Possible exploit?

[Kaleb]

It will kick up a right stinker if I post this on the main forums so I thought I'd air my suspicions here first.

I hadn't realised that new andom seed on reload is not turned on here (which is a big mistake!). This means it is possible for the RNG system to be exploited and it's actually quite simple.

You just need to load the save from the game in single player and then play out the combat. So effectively you know what the result of a combat will be before you do it in MP.

If it's just a one off combat (e.g. a Trireme vs a Trireme or an slightly wounded Axe vs a Chariot) try the combat out in SP. If it works, do it, if not - don't attack.

If it's a larger combat, and you have several units to attack with - especially some cheaper throw-away units like archers or Chariots - then when you know you have a bad roll coming up, attack with your expendable units so they take the hit.

When you have a good roll coming up, try to time it so you can take down a big unit.

e.g. in this last round of combat SANCTA attacked our full strenght WE with a 1.5(ish) strength Chariot (who withdrew). Now why would you launch an attack like that unless you knew you had a bad roll coming up (and you knew the chariot would withdraw!) and needed to get rid of it?

(I'll post full combat log up tomorrow, as usual they've killed a load of our units and only lost 2 of theirs, winning all combats including one of 25%).

Of course we cannot prove any of this. But these good combat outcomes just seem to be keep happening and the more they happen the less the liklihood of it all just being down to luck and the more suspicious I am becoming...!

 

[donsig]

Well, like you said, there's no way to prove anyone has done this. If we bring it up publicly we look like sore losers and we teach others (who may not have figured this out) how to do it. Doesn't the MTDG ruleset cover this sort of exploit in a general way?

How do you open a pitboss game in SP? Do you have to save it to your own computer somehow? If so, is there any way the host would see the game being saved?

EDIT: Oh, and is it really so odd that they attack with a chariot first? In my SP games I used suicide attacks alot sacrificing units in order to weaken enemy units so I can kill them off.

 

[Sommerswerd]

If SANCTA has an exploit that we can not beat and can not prove, then we should not be warring with them anymore... at least until we have either
1. determined a way to neutralize the exploit (with larger armies or by some other means) or;
2. Proven the existance of the exploit / proven SANCTA uses it, and gotten the exploit banned by the UN

If SANCTA is using the exploit and we cant stop them, then it seems that we can't beat them in a war, so it is foolish to continue the war... right?

On the other hand, if SANCTA is not using any exploit, and they are just outplaying us, then we need to figure out a better plan to fight them (since our current plan is not working).

Either way, we need a new plan.

 

[Sullla]

Kaleb, I understand that you guys are hurting right now over the combat results, and I can sympathize with that. However, I can't say that I've seen any evidence of foul play at work on SANCTA's side. In their forum, they have screenshots showing (with the clock displayed for timestamp purposes) that they made the exact same moves both before and after the server reload. I'm not convinced that they did anything differently the second time around.

Now SANCTA's first attack was just plain lucky, but this one was entirely expected. Let me walk through the course of actions that transpired:

Chariot vs. War Elephant (odds less than 0.1%): result was a retreat with no damage inflicted. I suppose you could claim that this was an attempt to reroll a random seed, but it seems far more likely that SANCTA was trying to damage the war elephant with a throwaway unit. (How useful are chariots in the post-axe age?) SANCTA got no benefit from this attack whatsoever.

Longbow vs. War Elephant (13.5% odds): result was a dead longbow and the elephant reduced to 2.9/8 health. This is pretty much exactly what you would expect, the longbow dying and doing some damage to the elephant in the process.

Chariot vs. Maceman (odds less than 0.1%): result was a dead chariot, again doing no damage whatsoever. SANCTA might as well have deleted this unit, because it did nothing but give your mace XP. Again, nothing particularly strange or unusual to this point in time.

Longbow vs. Maceman (26.7% odds): result was a SANCTA win. Here is the only combat result that really broke in their favor, winning at fairly low odds. And yet this is one place where the combat "odds" don't show up properly; despite the much higher base strength on the mace, the mace has to win 5 combat rounds to kill the longbow (22 dam each round) while the longbow has to win 6 combat round (17 dam each). Thus the longbow actually has better odds to win than the number suggests; it's actually more like 40-45% because of the way Civ4's combat system works. You can try simming this out, if you're curious. So SANCTA's result was somewhat unlikely, but hardly out of the range of possibility.

Besides, I don't see any evidence of conspiracy here. SANCTA was doing the same thing as the longbow/war elephant battle, seeking to damage the top unit so they could kill it with another unit later. They just got a bit lucky and won the combat entirely instead of crippling the Kaz maceman. Onwards:

Spearman vs. Horse Archer (73% odds): I think we can all agree that this was to be expected.

Longbow vs. crippled War Elephant (99% odds): ditto

Axeman vs. Maceman (20.5% odds): SANCTA loses the axe vs. mace battle, as expected. But they damage the mace into the redline area, which allows them to follow up and kill it:

Axeman vs. Maceman (99% odds): SANCTA kills the mace.

So honestly, what was so unexpected about those results? They attempted to sacrifice two chariots (which both failed), sacrificed a longbow to kill the war elephant, sacrificed an axe to kill a mace, and one of their three sacrificial units got lucky and won outright. Hardly beyond the pale. The real problem was that the Kaz units were split up into multiple stacks, which meant that your injured units couldn't cover one another and prevent SANCTA from getting favorable matchups, and in a part of the map with no defensive cover.

Kaleb, are you absolutely sure that "Preserve Random Seed" is turned off for this game? I created the original Worldbuilder save file, and I made sure that the default option (where reloading doesn't change combat results) was in place. I'm a little skeptical of that claim, because I don't see how SANCTA could have reproduced the identical combat results twice if Preserve Random Seed wasn't in place. Not to mention the fact that I made sure that it was set when I created the game at the start!

But something could have been changed in the setup phase too, so who knows, maybe I'm wrong.

 

[Indiansmoke]

Kaleb a pitboss game cannot be loaded in single player and to load in in multplayer you would need the game admin password. So unless Sancta have the admin password I cannot see how they could load the game and try out the battles.

@ Sulla..there is no option preserve random seed! The option is random seed on reload!

 

[Sommerswerd]

Thanks for your analysis Sullla! I for one would appreciate having your objective take after every battle.

It seems clear from Sullla's post that we just lost... period. We were mistaken to split up our forces, and probably mistaken to continue advancing into SANCTA land. Unless we are going to accuse SANCTA of cheating outright which, (as Donsig correctly points out) will be seen as pure sour grapes/ sore losering, we should withdraw/ ask for a cease fire.

 

[Kaleb]

Kaleb a pitboss game cannot be loaded in single player and to load in in multplayer you would need the game admin password. So unless Sancta have the admin password I cannot see how they could load the game and try out the battles.

Well there we go then... I guess it's not possible. was this always the case btw? or just with the new patch?

Just wondering though why we have a rule that people can't load up a game and play ahead?!

 

[Indiansmoke]

Well there we go then... I guess it's not possible. was this always the case btw? or just with the new patch?

Just wondering though why we have a rule that people can't load up a game and play ahead?!

Good question...not sure why we have this rule.

Are you in a game with 3.17 still? If yes just take a save and try to load it in single player...and see what happens...if it loads then we have a major issue. With 3.19 it definattely does not work.

 

[donsig]

Just wondering though why we have a rule that people can't load up a game and play ahead?!

The ruleset we have could have evolved from the MTDG ruleset where the games were (are) played via email. The previous MTDG may have been email as well. I don't know. Yeah, I know you can't play that far ahead in an email game but you can for the one turn you have. The play ahead rule is also in the democracy games and elements of that rule set may have formed the basis of the MTDG rules. too. I guess now the only purpose of the rule is to ensure we all recreate moves exactly if we have to reload.

 

[Kaleb]

Is there any way to edit a save so it can be opened in SP?

 

[Sommerswerd]

You can not load an MTDG game in SP. It will not load, the game tells you that you must load in MP. I mentioned that I tried this several months ago, to create a test map. Anyway, I think we can put this issue to rest. SANCTA just beat us, because our attack strategy was inferior to their defense strategy. Can we decide what to do next now?

 

[classical_hero]

Well I have lost to odds of over 99% (with me being the overwhelming favourite, it was a chariots against an Elephant), so don't feel too bad about that, since worse things have happened.