Single Player bugs and crashes v37 plus (SVN) - After the 24th of December 2016

[JosEPh_II]

Glad it helped. Now that game will be even more of a challenge once I lose my "bonus" research points.

 

[JosEPh_II]

Why do Units drop to 1/100th of their new strength after upgrading? See screenshot, Upgraded Musket men to Riflemen.

This happens to even LE and Healers as well as Mil units. Not sure about worker upgrades yet.

Hopefully this is Not intended.

 

[Toffer90]

Why do Units drop to 1/100th of their new strength after upgrading? See screenshot, Upgraded Musket men to Riflemen.

This happens to even LE and Healers as well as Mil units. Not sure about worker upgrades yet.

Hopefully this is Not intended.

↓ Something went wrong here, I guess. ↓

9780

• Upgrading units should no longer be able to kill them when the unit is severely hurt. There could still be some odd places where it can happen but I'll need to follow the processing on that with a savegame.

 

[JosEPh_II]

↓ Something went wrong here, I guess. ↓

Good thing I'm not at war. This would be devastatingly bad on the battlefield. And it's taking up to 4 turns to heal up even in a City with lots of Healers.

 

[Thunderbrd]

yeah I'll fix it asap. sorry.

 

[JosEPh_II]

Glad I caught it before a player reported it.

 

[Thunderbrd]

You want to put a warning up on the last commit notes in the SVN thread? I won't be able to repair it for some hours yet.

 

[JosEPh_II]

You want to put a warning up on the last commit notes in the SVN thread? I won't be able to repair it for some hours yet.

Did that.

 

[Thunderbrd]

Thanks Joe. I think I've got this right this time. About to test it once it finishes compiling. I'll try to make sure it gets committed tonight.

 

[JosEPh_II]

New consideration for Diplomacy and Trading units. I know there were some changes done recently but I now want to show some examples of in game results.

1st screenshot is Asoka wanting to trade Constitution plus getting 499 for 89 of my LE, Health and War dog units. the research points for Constitution is 12501. Now if 1 = 1 , then his side of the equation is = to 13000. Thus my 89 units individual value is 146.07 ;gold:.

2nd screen shot is Darius wanting to trade for a single Elephant worker which he values at 1188 . Which has been pretty standard price in this Normal, Immortal game.

My contention is that the 146 that Asoka is assigning to my units is too low. And the value for a worker too high.

ATM I don't know where these evaluations take place. I'm assuming it's in the .dll.

Also if I wanted to buy Constitution from Asoka he will ask for ~17,xxx . I'm of the old school where I think this too is too high. the ratio is ~ 1.35 for 1 . A ratio of 1.2 to 1 is more palatable imho.

What do you the other Modders think about this?

The 1st scenario is a no go from the start for me. It would strip 6.84 units from my home continent Cities, per city. This would also devastate my crime and health levels. It would also stop my production of other things to solely concentrate on rebuilding all these units. Which is a viable strategy but does the AI really "know" this? Plus Asoka is starting to feel the pinch for neglecting his own cities crime and health issues. I'm killing scoundrels coming from his territory into my 2 cities I've placed on a continent we now share.

And as you can see I'm middle of the pack in overall strength. That came from being isolated on a small continent for the 1st 3+ Eras. I'm the blue Empire in the game.

 

[Noriad2]

I think that workers should be worth half their price in hammers. Why buy 2nd-hand if it is just as expensive or even cheaper to build one new (considering hammer production bonuses)? AI should only willing to buy a worker if it is a bargain.

 

[Toffer90]

I think that workers should be worth half their price in hammers. Why buy 2nd-hand if it is just as expensive or even cheaper to build one new (considering hammer production bonuses)? AI should only willing to buy a worker if it is a bargain.

If it takes 40 turns to train 10 workers, you can save a lot of time by buying them. Though, IMO it should be close to: 1 : 1 (gold : hammer).

 

[JosEPh_II]

What is your thoughts on the unit values vs Tech? If the needed to make the unit matched it value in and then also = in , the value of my units would be closer to 400+. Not the value of 146 as Asoka wants it to be.

 

[Thunderbrd]

Y'all should see the list of considerations the code gives to evaluating these trade values. It's crazy complicated but it actually makes some sense until you see what it comes out with. They start with comparisons like what you're talking about but then take into account how desperate the AI is and so many other factors. I needed some feedback on the end results, yes. And this helps. I'm kinda not disliking how little the AI is valuing normal units (because the AI always tries to give you a crappy deal... it's their nature. But yeah, we could do to crank down the worker trade values further than by just a 50% reduction as it currently is set to. There's a global for that

 

[Thunderbrd]

Also if I wanted to buy Constitution from Asoka he will ask for ~17,xxx . I'm of the old school where I think this too is too high. the ratio is ~ 1.35 for 1 . A ratio of 1.2 to 1 is more palatable imho.

This could also be influenced by the particular leader. And I'm not thinking that's too high unless we have made gold super challenging. There's a lot of ways to make extra gold too. Maybe we finally have good cause to make that a high priority for the player.

Maybe eventually we should add more layers to the diplomacy chart and vary it more by how solid your alliance is.

 

[JosEPh_II]

Most of the AI only have a treasury of a few hundred to a couple thousand . I have used the trade ships and every civic and building that gives more trade routes and higher trade %'s. The AI is also using Trade ships but they spend their money as fast as they make it. I upgrade in batches when my Treasury get to be 30,000+. I was at one point only a 0.3 vs Tokugawa and 0.5 vs Nebuchadnezzar and Ivan. I'm finally making progress in this area as you can see in the screenies.

Tokugawa has been very agressive in expanding and fighting. I'm surprised Ragnar still has 15 cities at this point in the game. He had 4 AI at war with him most of the time.

And I'm actively helping Wilhelmina stay alive. Gifting her lower level tech while making small resource trades and 1 big tech trade with Ivan to to get him to stop invading her.

If it takes 40 turns to train 10 workers, you can save a lot of time by buying them. Though, IMO it should be close to: 1 : 1 (gold : hammer).

I agree with Toffer on this one.

But there is no way an AI should want 89 units (and they are not low level units for the time frame and tech level of the game) for a 12501 tech. Even if they are in an Open Borders agreement with you.

 

[Toffer90]

Rewriting diplomacy is something that is crucial that we do one day, the magnitude of such an endeavor is huge, and therefore it must wait for now.
Small adjustments made in the meanwhile, if TB feels so inclined to look at that code, is cool and all; but it will only be a treatment, not a cure.
Well, that's my opinion on the matter at least. ^^

 

[Thunderbrd]

But there is no way an AI should want 89 units (and they are not low level units for the time frame and tech level of the game) for a 12501 tech. Even if they are in an Open Borders agreement with you.

Why not? The AI should love to get such a deal if you're willing to do it. What they bid is not always the best deal they are willing to accept either. You can sometimes take some of what you're offering off the table and get them to agree to something more fair. Diplomacy status makes a difference. Are they friendly or do they just like you enough to be willing to negotiate? At least he knows what he's looking for. And the opportunity to catch up on tech is worth far more than units you might not need. Apparently he at least knows what you see is true, that he doesn't have as much of what he needs as he should.

When you don't have a tech, they are worth a lot more than their research cost because you don't have the benefits they bring and the launch points towards the techs beyond them. Any AI willing to trade a technology is selling off their edge, something that its very rare that units alone can contribute as much towards. Theoretically you'd think it should be that research compares the same as any other yield or commerce but the reality is that it doesn't. Research is strategically worth nearly twice as much as even a normal yield (as long as you've met your needs in other areas at least.) This is because it can often mean getting a better income OF all other yields and commerces.

In a competitive game against AI alone, with technologies, I strive to pull ahead and then NEVER trade a tech. In fact, the value of a tech is only barely based on its beaker cost for an AI. Far more important to them is the strategic value of the tech specifically to them and to you and to them to keep you from having it but then taking into account any other AI's and the likelihood that they would trade that tech to you and be the one to benefit from it.

All human games are much harder to negotiate where tech trading is concerned. If you don't let other human players get in on trading agreements, they will put you out in the cold and a group of them will form a trade conglomerate between them until you are left in the dust and have nothing to offer anyhow. And usually as a part of this kind of agreement, the value of the tech or who gives more or less is much less important than the value of the alliance.

So a truly intelligent AI doesn't only look at rates of one thing vs another. And our AI players do look at things quite deeply. Which is why its usually best to let them consider all they do and then have a dial we can twist to adjust final results. (Because sometimes math might look at the right ratios of concern, but not always be good at bringing it back to a good balanced result.)

 

[JosEPh_II]

Why not? The AI should love to get such a deal if you're willing to do it. What they bid is not always the best deal they are willing to accept either. You can sometimes take some of what you're offering off the table and get them to agree to something more fair. Diplomacy status makes a difference. Are they friendly or do they just like you enough to be willing to negotiate? At least he knows what he's looking for. And the opportunity to catch up on tech is worth far more than units you might not need. Apparently he at least knows what you see is true, that he doesn't have as much of what he needs as he should.

When you don't have a tech, they are worth a lot more than their research cost because you don't have the benefits they bring and the launch points towards the techs beyond them. Any AI willing to trade a technology is selling off their edge, something that its very rare that units alone can contribute as much towards. Theoretically you'd think it should be that research compares the same as any other yield or commerce but the reality is that it doesn't. Research is strategically worth nearly twice as much as even a normal yield (as long as you've met your needs in other areas at least.) This is because it can often mean getting a better income OF all other yields and commerces.

In a competitive game against AI alone, with technologies, I strive to pull ahead and then NEVER trade a tech. In fact, the value of a tech is only barely based on its beaker cost for an AI. Far more important to them is the strategic value of the tech specifically to them and to you and to them to keep you from having it but then taking into account any other AI's and the likelihood that they would trade that tech to you and be the one to benefit from it.

All human games are much harder to negotiate where tech trading is concerned. If you don't let other human players get in on trading agreements, they will put you out in the cold and a group of them will form a trade conglomerate between them until you are left in the dust and have nothing to offer anyhow. And usually as a part of this kind of agreement, the value of the tech or who gives more or less is much less important than the value of the alliance.

So a truly intelligent AI doesn't only look at rates of one thing vs another. And our AI players do look at things quite deeply. Which is why its usually best to let them consider all they do and then have a dial we can twist to adjust final results. (Because sometimes math might look at the right ratios of concern, but not always be good at bringing it back to a good balanced result.)

Sorry T-brd that was a rabbit trail post.

If that's the way you want it and there is not an xml value to adjust, then I've stated my case. And since I can't code thenthat is the way it will be till a player complains.

 

[Thunderbrd]

Do you mean it was just too long? I was trying to explain how the code works and why it should remain working that way. There are a TON of factors the AI takes into account and should and it was all well thought out a long long long time before I ever got here or C2C was a thing.

There IS a final xml value to adjust a % modifier to the final result. At least for worker trade values. I'm just saying that unless you take a deep look at why an evaluation comes up with a particular value for a thing in the game, you may not, as a player, have a clue what's being taken into consideration.

I know this "Far more important to them is the strategic value of the tech specifically to them and to you and to them to keep you from having it but then taking into account any other AI's and the likelihood that they would trade that tech to you and be the one to benefit from it." sucks as a sentence. Let me try again:

It matters what the tech will bring to you, not just the cost of the tech.
It matters what the tech will bring to them, not just the cost of the tech.
It matters if they can force you to have to research it or if another AI player has it and may sell it to you anyways. (Are they the only civ with this tech?)
It matters how far behind you or they are.
I could go on and on with considerations that have been programmed into the AI's determination.

As for units, they are never going to give a fair deal for units and they shouldn't. Unit production modifiers are common in the game. Units can be drafted, slaved for, and can usually be summoned up quickly by any player who really wants to get them fast. If a Player is short units, trading is one often substandard way to get them in a hurry.

If you don't like a deal, you're probably smart and the game is running intelligently. The AI is not supposed to be cutting fair deals because the concept of fairness is, in their appropriate view for any negotiator, a matter not of solid value but a matter of what you need from your gameplay imbalances, vs what they need from their gameplay imbalances.

Anyhow, I'll stop before I go on trying to explain what is too complex for some to care to follow. This is not my design, just one that makes sense as I read it.

The global adjustment for worker end results: WORKER_TRADE_VALUE_PERCENT_ADJUSTMENT in GlobalDefines.xml I can add another for normal units but again, it seems to me that the AI is spot on with its valuation. I'm actually surprised anyone would think that any amount of units should be able to be traded for a tech at all but I guess if the tech is a LOT more valuable than the production costs of the units then it might be kinda almost fair in some cases.