According to what I saw, the current number has no influence on the next number to come, meaning that e.g. after 100000 it will generate another random number each time. I would have to take a look at that again though to make a definite statement. But as already said as well, even when knowing the RN, you cannot predict the outcome of a battle. Meaning you note the RN one time, look at the results and then input it again before an exactly similar battle, it won´t generate the same results.

	21-02-2004 00:28									#18
	Support Apolyton, buy Alpha Centauri

From Lucky, Section #3.

quote: Originally posted by Lucky quote: Originally posted by wervdon Few points: 1) He (Lucky) said in one of the PM's that he has not looked that Random Seed in the ISDG but that we'd have to take his word for it. 2) He (Lucky) said he's not hacking the save to get the seed, instead he's using a program that monitors the memory space allocated to a running program and that looks for changing values. In effect he's hacking the running exe to get his data. 3) The thought that you can't get any future results out of the starting seed value (beyond the next combat round) is BS. Anyone who knows enough to extract a random seed in the manner he's describing should know that. I've already pointed out pretty well above how easy it would be to go from 1 random result to as many as you want just by writing a 10-15 line C program. Also anyone who thinks that the seed stored by rand() and srand() wouldn't be in civ3's allocated memory (it is a call to a library function), should keep in mind that preserve random seed likely stores a copy of the value in a global variable somewhere in the program itself for the purpose of saving it somewhere in the save file. Which also means if you know the last value generated, it would be pretty trivial to find it in the save and then hex edit it to a value that got you good results (say great leaders) in the past. Of course, Mongoose couldn't then reproduce your results, unless of course he used a hex-edited save that you sent him yourself. So this can probably be ruled out, at least. #1. Yes, unless you have a way to proove it! However, would I really be telling you all this if that would "greaten the mess" or whatever for me? I think not. And again, knowing the actual RN gives you no further knowledge about anything. #2. Yes, that´s what I did when conducting my tests. Last time during the C3C BETA, however I might be forced to do it again to verify some results now. #3. And I pointed out that your statement is BS (using your words). To my knowledge this is simply not possible by just knowning the current RN. And no I haven´t bothered writing any program as you described, but thanks for the idea. #4. And I have also not bothered looking for the values in the save. I think I read somewhere (CFC IIRC) that in the uncompressed save several of those that picked it apart noted a value or multiple values they consider the RN. Haven´t followed that though.

	21-02-2004 00:28									#19
	Browse Apolyton AD-FREE

Send him my tests in which I DO use 1000 units of each to test this. What a moron he is if he thinks we will buy this load he's trying to sell.

	21-02-2004 00:29									#20
	Support Apolyton or Terrorists Win

From Lucky, Section #4.

quote: Originally posted by Lucky quote: Originally posted by wervdon The only way I could be sure of how Firaxis coded the game is to reverse engineer the executable or ask them. I've done neither, nor am I interested in doing the former. What I can say is that I would be surprised if they (firaxis) went through the trouble of implimenting their own random # generator when one is available as part of the ansi standard. Also you couldn't just look at the # and say this value means a horse wins! What you would need to be able to do is figure out how firaxis took the results of a rand() call and transformed it into a dice roll. Generally speaking this is done via a modolus (sp?) operator, but there's other methods as well. If one was determined though, I don't think that much extra effort would really be required. Besides, if you're actually modifying the seed and not using it to predict, then you don't even need to do more than just record "good" seeds from trial runs that result in 20 victories in a row or great leaders (you'd gain this knowledge over time) and then before you start hex edit the save to start on that seed. If you did that though, Mongoose would have trouble reproducing your results if he had a "clean" save from before you got a hold of it. It´s definitely not a single rand function they implemented. Hmm, just what I was saying, you are contradicting yourself. I´ve stated again and again the the current RN means nothing. Since I cannot disassemble the game code I cannot find out how Firaxis is using the RN to generate battle results. Also I was always satisfied by finding certain similarities and general assessments that showed how certain aspects of battling worked and thus improved my experience about that. As I said myself, any form of manipulating or editing anything would immediatly conflict with any results mongoose would get.

	21-02-2004 00:35									#21
	Full PM-box? Change here!

As for Lucky's response to ME:

quote: Originally posted by Lucky I would prefer if the actual discussion would be kept as private as possible. At least for now. However, I was hoping that if I could at least partially convince a majority of the serious people involved in this, that you would have the integrity to spread that the doubts outweigh the suspicions the same way your spread the accusations before. Lucky

That's ok, if we come to the determination that we still believe he did something wrong, we need only share it with Mongoose and keep all of this information around for when and if GCA uses whatever Lucky does again in this game (same in the C3C game). The first time it happens, you tell the admin about much of this as the basis of just HOW you believe Lucky would be doing something to obtain extreme combat results. A reasonably admin will have to take that even more seriously than Mongoose took my or another's probability analysis (though we might claim that that could speak for itself).

	21-02-2004 00:37									#22
	Support Apolyton, buy Civilization III: Complete

quote: Originally posted by BigFree Send him my tests in which I DO use 1000 units of each to test this. What a moron he is if he thinks we will buy this load he's trying to sell.

I think what he's saying is to do it with preserve random seed enabled. Then start that scen a zillion different times, reloading for the horseman and swordsman to see how they perform the same or differently with the SAME RN values.

At least that's my assessment of it.

Could still be blowing smoke, though. This can be tested.

	21-02-2004 00:41									#23
	Browse Apolyton AD-FREE

quote: Originally posted by Arnelos I think what he's saying is to do it with preserve random seed enabled. Then start that scen a zillion different times, reloading for the horseman and swordsman to see how they perform the same or differently with the SAME RN values. At least that's my assessment of it. Could still be blowing smoke, though. This can be tested.

Well, duh. It's very easy to understand where there is a certain RN that would give the same results for a Swordsmen against a NM as it would a Horsemen against a NM. He says they RNG has 2 million of them. There's bound to be more that a few that would give the exact same results.

The real test is to not test it with the same RN; but to test propencity of one units ability to win over a period of combats versus nother. Which is what I did test. Lucky may be smart, but he's counting on us to be stupid if he thinks I will buy this which means he's really dumb. .

	21-02-2004 01:12									#24
	Support Apolyton or Terrorists Win

Put another way, if Horseman versus Pikeman on Plains always gives the same result as Swordsman versus Pikeman independently of the current random number, then what accounts for the statistical significance in BigFree's test?

I must be missing something here...

Perhaps the trials are not independent?


Dominae

	21-02-2004 01:35									#25
	Suffering from ads?

BF: Were your tests run in the same scenario with the same seed?

That is, did you have a stack of 1000 Horses and 1000 Swords next to each other attacking the same stack of Spears in the same scenario?

	21-02-2004 01:48									#26
	Support Apolyton, buy Call to Power 2

quote: Originally posted by Trip BF: Were your tests run in the same scenario with the same seed? That is, did you have a stack of 1000 Horses and 1000 Swords next to each other attacking the same stack of Spears in the same scenario?

There were 10 runs with 100 of each type of combat for a total of 1,000 for each combat type. I just reloaded the scenario 10 times.

The resluts and the scenario are posted.

	21-02-2004 01:52									#27
	Avatar Enlargement: We've got the solution

Here it is again:

quote: Originally posted by BigFree I have buit a SIM for all to test. I have not run it yet, but feel free to do so and post results. Thanks. http://apolyton.net/upload/view.php...nameto.bix).zip

Here's some results with a 1,000 trials for each type of combat.

code:

Tests

100 horse against 100 NM


#1  48 Horse left versus 83 NM left
#2  61 Horse left versus 78 NM left
#3  61 Horse left versus 68 NM left
#4  64 Horse left versus 74 NM left
#5  61 Horse left versus 75 NM left
#6  51 Horse left versus 83 NM left
#7  60 Horse left versus 80 NM left
#8  54 Horse left versus 92 NM left
#9  61 Horse left versus 74 NM left
#10 48 Horse left versus 77 NM left



100 Swords against 100 NM


#1  49 Swords left versus 53 NM left
#2  45 Swords left versus 56 NM left
#3  38 Swords left versus 63 NM left
#4  58 Swords left versus 42 NM left  
#5  51 Swords left versus 52 NM left
#6  49 Swords left versus 42 NM left
#7  46 Swords left versus 54 NM left
#8  44 Swords left versus 56 NM left
#9  48 Swords left versus 52 NM left
#10 34 Swords left versus 66 NM left  

	21-02-2004 03:08									#28
	Help yourself to an AD-FREE life

quote: Concerning the Num Merc example. Well, you can test it yourself. Create a scenario with 1 horseman and 1 swordmen on one side and 1 Numidian Mercenary on plains on the other side. Preserve Random seed needs to be on naturally. Load it, attack with the swordsman, note the exact results (e.g. won with 1 HP left and gained a level), reload the scenario, attack with a horsemen and witness the very same outcome. I have such a scenario if you want.

I tried this out yesterday. I actually played a game though and waited for a situations like this to arrive to test it. I do have a save file where a horse and a sword attacking a NM on plains gives a different outcome. However from the dozen or so instances where this situation was possible more than half did result in exactly the same outcome from attacking with the different units.

Although saying they are always the same is wrong and can be proved to be wrong.

	21-02-2004 03:18									#29
	Support Apolyton, buy Galactic Civilizations: Deluxe Edition

Well, I honestly didn't realize you'd quote me word for word, but I guess I'm ok with it.

In any case, I don't have time to read all of his responses, and I've pretty much given all the technical information I can on the subject, so anything further I state would have to be mostly opinion.

So I hope no one takes it as bad form if I don't try to respond to his responses to the quotes you showed him.

	21-02-2004 06:37									#30
	Got spare money?

I was playing a SP game at a friends house earlier today and I had a sword and an archer able to attack an enemy spear. Remembering this little debate, I attacked with one, then reloaded and attacked with the other to see if the results were different.

And yes they were. Both of my units lost but the sword suffered less damage than the horse.


Also a little anecdote. Shortly after the Beta, me and Lucky were chatting in ICQ about a bug he found with armies. He passed me the save and I noticed that his game was with random seed OFF, in other words, that each time you reloaded you'd get a different result (like Civ2).

Now, for someone so interested in the RNG, I find it odd that you'd decide to play Civ3 without the seed. In fact, I find it odd that you'd play Civ3 without the seed period. 99% of everyone I know plays with the seed on.

I say this is all one big stinky pile of poo.

-MZ