Ordering

[Cowmando]

I just tried adding to an order and the thing that tells you how much it will cost and how long it will take it told me that adding this order would take 6 ticks to produce. However once I made this order it turned out that it would take 7 ticks for the whole order to finish. So what I am suggesting is firstly to make sure the thing that predicts the production time is accurate to what will actually happen. Secondly also maybe is should be possible to cancel additions to old orders (perhaps only for the first tick they were added) separately from the whole order so if something like this happens to someone else again they can actually rescue the situation.

[Mzyxptlk]

The entire production system needs recoding, tbh. But this has been bitched about a lot already, don't think it's needed to add yet another topic.

[Ultimate Newbie]

I think its valid, at least the canceling of an addition to the order at the full cost (if its done that tick), because you can then fiddle to work out how much to actually produce without pushing it over x number of ticks.

Its a good idea, given the current system, which i imagine will stay for the near future. Gradual improvement in the user friendliness is a good thing.

[Cincinnatus]

Originally Posted by Cowmando

I just tried adding to an order and the thing that tells you how much it will cost and how long it will take it told me that adding this order would take 6 ticks to produce. However once I made this order it turned out that it would take 7 ticks for the whole order to finish. So what I am suggesting is firstly to make sure the thing that predicts the production time is accurate to what will actually happen. Secondly also maybe is should be possible to cancel additions to old orders (perhaps only for the first tick they were added) separately from the whole order so if something like this happens to someone else again they can actually rescue the situation.

to even have a chance to fully debug that "issue" you had, i need a much more detailed information on what you did and about the order...
like how much was in the order at first (all ships exactly), factories assigned, factory costs (incase they don't match factory count), the original production time,
how much it had progressed (not in % but in PUs), race, pop assigned to production, government, and how much you added to the order.

Originally Posted by Mzyxptlk

The entire production system needs recoding,

why?

[Mzyxptlk]

Because it's shit?

[Almeida]

Originally Posted by Mzyxptlk

Because it's shit?

QFT!

how many more ppl have to tell you (=the PA crew) that the current poduction system sucks ass, before you consider to change it?

another question: was it intended or did you just not think about it before implementing it: if you got enough resource income you can hide your resources until the end of the round in the production queu, simply order some ships, assigne 1 factory, fire all your shipwrights and keep adding to this order everytime before it finishes. this pretty much opens the way for another Greenhills surprise ending.

[Mzyxptlk]

That's probably the only good thing that this production system offers: a surprise at round end. I look forward to it already.

[Gio2k]

It's quite impractical anyways. It means you have to order constantly and be online to make sure your ships are there. Also once you ordered a specific type of ships you are stuck with them. What if you keep ordering loads of anti-de and all at once you need to cover for bs?

[Gio2k]

The changes in the production are the single most dumb idea in my opinion. But since the geniuses behind the coding decided they already had changed everything to work with production/construction/research units, and didn't want to change back, they just kept patching the flaws of this new system.
But i think the covop problems are the more serious issue in this round, and i hope someone changes the effectiveness of them asap.

[Mzyxptlk]

Originally Posted by Mzyxptlk

My suggestion (stolen from dec) for next round would be to reduce the stealth boost for planets that choose dictatorship to 15%, and that stealth regeneration should not be influenced by it. This should reduce cov op strength by about 55%

This should solve the problem without totally destroying them, as is the habit of PA Team (see XP).

[Heartless]

Originally Posted by Cincinnatus

why?

Because it is unnecessarily overcomplicated. Things like ordering ships should NOT be complicated.

[Gerbie2]

I’d like to see it improved so that:
1. keeping production going forever will no longer be possible
2. the formulae are easier to understand

I already proposed a change to solve the first point in another thread. Simply make it so that each time you add new ships to an existing order the amount of Production Units is recalculated for all the ships in the order, rather then adding new Production Units calculated for the addition.
You don't even have to change any formulae in the game, just how one of them is applied. A minor change that helps solve a significant problem.


The production formulae are unnecessarily complex. I would change them to make it easier to understand how the formulae work.

The manual displays the formulae like this:

The PU for an order depend on the amount of resources it is as follows:
Production Units = (total_resources_spent^1/2)*LN(total_resources_spent^2) where LN is the "natural log".

Each tick, your planet produces the following amount of PU for each of the Light, Medium and Heavy factories. This can be calculated by:

Output = int(((4000 * # factories of that type)^0.98) * (1 + (pop_bonus + gov_bonus + race_bonus) / 100))

where int() means "take the whole number part of".

Production time = (Required Production Units + (10000*# factories of that type allocated to that order))/Output, rounded up to the nearest number.


I would replace PU by a new measure of unit I call Production Time Units (PTUs). When ignoring bonusses, each factory produces 1 PTU/tick.

The formula looks like this:
Production Time Units = (total_resources_spent^1/2)*LN(total_resources_spent) / 2000.

This is a more complex formula than above. I made two changes:
1. replaced LN(total_resources_spent^2) by LN(total_resources_spent) * 2
(ln(a^b)=ln(a)*b)
2. devided the amount of PUs by 4000 (≈ the output of a factory). 1 PTU = 4000 PU


In the next formula I would remove the int() and ()^0,98:
Output = # factories of that type * (1 + (pop_bonus + gov_bonus + race_bonus) / 100)

Removing int() is obvious. Why round off in the middle of a calculation?
Removing ()^0.98 becomes clear when you realise that the effect of this complication is very small. Occasionally this will lead to an additional tick production time.


Now this leads to the following production time function:
Production time = (Required Production Time Units + (2.5*# factories of that type allocated to that order))/Output, rounded up to the nearest number.
The 2.5, rounded up to the nearest number is the minimum production time for every order (3 ticks).

edit: The effect of the ()^0.98 was somewhat larger than I first estimated. I changed the description. Removal could accellerate production by 15-20% (depending on the amount of factories). If this is unwanted it is possible to decrease the 2000 in the first formula as compensation.

[Mzyxptlk]

Near-identical formulae, but a lot easier to understand. Looks good to me.

[furball]

Alternatively, get rid of this shitty idea and go back to old-style production.

[Veedeejem!]

Originally Posted by furball

Alternatively, get rid of this shitty idea and go back to old-style production.

This can't be said enough i think

[Gerbie2]

That would solve the problem as well, but would not enable the features PA Team seems to want to introduce. It seems like they want to promote micromanagement and an incentive to have multiple factories.

If we look for a compromise between both systems, we can make the following mixture between the old system and the current system:
- place as many orders as you like
- don’t assign (1 or more) factories to specific orders
- instead assign 1 or more factories to all orders simultaneously to determine the output
- each order gets an amount of PUs assigned to it (or PTUs) when ordered (calculated as current or as I proposed)
- each tick the output is devided over all orders, based on their PU size
- you cannot add ships to existing orders

The result:
- no more delaying orders by adding new ships to it
- you can influence Output by adding or removing population or factories (micromanagement)
- large orders take longer to produce than small orders: there is an incentive to have multiple factories

Problem:
- you can slow down all production by adding a large number of small orders (this paralizes the type of factory)

To solve this you could limit the amount of orders people can have going. Right now the limit is the amount of factories you posses. This would be impractical if you cannot add to existing orders, especially at the start of the round when you have only 1 factory. For this reason it could be changed to a fixed number or something like: 5 + the amount of factories you posses.

[furball]

Originally Posted by Gerbie2

That would solve the problem as well, but would not enable the features PA Team seems to want to introduce. It seems like they want to promote micromanagement and an incentive to have multiple factories.

There are many good ideas for micromanagement. I have personally proposed many in the past, and will continue to do so. But this was never a good idea: there are many superb proposed modifications of 'classic production', but this is one of the terrible ones.

[Tomkat]

I actually quite like the new production methods. I don't understand what everyone's problem is with it

[Veedeejem!]

Originally Posted by Tomkat

I actually quite like the new production methods. I don't understand what everyone's problem is with it

Why "fix" something that wasn't broken in the first place?

[Wandows]

Originally Posted by Veedeejem!

Why "fix" something that wasn't broken in the first place?

By that you're saying that once it technically functions, it should never be subject to change? That would make game development/progess somewhat hard is it?

I think the change to "units" rather than ticks to control the production (or research or construction) time is potentially alot more interesting as its easier to influence through (other) game features, hence a nice step forward. Although i think the coding time spent on it could have been used better on other features that are a problem (i.e. combat engine bugs), i like the changes made (except for the hidden stack thing).

[Mzyxptlk]

The idea that a factory is limited in its output is good. The implementation however is terrible.