SGOTM 17 - Pre-game Discussion Thread

Status
Not open for further replies.
Has anyone used this trick during SGOTM?? It would of been easily possible last game really?

Kaku did it in SG16, but we asked permission before reaping the benefits, and we were approved.

I assume AlanH consulted with the team captains or simply made a ruling.
 
Kaku did it in SG16, but we asked permission before reaping the benefits, and we were approved.

I assume AlanH consulted with the team captains or simply made a ruling.

Well if other teams start to use it I guess all teams will. Not really in the spirit of the game in my opinion. Although looking at the start we may not have many resources to trade anyway.
 
Yea, I learned that trick a couple years ago, but never personally used it. I was not involved in Kaku's (actually quite small) use of it. I agree it seems 'not in the spirit of the game'.
 
Kossin, thanks for bringing up the Resource Trade Bug. I'm against its use, just as I was against the use of the flying camera, even if its use is hard to detect or prevent. That graph shows that exagerated use is easy to detect, plus inspection of the saves will show the same.
 
Not sure if this is the correct thread, but I thought I'd raise the issue.

I remember reading about it a long time ago but couldn't find any ruling on it so...

Resource trade bug
While not overwhelming in the last SGOTM in terms of research, it does have one effect: the AI which pays the bills will be slowed down from all the gpt he's spending on a few resources. (just how much is difficult to say)
Spoiler :
sgotm_graph_2.php


Link to how it works:
http://forums.civfanatics.com/showpost.php?p=12133596&postcount=4625


Basically, you can get multiple deals of x gpt which the AI wouldn't normally offer.

Given that the game is now Immortal difficulty, AIs are bound to have a bit more money around.

~~~

So, is it good and everyone can use it? Or will it be handled the same way as Flying Camera?

Thanks for clarification.

This is exactly the right time and place to discuss this.:goodjob:

I am of the opinion that if this became viewed as something necessary to use heavily in order to be competitive, the games would become more tedious and less fun. So I lean towards banning abusive gpt trades to the AI for everyone. There are a number of variations on this, involving pillaging your own resource to use it over and over this way. Unless someone has some good arguments why this ought to be allowed... lets hear it.

We'd need to word the ban carefully, so you are all invited to help out. Would it suffice to stipulate that you may not accept gpt trades/gifts/tributes from any AI that you are trading gpt to, and vice versa?

Well, I'll wait for more input before a final ruling, which will be definitely made before the game starts.
 
I think that is to harsh. I think giving gpt for nothing should not be allowed.
 
Is gifting gpt ever a legitimate way to gain diplomatic or other valid advantage? If not, then that can be banned. If it *is* legitimate, then maybe it should not be permitted to trade with that civ for gpt in the other direction?
 
I think that restricting gpt trades to only one direction will solve the problem. (AlanH's idea)
Eg, you cannot trade gpt to sy you already get gpt from, and vice versa.

We would still have the diplomatic uses this way.

I don't see how resource pillaging could be abused.
 
Eg, you cannot trade gpt to sy you already get gpt from, and vice versa.
While using a relatively strict rule like this one would be easier to define and easier to enforce, it can cut away some strategic options.

Players who do not use the trade subsidies approach will still very easily get into a situation where they are selling a Resource for Gold per Turn.

Yet, if an AI offers up a Resource that you really want (say, Marble), you may have already traded around a lot your spare Resources and you might end up having to give away, say, Corn + 15 Gold per Turn just to get that Marble Resource.

Do I have to cancel all of my existing sold Resource trades with that AI, which, if you are not using trade subsidies, are actually harder to renegotiate/reappear in the future? I might only want that Marble for 10 turns, but good luck in getting those earlier Gold per Turn trades back later.

Do I have to avoid getting Marble for 8 more turns because I didn't predict 2 turns ago, when I traded away my Fish for a small amount of Gold per Turn to the same AI, that they would connect up Marble?

So, if we were to use too strict of a rule, we'd actually be removing strategic options from players that do not use trade subsidies.


From my understanding, BtS cleaned up some of the avenues for abuse by having the AI be willing to alter their Commerce Sliders after every trade.


According to what bcool wrote, it seems possible to still partially implement this tactic, but I'm not certain that there is not a risk of the AI deciding "oh, look, I have +1 Happiness from that first trade, so now I'm able to work one more citizen in City X and thus if I were to get another 8 Gold per Turn, the extra Commerce in City X would allow me to raise my Science Slider and thus I won't actually offer 8 Gold per Turn to the silly player." In other words, do we know for certain that this tactic works 100% of the time as described, or is there an inherrent risk involved, such that the AI might decide to make some of that gifted Gold per Turn unavailable to the player?

So, it would be nice to have some proof that this tactic is 100% consistently exploitable.


Also, this issue has come up and was discussed to death previously and the ruling was not to ban it... and I'm pretty sure that a good part of this discussion happened before the patch that made an AI evaluate if they want to rework their Commerce Sliders after every single trade. That's not to say that we can't ban something now that was decided not to be banned in the past, but if it wasn't worth banning then, what makes it so much more powerful now that would make it worth banning?


Here are some relevant arguments that I have seen posted which argue against a ban:
1. You still need to own Resources to give to an AI.

2. You still have the maximum cap that the AI is willing to pay (8 Gold per Turn for the Resource, in bcool's example) such that the AI considers that amount of Gold per Turn to be a fair value for the Resource and won't pay any more for it.

3. You also have to accept a 10-turn no-net-gain in order to set up the trade. (Arguably, you are losing out for those 10 turns... net-neutral in terms of Gold per Turn... but you don't have access to the Resource, which could have been sold to another player, traded to an ally, used as a backup if you lost your other copy, etc).

4. Yes, you can argue that you could be trading away Resources that an AI "doesn't need," but AIs (from my experience) will be more willing to grow their Cities (grow them into their new Happiness/Healthiness caps) when they obtain additional Resources either via trade or in their own lands.

Yes, you are taking advantage of the fact that an AI is unlikely to cancel a deal that they think is fair for them, but what you are potentially abusing is the AI's inability to evaluate whether it is better for them to have that extra Happiness/Healthiness Resources or the Gold per Turn at the current time in their game (and who is to say that the AI isn't coded such that they are more forward-thinking than the current time?).

Yet, I make trades like these all of the time for my own Civ... an AI comes up with a Dye Resource (and that's a Resource that I don't have in my lands--or is a Resource that I do have but the Calendar tech is far away) and I'll do my best to trade for it, even if I'm 3 population points away from my Happiness cap in my largest City. I'm investing in the future with the potential to grow into additional Happiness (or to better withstand War Weariness, to be able to run different Civics, or a whole host of other strategic options). Basically, just because the extra Resource doesn't appear to benefit a Civ "right now" does not mean that it is not worth it for that Civ to seek out that Resource when the trade is available to them.

If an AI invested in a Dye Resource and later connected their own source of Dyes, they would come to you and tell you that they are cancelling that deal. So, code does exist to prevent abuses (i.e. code exists such that an AI will seek to cancel a trade where they are getting a Resource that will provide it with zero additional benefit due to already having a copy of that Resource, combined with the AI's inability to give away their only own copy of a Resource).


EDIT:
Is gifting gpt ever a legitimate way to gain diplomatic or other valid advantage? If not, then that can be banned.
Sure, legitimate examples exists for gaining a diplomatic advantage. Maybe you don't have a tech or a City to trade to an AI but you want to obtain up to +4 positive Diplo for Fair and Forthright Trading relations. Picture a game where you got Currency but not Alphabet or where you're behind a tech leader that doesn't have any techs that you want. Maybe the only tech that you have to trade is a strategic one that you don't want the AI to have but you don't mind the AI having some of your Gold each Turn.
 
The fact that it was not banned previously after discussion was the reason I told Kaku that they could use it. When they asked, I was not aware of any change in the situation that would affect the original decision.

The exploit is not the trade, but the fact that you deliberately, prematurely cut off the subsidy. It seems to me that this is the action that is most worthy of a ban.

LowtherCastle said:
That graph shows that exagerated use is easy to detect, plus inspection of the saves will show the same.
I'm not convinced that the graph alone shows that such a deal was done. Kossin's snapshot of the SGOTM 16 graph shows that several teams had steps in Foreign GPT during that period. Were those steps all due to similar deals? If we have to manually crawl through saves to find and examine everyone's deals, I would welcome some volunteer help!

Fact is, we ultimately have to trust each other to play by whatever rules we agree.
 
I can see situations where you get gpt from an AI in one trade and give the AI gpt in another future trade as Dhoom suggested. So I agree that we can't have a simple rule stating that gpt trades can only be made in one direction with each AI.

However, in my opinion, it is an exploit to gift an AI gpt with the intention of taking it right back in a subsequent trade and then to cancel the gpt gifts 10 turns later.

I'm fine that it's been used in past SGOTM/xOTM games since it's never been explicitly banned. However, I hope that we are able to correct this "exploit" going forward...
 
Is gifting gpt ever a legitimate way to gain diplomatic or other valid advantage? If not, then that can be banned. If it *is* legitimate, then maybe it should not be permitted to trade with that civ for gpt in the other direction?
Unfortunately, there are ways around your solution, that is, the bug is not strictly a function of gifting gpt.

The only simple check that AlanH has, iiuc, is the Net Foreign GPT Trade graph.

I don't know the code, but clearly there are constraints to how much an AI will give at any given point in the game. Suddenly, the AI offers a bunch more than before. The Bug problem is compounded by the possibility that multiple AIs suddenly offer a bunch more gpt, so a team might legitimately increase their their Net GPT by 40 gpt.

So howis AlanH to know whether the 40gpt increase on the graph is kosher or not?

As far as I can tell, what each AI shows is what their new max limit is. This might be slightly adjustable, for example, if they bump it to 8gpt and you cancel a 1gpt deal, then it might go to 9 gpt. But in that case, he'll show the 9gpt after you cancel the deal.

When abusing the Bug, he never shows more than, for example, the 8 gpt. The catch is that he'll keep showing 8 gpt over and over for each trade, rather than it dropping to 0 after the first trade, which is what happens when you don't abuse the Bug.
What we're not considering legitimate is tricking him to give more than the code says he can afford and that's the max he'll show on the board, before or after you cancel trades.

Thus, the only "rule" I can see that makes any sense would set an arbitrary max threshold for "unreported" gpt trades, such as 9 gpt cumulated from all AI gpt trades on any given turn. If the AIs suddenly cumulatively offer more than this arbitrary threshold, then upload the file in which the AIs offer more than 9 gpt on a given turn BEFORE making any trades. Then you're covered. The save is on record, proving that the 22 gpt increase in Net Foreign Trade came from 22 gpt from Brennus, or from 12gpt (Brennus) + 10 gpt (Hammy), or from 8 gpt (Brennus) + 8 gpt (Hammy) + 6 gpt (Qin).

Codedivers and/or those who build the files that AlanH uses for making those graphs might have a more elegant solution.
The exploit is not the trade, but the fact that you deliberately, prematurely cut off the subsidy. It seems to me that this is the action that is most worthy of a ban.
Actually, that's just the AI stupidity. What Tachy did was make an even trade for 10 turns, then legitimately cancel it. Qin could have likewise canceled all 5 of his deals but "chose" not to. At that point, it became a Bug, because Qin's economy was giving up ~5x too much gpt per the code.

That there is worse potential abuse, starting from T0 of the trade sequence, is the larger problem, but either way, after 10 turns, Qin's giving up more than he should.

I'm not convinced that the graph alone shows that such a deal was done. Kossin's snapshot of the SGOTM 16 graph shows that several teams had steps in Foreign GPT during that period. Were those steps all due to similar deals? If we have to manually crawl through saves to find and examine everyone's deals, I would welcome some volunteer help!
You're right. I forgot about the possiblity I mentioned above, that conceiveably 5 AIs suddenlyl offered 8 gpt, so the graph doesn't necessarily show Bug abuse. It's an indicator and I believe my solution above should be relatively simple for you to implement, given that we can assume the bug's not going to be exploited now that we are agreeing not to.
 
The Shawshank Redemption's gpt shot up once we learned Astronomy and were able to start making resouce trades with the over-seas AI. We did not use the "Resouce Trading Bug" or whatever we decide to call it.

So, I agree that by looking at the graphs alone, it's hard to tell when a team has used it.
 
AlanH: One more detail.

Another issue is that the graph shows the change between one save and the next, which is usually more than one turn. So a thrid possiblity is that a 40gpt change that shows up on the graph came from, for example, 5 trades of 8 gpt to 5 different AIs, on 5 different turns. So according to the rule I proposed above, there was no need for the team to make a save, because each trade was under the threshold for that turn. Well, you can check that by looking at the final save and noting that there are 5 separate trades to 5 different AIs.

There might be more variations and potential for abuse, but like you say, this is a friendly game, based on trust. Most important, imo, is that we agree to some rule and trust one another.

xpost with Mitchum: In the final analysis, the tell-tale sign of the bug is a pile of X gpt trades with an AI that are inordinate for that point in the game. This can also be resolved with judgment.
 
The exploit is most viable on the last turn of the research of the AI. It depends if the AI has gold laying around and how many units, if it's currently in war. Yet most stuff is quite predictable if you wish to exploit it.

The resource pillage export/exploit goes like that: give the AI a lump of gold + a single resource. Exploit the AI by selling all extra (spare) resources, pillage the road or the resource. The 1st deal is canceled and the AI - screwed.

Yet, even w/ AI_doCommerce when the gold/Turn is changed, giving gold/turn and then taking it back restores the equilibrium. The exception would be if the AI has enough gold to bump the science (or culture/espionage) slider w/ the extra gold - if the initial gold/turn is closing to zero, the odds are very odd. Side note: it's unfair (or lame) the AIs to assign tiles outside their turns but well.

my opinion: this is a blatant exploit and I'd consider it heavily frowned upon.
 
In the final analysis, the tell-tale sign of the bug is a pile of X gpt trades with an AI that are inordinate for that point in the game. This can also be resolved with judgment.

Agreed. Getting 30 gpt total in trade from multiple AI when Astro is learned, even within the span of 1 turn, is not unheard of in a typical "continents" game. Having 30 gpt from one AI at the end of the game is also possible as multiple 8 gpt trades are made throughout the game. However, getting 30 gpt from a single AI in the span of 1 turn (or even 10 turns) is highly suspect.
 
We're all here to "Kill Them All"! Now players want to redefine the game so certain ways of "Killing Them All" is illegal. What kind of insanity is this?

We should ban glaringly obvious bugs like infinite technologies from The Oracle or Liberalism or liberating a City an unlimited number of times to an AI while at War with it. Other "bugs"/"design flaws"/"I just don't like the way this game mechanic works" game elements that are limited in the advantage they provide should be allowed to be used by all teams.

The list of "exploits" is long enough as it is, without adding things that simply put the AI at a disadvantage. Everyone has a level playing field with or without banning these marginal means of advancement, so why complicate things by adding yet something more to ban unnecessarily.

Perhaps we should ban the use of strategic resources when we "Kill Them All"! ;)

Changing the game is fine, if we could change the executables to enforce it, but we can't. So let's just play the game as it is, banning only things that give unlimited advantage that consequently destroy game balance.

This little Wpt gifting tactic will not destroy game balance.

Worst of all is the burden of enforcement placed on the SGOTM staff of all these petty bans that don't affect the fairness of the game between Teams. We have no obligation to be fair to the AI. Our obligation is fairness between the Teams!

Sun Tzu Wu
 
Sun Tzu Wu may have a point. Kakumeika was cultivating Qin as a tech trading partner. We got Physics and Combustion (iirc) from him and wanted Rifling as well, but he was too slow. It could be that we shot ourselves in the foot with our 32gpt from him.

In SG17, a team would have to half-cocked to research Currency just so they could take advantage of this bug to slow down the enemy's economy.
 
Is gifting gpt ever a legitimate way to gain diplomatic or other valid advantage? If not, then that can be banned. If it *is* legitimate, then maybe it should not be permitted to trade with that civ for gpt in the other direction?

Gifting gpt does not (AFAIK) grant you diplomatic gain whatsoever.

One of the uses for it is when the AI is in "financial trouble" and you want to gift the AI some cheap units to take advantage of their upgrade discounts.

~~~

Personally, I don't mind if the decision goes either way, I just wanted to clarify so that everyone was on the same page.

Sure, it can be used to slow down an AI (which you may want to conquer in, say, 40 turns) but it also diminishes their research rate hence trade opportunities as raised by the above posts.

There's no simple workaround for it (if a ban is implemented) similar as Flying Camera. Honor system should be sufficient if necessary imo.
 
Status
Not open for further replies.
Back
Top Bottom