What do you consider an exploit?

Sadly, that wasn't a joke post. It actually happened.

Given that the AI will take a 'risk' and break a LSG for lux deal shortly after signing it, by DoWing me, I'll give the dev. the benefit of the doubt that a risk assessment was part of the deal. they just had a problem of understanding the numbers.

 

The costs are opportunity costs, and depending on the situation they can overpower any incentive to spam war decs to farm money:

- You can't sign RA with that civ while at war, and possibly after
- You can't milk said AI for DoF mutual friends which lets you get away with murder elsewhere for a while
- Any alternative benefit (such as trading for their resources/more gold using something other than war) is gone
- You take a permanent relations hit every time you declare, and AI who like your target won't like war shenanigans
- You have to actually be strong enough to at least survive the war! Not hard, but it requires some up-front investment into defenses on high difficulties.

Doing this at the wrong time could easily cost you the game...which means one needs strategy to recognize and employ it properly.

Too bad firaxis never actually fixed the bug - instead of too much overflow, they removed all overflow -----> gold and reintroduced the micro necessary to avoid it. They said they were trying to "clean up" the code in unofficial 3.17 where overflow was handled properly, basically admitting error here...

and then never lifted a finger to change it ever again

Don't think for a second that I'm only hating in firaxis civ V conduct; they have done very questionable things and left aspects of gameplay unfinished for years before civ V.

I agree its bad design. Tech trades were pretty bad too, but the choice between completely random outcomes or esoteric micro (in a game supposedly meant to be better for casuals!!!!) really grates.

There's no logic for putting the break point there. The AI is incapable of consistently evaluating good sites for national wonders, it is incapable of determining when a war is in its best interests, and it is incapable of using its military properly vs a tactical human who can "farm" it for xp and great generals. Whether the AI does any of that semi-ok or terribly is RNG. Whether the AI makes a lump sum deal with you and breaks it by declaring on you (!) is also RNG.

Again, for variant play it can be fun to make up extra rules or handicap yourself, but I've yet to see someone present a logical reasoning to draw the line here rather than somewhere else; the AI is bad at a LOT of things or simply isn't coded to do a lot of things; many of these things people would not consider to be exploits!

It's pretty dubious to even believe the AI was intended to play like a human; I personally think it should, but do you think the devs intended it that way? We see a lot of evidence they did not. Why, then, hate on strategies that abuse deliberately allowed AI ignorance? Isn't that what the bonuses were placed into the game to fight against? Why not draw the line anywhere? Flip a coin, one game you must auto workers...it's all the same :/.

Unfortunately, I know way too much about civ IV to give them a benefit of the doubt in AI code for V . I'm betting you got sheer RNG screwed where the AI just doesn't check for PTO obligations inbound when declaring...possibly so that humans don't game anti-war by selling PTO, but more likely because they didn't feel it was necessary (or didn't consider it).

Kind of silly that the AI committed a HoF violation on you though . I'm sure you dropped the ban hammer very hard on that one .

 

I'm almost tempted to try to upload the game to the HoF to see if it gets in or not, considering the AI really did break the HoF 'rules'.

 

The problem is that the AI never "thinks" in the traditional sense, it's all a long set of probability-weighted dice rolls. When it has a chance to execute a tactic by sheer chance, it's hard to reliably distinguish that from other sheer-chance decisions the AI makes.

Technically speaking, the AI can do a whole heck of a lot more in-game than us and has been able to do so at least since civ IV but probably in all of them.

In this context, the AI has the same "tools" (DoW, lump sum for PTO then break it, etc) as we do, and similar to its heroic epic or oxford placement, simply uses them poorly. It's all how you frame it, and unfortunately very arbitrary when one goes to draw a line.

That's why I have a hard time accepting anything to be called "exploit" unless it directly violates in-game rules based on UI descriptions/etc (examples: oracle infinite tech bug in civ IV, anarchy to produce units via switching into liberty endlessly in civ V).

 

Well, there's the easy test and then the complication. The easy test would be to find out if the AI gives significantly less money in a straight cash deal than the per turn equivalent. The problem is then you have to figure out the declining value of money when spread out over time (it's always better to have money up front than per turn). Of course, the real world calculation is easy, but I don't have a clue what the inflation rate is in a Civ game

 

A present value calculation in civ WOULD be useful if one could come up with some basis in its application . Seems silly at first, but not so much given the kind of game this is and such a thing could actually make the human (and the AI if tweaked around it) better players.

IIRC the AI does not value 10 gold/turn the same as 300 gold, though I haven't tried that and compared exact values in a long time as they usually want something ludicrously specific like 351 gold (as opposed to 348).

 

Yeah, I think a present value calculation would be useful, but, given the extreme complexity of factors to consider, not useful enough to justify it. We live in a market system, where a calculation like that almost makes itself (interest rates vary, but there's a general range that's stable). In Civ, you'd have to consider how much infrastructure you have, the cost of things to purchase, and the chances of war. All of these factors would change the value of cash in hand dramatically.

 

The AI's present value system is laughable. It would rather give you more gold now (300 for a lux) than less gold later (9/turn +17).

I've considered trying to curve-fit my empire's total output as an exponential, to get an estimate of the "fundamental time constant". The difficulty is deciding how to aggregate output (beakers, gold, GPPs, etc) into a single number; though hopefully, if you averaged over enough games, the specific aggregation formula wouldn't matter too much.

 

I used to consider Tech Blocking exploitive, but the devs have patched it from 25% to 33% rather than removing it. Thus, they obviously *want* it in the game at this point.

Selling open borders- I don't consider it an exploit. You are getting gold for a commodity. The AI will use those borders. I've seen them use my open borders to wipe the civ behind me before.

Barb re-capture- I do consider this an exploit. At the very least there should be something like a 10 turn cool down before you can get the benefit from the CS again. A real city-state would be mad that you didn't escort their worker to safety after freeing it if it got recaptured.

Siam's wat's- Not an exploit, IMO, just imbalanced.

Cutting a deal for gold right before DoW- I don't consider this an exploit. It seems like something that a real life warmonger would do. Just look at how many countries that hate USA that they give billions of dollars to each year.

 

that's a logical leap you cannot really make.
given that they have limited dev power, and simply changing the blocking percent from 25 to 33 or whatever they did is a one liner code change, my guess is they dislike RA blocking but were unsure how to fix it easily, so made it slightly harder.
i won't be surprised if they keep revisiting it.

removing blocking completely isn't viable, as they wouldn't want casual players who know nothing of it to occasionally get the tech they're 90% done with as a result of a RA.

 

The LOGICAL tweak to RA would be to allow people to just choose their tech, and then simply raise RA cost.

Super random chance elements are bad for gameplay, and if it's not a bonafide random chance, get rid of the obscure micro that does not exactly "cater to casual players", which was their excuse for leaving the UI unfinished.

 

Or fixing a supplement amount of beakers given depending of many factors of your civ and your associate like population, litteracy, etc.

Can also be associate's factors like diplo status, for example a RA signed with a guarded civ can gives less beakers than a friendly civs with equal powers.

An RA is by definition a pact of research. In true life, it gives a sort of acceleration of research of many discoveries, but nothing like we can see actually in this game, which are one tech at a time, and entirely choosen by you.

It's really weird to see a civ discover nanotechnology with the ENTIRE help of another civ who just discovered education.

 

RA blocking cost is still 25%, they only added a +3 beaker rounding error.

 

The things I can be bothered to do, those are strategies. The things that take too much micro for me to bother with, those are exploits.

I agree that they can be defined as taking advantage of an oversight in the game design. So selling resources for gold, even if you probably get too much ROI, I'd consider a very strong strategy; selling 6 resources for 1800 gold the turn before a barbarian pillages one of them and cancels the whole deal I'd call an exploit since it's design oversight (and actually one that was created in a patch IIRC, since resource deals used to keep going even if you lost access or declared war partway through the deal). I'd also call DoWing after making a trade an exploit for the same reason - TMIT pointed out the opportunity costs, but I'd consider it an oversight because there is no way in hell you could realistically expect anyone to be willing to make a deal like that after seeing you pull that even once or twice. It just sounds like something Firaxis didn't consider when they changed how deal cancellations worked.

RA blocking is a toughie. It's taking advantage of a mechanic that was a failsafe intended to make RAs more desirable, since if you got the tech you were 10 beakers from completing, you'd be furious. Yes, tech trading had to go since it broke CIV in all kinds of ways, from whacking out power levels randomly/arbitrarily to overpowering vassals (probably shouldn't even mention vassals on these forums; TMIT may be watching!) to punishing the hell out of isolated starts. But they could have had it both ways - less exploitable/imbalanced but without arbitrary rules that just beg to be exploited. Personally I think they should have done RAs in the vein of Master of Orion II and just made it an over-time cash investment for an over-time raw beaker return. Perhaps you could even give them flexible rates and give the player a little bit of control over turn-by-turn beaker rate. (I enjoyed being able to get that important tech a turn faster in CIV by maxing the slider, but I think the straight equation of "gold=beakers" was awful; this would be a nice compromise between gold to beaker conversion, tech trading, and the RA system.)

 

I think the best solution to RA blocking would be to base the chance of receiving a given tech on the amount of beakers already invested in it. You could do some tech blocking to hedge your bets, but it would never be a sure thing

 

this is how RAs were described in the original civ 5 pdf manual.

 

If that's the case, the coding may be around for it. Hmm...

 

From what I learn from jobs... any good use of opportunity is an exploit.