 |
| |
|
|
MrWhereItsAt
|
 |
New Year's Resolution: 2005 is the year of Where It's At - come get some.
Nov 2001
time: 17:31
|
|
Here is a list of all the losses we have incurred in the war to date, with the terrain they were on at the time, and probabilities of GCA victory.
Attacking losses:
2 Vet MIs - Hawk and Grog (Turn 1)
1 Vet Horse - Trip's Guard (turn 2)
Defending losses:
1 Vet NM (full 4hp) - Numidian Master (GCA turn 2) - attacked and killed by a Horseman who then retreated.
Was on grassland in a large stack. Odds: 15% if Reg Horse, 24.8% if Vet Horse, 35.5% if Elite Horse. [EDIT by E_T] No Catupults were present[/EDIT]
1 Vet Horseman (2hp only) - NuclearHorseman (GCA turn 3) - attacked and killed by a Horseman, who then retreated towards IDG or Algalord. Was on hills. Odds: 78.7% if a Reg Horse, 89% if a Vet Horse, 94.5% if an Elite Horse.
1 Vet Sword (full 4hp) - jagjef (GCA turn 3) - attacked and killed by an Elite Horseman who lost 0-2 hp,
that generated a leader. Was on grass. May have been unsuccessful Horse attacks before the winning battle. Odds: 58.3%, higher if there was damage done by an earlier attack.
1 Reg Sword (full 3hp) - Flash (GCA turn 3) - attacked and killed by an Elite Horseman who lost 0-1 hp. Was on a Mountain. Unsure if there were other attacks. Odds: 43%. More if there were earlier unsuccessful attacks.
1 vet NM (full 4hp) - Y (GCA turn 5) - attacked and killed by a Horseman, who then retreated to Algalord. Was on a hill. NO other attacks. Odds: 13.6% if Vet Horse, 20.8% if Elite Horse. [EDIT by E_T] Catapults were present.[/edit]
These calculations were made with CFC's Combat Calculator http://www.civfanatics.com/civ3combatcalc.html
and may not take into account retreat odds. Can someone who knows try to sort this out please? I have
done all the hard work, now someone else can bloody well fix my stuff-ups 
This is worth updating turn-to-turn as well, so we can keep track of exactly how "lucky" GCA may be.
Last edited by E_T on 09-10-2003 at 20:29
|
|
| |
|
|
|
MrWhereItsAt
|
|
New Year's Resolution: 2005 is the year of Where It's At - come get some.
Nov 2001
time: 17:31
|
|
The odds of the GCA attack succeeding. In many cases the health of the attacker is unknown afterwards (retreating out of view), and in any case, the chance of a victory of any kind is IMO the major element. In some cases it is clear that the attacker has been left with some damage, and GCA is hardly likely to care all that much about how much damage they incur as long as they win and can retreat out of harm's way.
We may end up in a position to show that GCA is attacking with a tiny chance of success. I hardly feel it is relevant whether they win with 0 or 3 hp loss - the point is they win, and how improbable this is. Besides, to show that that specific amount of hp loss is unlikely is not likely to be taken very seriously - ANY individual combination of hp left for attacker and defender is unlikely compared to 100% - we want to show what they are getting (a victory) is suspicious enough.
I haven't explained that very well.  I'll try an example.
Let's look at their attack on jagjef. The attack was likely to succeed overall (58.3%), so this victory in itself is not suspicious at all, and I can see us trying for the same odds. However, the chance for the Horse to win with exactly 2hp loss (as it did) is only 14.1%. If they were aiming for exactly this result, then it is improbable the rng would give them this without some.... coaxing. BUT, since they could have tried until they got 0 or only 1hp damage, I don't believe they are trying to get as overwhelming a victory out of every battle as possible, and we could not prove as such unless there are lots of battles soon where they lose no or only one hp. Of course they COULD be trying for this and 2hp damage was the best they could do with all the times they reloaded, but I challenge anyone to prove that much to me, let alone the UN.
But, I guess we should note the places they win with poor odds and lose 0 or 1hp in return.
|
|
| |
|
|
|
Dominae
|
|
Thanks for posting a thread on this, MrWhereItsAt.
I think we should only consider Merc. losses in any calculations to determine of GCA is cheating, since 1) we know GCA cannot be ganging up on those because they're in our big stack, and 2) it's strategically strange enough to just attack those (and never lose) that this is where our "proof" will lie.
Could you please update the original post with mention of which Mercs. were in stacks with Catapults? Obviously that factors into the odds.
Incidentally, I'm very surprised that the/my "combat report" suggestion was never adopted at the UN. This would make the game a lot more fun, IMO, and would give us more information to work with against would-be cheaters (GCA's results may seem a lot more plausible - or not - if we knew exactly what was going on).
Dominae
|
|
| |
|
|
|
E_T
|
 |
Orlando, Florida
Mar 2001
time: 00:31
|
|
I'm geting these figures from Zachriel's Battle CIVULATION, where you can have it set to figure Retreats.
quote:
Originally posted by MrWhereItsAt
Here is a list of all the losses we have incurred in the war to date, with the terrain they were on at the time, and probabilities of GCA victory.
Defending losses:
1 Vet NM (full 4hp) - Numidian Master (GCA turn 2) - attacked and killed by a Horseman who then retreated.
Was on grassland in a large stack. Odds: 15% if Reg Horse, 24.8% if Vet Horse, 35.5% if Elite Horse. [EDIT by E_T] No Catupults were present[/EDIT]
|
With an Elite Horse: They win 32.51% of the time with an average of 1.66 HP left. In 10000 trials, with 1 units, the Attacker had 0 promotions; while among the 1 Defenders, during 10000 trials, there were 65 promotions, all Elites. An average of 0.43 Attackers retreated. Attacker got a Great Leader! (194)
With a Vet Horse: They win 21.32% of the time with an average of 1.37 HP left. In 10000 trials, with 1 units, the Attacker had 131 promotions, all Elites; while among the 1 Defenders, during 10000 trials, there were 88 promotions, all Elites. An average of 0.47 Attackers retreated.
As we had later seen a 3/5 Elite Horse, we can conjecture that they attacked with him and with not too bad Odds.
quote:
1 Vet Horseman (2hp only) - NuclearHorseman (GCA turn 3) - attacked and killed by a Horseman, who then retreated towards IDG or Algalord. Was on hills. Odds: 78.7% if a Reg Horse, 89% if a Vet Horse, 94.5% if an Elite Horse.
|
Horse against Horse, no Retreat possible. If they used that same 3/5 Elite Horse, it was still a possibility of a GL. Again, Highly likely.
With a 3/5 Elite Horse: They win 78.92% of the time with an average of 2.24 HP left. Regular Possibility of a GL.... We had seen it afterwards, again with 3/5 HP. Again, not too terrible of odds.
quote:
1 Vet Sword (full 4hp) - jagjef (GCA turn 3) - attacked and killed by an Elite Horseman who lost 0-2 hp,
that generated a leader. Was on grass. May have been unsuccessful Horse attacks before the winning battle. Odds: 58.3%, higher if there was damage done by an earlier attack.
|
With an Elite Horse: They win 53.76% of the time with an average of 2.18 HP left. In 10000 trials, with 1 units, the Attacker had 0 promotions; while among the 1 Defenders, during 10000 trials, there were 66 promotions, all Elites. An average of 0.29 Attackers retreated. Attacker got a Great Leader! (334)
quote:
1 Reg Sword (full 3hp) - Flash (GCA turn 3) - attacked and killed by an Elite Horseman who lost 0-1 hp. Was on a Mountain. Unsure if there were other attacks. Odds: 43%. More if there were earlier unsuccessful attacks.
|
With an Elite Horse: They win 41.36% of the time with an average of 1.95 HP left. In 10000 trials, with 1 units, the Attacker had 0 promotions; while among the 1 Defenders, during 10000 trials, there were 81 promotions. An average of 0.32 Attackers retreated. Attacker got a Great Leader! (270)
Again, not too terrilbe of odds.
quote:
1 vet NM (full 4hp) - Y (GCA turn 5) - attacked and killed by a Horseman, who then retreated to Algalord. Was on a hill. NO other attacks. Odds: 13.6% if Vet Horse, 20.8% if Elite Horse. [EDIT by E_T] Catapults were present.[/edit]
|
I don't know what the % of hitting for a Cat in defencive mode, while not forted, so we will look at 3 runs, with Elite (not hit by cat), Vet (either Vet not hit by Cat or Elite that was hit by Cat) and Reg (Vet hit by Cat).
Elite Horseman: They win 18.56% of the time with an average of 1.36 HP left. In 10000 trials, with 1 units, the Attacker had 0 promotions; while among the 1 Defenders, during 10000 trials, there were 101 promotions, all Elites. An average of 0.52 Attackers retreated. Attacker got a Great Leader! (102).
Vet Horseman: They win 11.97% of the time with an average of 1.2 HP left. In 10000 trials, with 1 units, the Attacker had 68 promotions, all Elites; while among the 1 Defenders, during 10000 trials, there were 115 promotions, all Elites. An average of 0.53 Attackers retreated.
Reg Horseman: They win 6.24% of the time with an average of 1.08 HP left. In 10000 trials, with 1 units, the Attacker had 40 promotions; while among the 1 Defenders, during 10000 trials, there were 139 promotions, all Elites. An average of 0.52 Attackers retreated.
The Main thing that I can see and from what I have seen from past Info from them on their battles with CDG, they attack if there is a better for them to still be a live, and that retreat capability gives them that advantage, while in the feild and counterattacking.
Even if they wear us down some, we still have a large force of NM's that are just now coming into the theater. But Also, they will weaken themselves in the process and not be able to feild units as fast.
We know that they have made a couple of mistakes, with the road razing, they should have done one more but they didn't. They attacked us instead. Lets hope they make a few more mistakes like that.
E_T
|
|
| |
|
|
|
Dominae
|
|
quote:
Originally posted by Arrian
Too bad there is no "eye" in this game.
|
There is an eye, but he/they are lying about not doctoring the map. What's the good in an arbiter you do not trust?
The whole fact that you have to petition the UN to get anything done in the middle of the game is silly to me. We should have agreed on a very clear set of rules right from the start (including what's considered cheating, which exploits are alowable, etc.).
Dominae
|
|
| |
|
| |
|
|
|
asleepathewheel
|
|
listening too long to one song
Mar 2002
time: 00:31
|
|
Great stuff
Whatever happened to Zachriel, he had some great posts, insights, but I haven't seen him around for quite some time.
I seem to remember reading that Mongoose repeated the attacks by GCA against the immortal stack and that it was done without cheating. Obviously it would be easy enough to reload and come up with a sequence, especially easy in our case as its one unit attacking, but I would hope in a large scale attack, Mongoose would notice and reveal if they had a ****ed up sequence like-attack, move worker, attack, attack, move worker, attack, settle. One would think that odd at the very least.
But it would be very easy to hide one attack/kill.
Do we know anything about any of the GCA players? It would probably be quite easy for Lucky to even hide this from his team if they are loosely playing it.
|
|
| |
|
|
|
Dominae
|
|
quote:
Originally posted by Arrian
Who is this "eye?"
|
Mongoose, the person who set up the map.
Dominae
|
|
| |
|
|
|
wervdon
|
|
Moving workers would have NO effect on the outcome of a series of battles. The only moves that would need to be in sequence or moves that have a random factor to their outcome.
ie)
Move worker
Attack Merc1 with Horse1
Attack Merc2 with Horse2
would be the same outcome as:
Attack Merc1 with Horse1
Move worker
Attack Merc2 with Horse2
I don't know if any actions other than Attack and bombardment use the random # generator though. I suspect that in single player at least that initiating diplomacy with an AI also would.
I agree though the more mercs they pick off out of our stacks guarded by mercs without triggering our golden age (meaning we know they didn't lose a unit) the lower and lower the odds of go.
I can believe that they might be desperate enough to take *bad* moves out of desperation (such as attacking a merc with only a 20% chance of success). I can't believe that they can do it every chance they get on and on and never get unlucky once.
So what we need to do:
Every time we lose a merc, make a note of it. If we don't know which unit killed it, we should probally assume its an elite horse that escaped with 1 pip. Horse because they are the only units with the movement to get away, and elite and 1 pip because that makes all the error in their favor and not ours.
|
|
| |
|
| |
|
| |
|
|
|
WhiteBandit
|
 |
Southern California
Jul 2002
time: 21:31
|
|
quote:
Originally posted by asleepathewheel
Great stuff
Whatever happened to Zachriel, he had some great posts, insights, but I haven't seen him around for quite some time.
I seem to remember reading that Mongoose repeated the attacks by GCA against the immortal stack and that it was done without cheating. Obviously it would be easy enough to reload and come up with a sequence, especially easy in our case as its one unit attacking, but I would hope in a large scale attack, Mongoose would notice and reveal if they had a ****ed up sequence like-attack, move worker, attack, attack, move worker, attack, settle. One would think that odd at the very least.
But it would be very easy to hide one attack/kill.
Do we know anything about any of the GCA players? It would probably be quite easy for Lucky to even hide this from his team if they are loosely playing it. |
The thing is, in order to get the same exact results that GCA had gotten, Mongoose would need to know EVERYTHING that GCA did that turn.
If they are writing everything down, they just send Mongoose all the actions they did that resulted them winning the turn (moving worker at this time, moving settler at this time, changing production here, THEN attacking). Then Mongoose will just replay the save and get the exact same results that GCA did. Surprise! Obviously they aren't cheating then...
As long as Mongoose plays the moves exactly the same as GCA did their turn, there will be no discrepancy. They could have reloaded a thousand times, and it wouldn't matter.
Mongoose wouldn't be able to accurately replay GCA's results *without* having all of this information anyway. So of course it would make sense that Mongoose could "verify" that they aren't cheating. He's making the exact same moves that GCA played (and he can't use anything different).
I.E., to further clarify this example, look at the problems Vondrack has when he tries to replay the save to write the turn reports. If one small thing is out of order (such as BF writing down he changed a city production instead of moving a work, rather than vice-versa as actually would have happened), notice that Vondrack gets different results. Taking out a barb costs 2 pips instead of the 1 that was actually played, or the warrior attacking actually dies instead of winning...
The point of all this is, in order for Mongoose to accurately sim GCA's moves to "make sure" he is getting the same results, he has to have GCA's entire turn list.
I think the actual thing that we need is something along the lines of what CFC uses. Apparently for their GOTM, they can figure out exactly how many times a save has been loaded. Perhaps we should open up secret dialog with Octavian/Chieftess and the gang about this.
If Mongoose can get it and track the number of loads each time the save is passed on, you can quickly figure it out.
We need to keep this a SECRET though as this would be easily able to nerf as well. Just create a copy of the save, reload a million times to get what you want... record results, open up copied save and replay the right results there...
Perhaps GCA is already doing this. We may never know, at the moment though, way to go for continuing to compile statistics! It is somewhat suspicious of their luck.
|
|
| |
|
| |
|
|
|
asleepathewheel
|
|
listening too long to one song
Mar 2002
time: 00:31
|
|
quote:
Originally posted by BigFurryMonster
True, but if this involves:
- Open the save
- Change all production to Forbidden Palace
- Change production in capital to warrior
- Fortify Swordsman
- Change production in capital to spearman
- ... (insert silly production/unit-movement action here) ...
- Attack Poly
...it'll be VERY suspicious. |
That's what I was talking about. Some odd sequence of moves would be a definate tip off. Now if Mongoose would reveal that...is another question.
|
|
| |
|
|
|
vondrack
|
 |
Praha, Czech Republic
Jan 2002
time: 06:31
|
|
quote:
Originally posted by WhiteBandit
I think the actual thing that we need is something along the lines of what CFC uses. Apparently for their GOTM, they can figure out exactly how many times a save has been loaded. Perhaps we should open up secret dialog with Octavian/Chieftess and the gang about this.
If Mongoose can get it and track the number of loads each time the save is passed on, you can quickly figure it out. |
I find this highly unlikely - there is, I believe, no way to tell how many times a save has been (re)loaded. Just try it - make a copy of a freshly received incoming save. Load the original incoming save, play the turn, end it without saving the outgoing save. Binary compare the "backup" copy and the save just played - you will find out they are identical. Means the incoming save is not altered in any way while being loaded. Means there is no way to tell if it's the first time it's being loaded or hundredth time.
I agree that Mongoose replaying the turn exactly the way GCA played it proves nothing. It cannot not prove cheating, it cannot prove the contrary. Reloading and replaying, choosing the "optimal" results prior to telling Mongoose the exact sequence would be the cheating we suspect here.
I have always been a vocal opponent of the cheating theory, but I have to admit I am beginning to hesitate a bit. The fact they killed our merc on a hill is, well, a fact - it did happen, we can see it. The fact they tried it in the first place, is a different thing, though.
I still say we wait for more odd events... but let's keep our eyes open. This thread is a very good idea.
One thing I find a bit odd is GCA does not have their units named... (or do they? I cannot recall a single named unit of them, not even the GL-generators). I am not really sure how they plan/discuss their moves accurately... Imagine planning our military moves without having our units named...
|
|
| |
|
| |
|
| |
All times are GMT. The time now is 05:31.
Apolyton Time is 00:31. |
top of page
|
|
|
Forum Rules:
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
|
HTML code is ON
vB code is ON
Smilies are ON
[IMG] code is ON
|
|
|
|
|
|
09-10-2003 14:43
.
|