Multiple units per turn, unit overflow gold, wonder fail gold

[vedg]

I have played K-Mod, but not Advanced Civ yet. I'd like to discuss my misgivings about documented Advanced Civ's changes to unit hammer overflow. Hopefully I misunderstood the manual and the situation is better than it seems to me.
123f disabled overflow gold for units.
064b disabled putting unit overflow into another unit of the same type.

Usually, when excess overflow occurs, K-Mod produces another unit only partially.
To redeem the production spent, the player will have to continue training that unit.
But this will result in even more overflow as it's going to be a cheap unit and
maximal overflow is already being stored. For example, if a city with 105 production
per turn trains a Spy (40 production), it'll store 40 overflow and put 25 into another
Spy. If the player finishes that second Spy to avoid missing out on the 25
production, 130 overflow will result in 40 overflow stored, two more Spies and 10
production for a fifth one. Ultimately, this leaves it up to the player to avoid excess
overflow in the first place.

I agree with the issue quoted from the Advanced Civ manual. But I am not sure disabling the overflow is an improvement.
In K-Mod a large-production city can produce several cheap units and leave some unit overflow hammers be; then switch to an expensive unit or building to get rid of the excessive overflow; then finish the cheap unit with no overflow put into the next unit of the cheap type. This is an unfortunate micromanagement, to be sure.
In Advanced Civ training a cheap unit in a large-production city for even two turns in a row is very wasteful. By the time the too large production rate becomes an issue, building research and wealth is long since available. So hammers are already at least as valuable than gold. Therefore, even overflow gold is usually a punishment. Simply lost overflow is a much worse punishment. This issue was likely a rationale for the K-Mod multiple-unit-per-turn change. It is aggravated a lot by Quick game speed.

---

Another suggestion.

Tbd. Perhaps convert only 50% of the invested production into gold and the rest into city
culture. That should be a separate change id though because it's a balance change
rather than just closing a loophole. For now, fail gold remains a valid tactic, and this
gives wonders with very weak abilities (e.g. Chichen Itza) at least some use.

Building Wealth and Research becomes available quite early. Therefore, unmultiplied fail gold is an exploit only in early game. Would it be possible to remember how many hammers where multiplied by a wonder-specific resource, leader trait (Industrious or Spiritual for Cristo Redentor) or building-only bonus (Org. Religion) and not compensate them? In that case National Wonder fail gold could be restored in order to let players change their minds about where to build a National Wonder without too harsh a punishment. Perhaps only hammers, invested into a National Wonder after building Wealth or Research becomes available to the player, should be compensated.

 

[f1rpo]

Sorry about the late reply, I had to find some time to refresh my mind about this subject. As a side benefit, I've found a minor bug in my calculation of the limit for overflow production – the production modifiers of the completed order had not been taken into account.

[...] I'd like to discuss my misgivings about documented Advanced Civ's changes to unit hammer overflow. Hopefully I misunderstood the manual and the situation is better than it seems to me.
123f disabled overflow gold for units.
064b disabled putting unit overflow into another unit of the same type.

The manual does say that overflow gold is disabled, but the situation is better than that – because I actually re-enabled overflow gold when I implemented change 064b. Apparently, I just forgot to update the documentation of 123f. (My ids aren't strictly chronological – they're not even good for that, unfortunately.) I guess my thinking was that, with multiple units producible per turn, overflow gold pretty much no longer occurs. So, when I went back to one unit per turn, I also brought back 1:1 conversion into gold – and added a tooltip predicting how much gold will be generated and an on-screen message reporting the actual amount.

By the time the too large production rate becomes an issue, [...] even overflow gold is usually a punishment. [...]

Agreed.

This issue was likely a rationale for the K-Mod multiple-unit-per-turn change.

I guess so; the punishment is a mild one, but getting rid of it entirely would mean that the player doesn't have to worry about excessive overflow at all. Maybe karadoc didn't see that his change falls short of such a goal. He may have also noticed that overflow gold lacks proper UI support. Don't know if he felt that faster production of Missionaries and Spies in the late game would be desirable. Production of multiple (high-level AI) city defenders per turn is imo not desirable.

In K-Mod a large-production city can produce several cheap units and leave some unit overflow hammers be; then switch to an expensive unit or building to get rid of the excessive overflow; then finish the cheap unit with no overflow put into the next unit of the cheap type.

OK, so it's not quite as bad as I presented it – the player doesn't have to produce Spies forever to eventually absorb the overflow. But it's just uncalled for to assume that the player wants two Spies when they've only ordered one. This can easily be more punishing than conversion into gold.

It is aggravated a lot by Quick game speed.

I haven't experienced that first-hand, but I assume that it is because per-turn production rates are, generally, as high on Quick speed as on Normal speed while the overflow cap is the total production cost of the completed order, which is only 2/3 on Quick speed, e.g. 25 for a Spy instead of 40. AdvCiv instead sets the cap to the size of the city's food store (i.e. 20+2*population) – which is also smaller on Quick speed, e.g. 33 instead of 50 at 15 population. Maybe this food store condition should get a game speed adjustment – to cancel out the adjustment of the food store size. Also, regardless of game speed, I still see some leeway for a higher cap if overflow gold occurs too frequently, e.g. 1.5 times the food store size. But, even as it is, it's (in most situations) already less tight than in BtS. Edit: Chopping production not counting toward the cap and never getting lost or converted might also be worth mentioning here.

[...] Would it be possible to remember how many hammers where multiplied by a wonder-specific resource, leader trait (Industrious or Spiritual for Cristo Redentor) or building-only bonus (Org. Religion) and not compensate them? In that case National Wonder fail gold could be restored in order to let players change their minds about where to build a National Wonder without too harsh a punishment. Perhaps only hammers, invested into a National Wonder after building Wealth or Research becomes available to the player, should be compensated.

I hadn't thought of (partially) compensating players when they change their mind about the location of a national wonder. Being forgiving like this is generally a good thing (though I rather doubt that this was the intention). Tracking the invested production without modifiers in a separate variable should not be a big technical challenge. Players wouldn't really have to be aware of such a change, but, well, some players want to be aware of everything, and it'll be an additional complication. And I guess it's pretty rare that a player will find a much better location only after having already invested substantially into a wonder. So I don't really like this – not nearly enough to bother with it at this point.

 

[vedg]

I haven't experienced that first-hand, but I assume that it is because per-turn production rates are, generally, as high on Quick speed as on Normal speed while the overflow cap is the total production cost of the completed order, which is only 2/3 on Quick speed, e.g. 25 for a Spy instead of 40.

Yes, a Spy and a Missionary cost 27 hammers on Quick. The problem is not limited to very cheap units, especially on Quick. Executives are relatively cheap (67 hammers) there. And the Heroic Epic city easily generates more hammers per turn than the cost of the most expensive unit. Given that multiple military advisors and possibly West Point can be present in this city, converting hammers into gold is usually undesirable. The same issue (somewhat less perhaps) applies to the Ironworks city. Even a cheap building, such as a Temple costing 53 hammers, is a challenge to produce without gold overflow in a production-rich city.

Production of multiple (high-level AI) city defenders per turn is imo not desirable.

The AI is disadvantaged in warfare anyway. If it is defending, then it must be at high risk already. Producing multiple expensive units per turn comes at a cost (population, chopping or hurry gold) and cannot be sustained for long. So I don't see it as a big issue.

But it's just uncalled for to assume that the player wants two Spies when they've only ordered one. This can easily be more punishing than conversion into gold.

Agreed. Sometimes it is more punishing. Maybe give the player control over this? For example, scheduling Wealth, Research, Culture or perhaps even some other building or unit puts overflow hammers into those? For simplicity, could even require explicitly scheduling multiple units of the same type to disable converting overflow hammers into gold. This might be more intuitive than the K-Mod behavior. Plus it would be easier not to micromanage the production order: just schedule what you need and it will be built in the specified order without any overflow gold side effects.

---

Would it be possible to remember how many hammers where multiplied by a wonder-specific resource, leader trait (Industrious or Spiritual for Cristo Redentor) or building-only bonus (Org. Religion) and not compensate them?

I hadn't thought of (partially) compensating players when they change their mind about the location of a national wonder. Being forgiving like this is generally a good thing (though I rather doubt that this was the intention). Tracking the invested production without modifiers in a separate variable should not be a big technical challenge.

The National Wonder suggestion wasn't my main point, just a side benefit. The most important advantage of not multiplying the fail gold is to combat rich wonder multipliers exploit. That way the fail gold will no longer be more attractive than the intuitive and intentional ways of converting hammers into commerce: building Wealth, Research or Culture. This was meant as an alternative to this idea in the manual:

Perhaps convert only 50% of the invested production into gold and the rest into city culture.

 

[f1rpo]

[...] The problem is not limited to very cheap units, especially on Quick.[...] Even a cheap building, such as a Temple costing 53 hammers, is a challenge to produce without gold overflow in a production-rich city.

I see. Setting the limit based on the food needed for growth rather than on the production cost of the completed order is intended to help when an average large city produces something very cheap. It's not going to accomplish much for the top production cities, even if the food-based criterion gets a game speed adjustment. I don't really mind if a Temple generates excess overflow in the Ironworks city, but this ...

And the Heroic Epic city easily generates more hammers per turn than the cost of the most expensive unit. Given that multiple military advisors and possibly West Point can be present in this city, converting hammers into gold is usually undesirable.

... is awkward. Military Academy also boost military production. Although, if there is no way to avoid excess overflow in such a city, then at least there's no incentive for tiresome micromanagement.

Maybe give the player control over this? For example, scheduling Wealth, Research, Culture or perhaps even some other building or unit puts overflow hammers into those? For simplicity, could even require explicitly scheduling multiple units of the same type to disable converting overflow hammers into gold. This might be more intuitive than the K-Mod behavior. Plus it would be easier not to micromanage the production order: just schedule what you need and it will be built in the specified order without any overflow gold side effects.

Yes, maybe cities should simply be allowed to finish any queued orders within a single turn; conversion into gold when there are none.

The AI is disadvantaged in warfare anyway. If it is defending, then it must be at high risk already. Producing multiple expensive units per turn comes at a cost (population, chopping or hurry gold) and cannot be sustained for long. So I don't see it as a big issue.

I'm guessing that it matters most for Catapult-vs.-Archer wars on Deity difficulty, perhaps also on Immortal. Maybe wars between AI civs are the bigger concern than human vs. AI. Should be less of an issue in AdvCiv due to reduced AI production discounts. Also, the AI could simply always go for conversion into gold; I think that would be the standard behavior anyway because the AI normally has no reason to queue up production orders.

The National Wonder suggestion wasn't my main point, just a side benefit. The most important advantage of not multiplying the fail gold is to combat rich wonder multipliers exploit. That way the fail gold will no longer be more attractive than the intuitive and intentional ways of converting hammers into commerce: building Wealth, Research or Culture. This was meant as an alternative to this idea in the manual:

Perhaps convert only 50% of the invested production into gold and the rest into city culture.

OK, that makes sense, treat fail gold as if the city had been producing Wealth all along.
Well, sensible ideas, but arguably not improvements in every respect, and I'm pretty much only doing maintenance these days ...