Dune Wars 1.9.4 Patch Feedback

Can I ask which these were?
There are a few which were deliberately expensive given how powerful they are for their tech cost, to avoid being overwhelming (most particularly the quad; quadspam can be devastating in pillage-choking an enemy, so its important that quads are expensive to build).

I know about the quads, don't worry they are still nice, beautiful, and expensive. These were only a couple of cases where a unit and it's upgrade were left the same cost. For instance the eagle thopter and buzzard thopter both had the same production cost, same with rocket artillery and the much more powerful missile launcher. I think there might have been one other pair like this, but I don't recall off the top of my head. An upgrade unit should always cost at least a little more than the previous unit, and I could only see this as something missed before? Those kind of things aren't uncommon when large scale XML changes are made to lots of units. If there was a specific reason for either of the cases above, let me know, but it seemed unlikely.

A related thing a saw that might also need to be changed, though I didn't touch it yet (might have been intentional) is the bombard damage of the rocket artillery/howitzer (Tier2 units). Both are 8%, same as Maula Mortars, while the missile launcher jumps up to 18%, then jonly up to 22% for the assault cannon. Intentional or do you know?

What was the idea behind this? First strikes are already pretty powerful on a collateral damage unit.

Not really. First strikes are a funny thing, because their usefulness is highly dependent on the combat ratio of the two units involved. When you already are much more powerful than who you are fighting, they have a larger impact, and where the other guy is more powerful, their impact is lessened. I covered some of the math about this using a few extreme cases in this post. This makes them hard to balance since they can be both overpowered and underpowered at the same time. However, since taking first strikes usually (though not with siege since they lack combat promos, etc.) means giving up more powerful +%combat promotions, the drill line has always been considerably weaker in 95% of cases than nearly any other option.

Siege units are all too often involved in attacking well defended cities where, unless you have a great tech lead, you are fighting at a combat ratio disadvantage. The traditional drill line gave so very little addition to survival chances in such cases, you were always better off taking barrage (at least do more collateral before dying/retreating) or accuracy (never attack with them, just take down the city defenses), drill just wasn't a good third option since it added so little to your odds. What I've mostly done is to condense the normal 4 promotions down to 3 so each is a little more potent and actually will be useful, but the total available remains mostly constant. If you really thought they were useful enough to take before, then I apologize, we disagree :). Of course if it's needs more balancing, I'm all ears on what people experience in game, particularly if it becomes a dominant option compared to the other two promotion lines.

Interesting to test.

Yep, this is of course coming from the discussion on the Organized trait. Hopefully I didn't overdue it, it's harder to test late game mechanics as I'm always having to start new games when developing. But my guess from the usual late game surpluses is it'll play out alright. And give more usefulness to the poor organized trait.

Which design did you go with here?

It's for now the extension of David's original system as it'll take some work to put in what we've discussed. That's slated for the next patch.
 
These were only a couple of cases where a unit and it's upgrade were left the same cost.
Ah, ok. These sound like good rational changse. There are probably lots of messy things left from various times where we shuffled around these units in the tree (or their art). Thanks.

First strikes are a funny thing, because their usefulness is highly dependent on the combat ratio of the two units involved.
Right, but my point is that there is some synergy with collateral damage, no?
With low strength units, first strikes are less useful in *how they alter the probability of winning the fight*. But with collateral damage units, the goal is not to win the fight, its to inflict the maximum amount of damage in the stack.
With collateral damage units, what I really want is to get a few combat round wins in, even if I eventually lose. I admit I haven't done the math, but it feels like this would mean that the value of first strikes to a collateral damage unit would be higher than it would be for a regular unit, because for a low-strength unit the best chance they have of getting a few rounds of wins in is early in the fight, before their strength is too degraded.

The other thing to consider is that if you have first strikes and I have first strikes, then they cancel out. Siege units are very often attacking cities, defended by guardsmen units with first strikes. So just as first strikes are less useful for low strength units by themself, they are more useful for high strength units by themself. Adding first strikes to the siege unit then is effectively the same as removing first strikes from the high strength unit - so imagine its strength 8 vs strength 12 with 2 first strikes, adding a first strike to the strength 8 unit has the same marginal impact as removing a first strike from the strength 12 unit. Unless I'm misunderstanding something?

Anyway, I'm happy to tweak and test I don't feel strongly that the status quo is balanced right.

It's for now the extension of David's original system
That's what I figured, just wanted to check.
I will see if I can find some time this weekend to do some testing.
 
because for a low-strength unit the best chance they have of getting a few rounds of wins in is early in the fight, before their strength is too degraded.

Not terribly important, but I wanted to point to this before getting to the real discussion, a unit's strength doesn't actually degrade during a fight. So while a unit that starts a fight damaged starts usually at a great disadvantage, the damage taken during a fight doesn't have the same effect because roundDamage and defenderOdds are constants after the fight starts (By the way, Advanced Combat Odds shows you both of these numbers for attacker/defender along with needed rounds for each for those who don't know).

Right, but my point is that there is some synergy with collateral damage, no?
With low strength units, first strikes are less useful in *how they alter the probability of winning the fight*. But with collateral damage units, the goal is not to win the fight, its to inflict the maximum amount of damage in the stack.

This actually points to why I wanted to make this change. There are two separate strategies that could be possible, emphasizing collateral damage with barrage promotions at the cost of survivability, or emphasizing survivability at the cost of collateral damage (where the accuracy promotions like I said emphasize avoiding combat). The promotions as originally set up weren't well balanced to allow a survivability emphasis. If the unit was roughly as likely to die taking drill or barrage, it was always better to take barrage and at least take more of the bastards down with you :lol:. It just can't be made too powerful where surviving becomes a given, as that would push players from ever taking barrage since it's the naturally riskier promotion line.

Depending on play testing (my next test game I want to try a siege heavy invasion strategy to test things out better myself) it might be good to reduce the effectiveness of collateral damage without barrage promotions to give greater emphasis to this choice of survivability or mass collateral damage.

Just as a side note, I'm thinking the sudden jump in bombardment damage between the Tier 2 and 3 siege weapons is probably justified because that's about when ablative shields will start to show up. If the Tier 2 units had a higher damage rating, it would take defenses down to quickly in the pre-shield era and it really makes the higher tier units necessary against a shielded city.

edit: Also wanted to point out that the amount of damage inflicted by the siege unit on the actual 'defending' unit has no impact on the collateral damage it inflicts on the other units in the stack. So a unit with first strikes that does more damage on average to the 'defender' doesn't do more collateral as a result. I imagine a lot of players don't realize this lack of relationship there.
 
but I wanted to point to this before getting to the real discussion, a unit's strength doesn't actually degrade during a fight.
Huh, ok, I misunderstood the combat mechanics then. My bad.

This actually points to why I wanted to make this change. There are two separate strategies that could be possible, emphasizing collateral damage with barrage promotions at the cost of survivability, or emphasizing survivability at the cost of collateral damage (where the accuracy promotions like I said emphasize avoiding combat). The promotions as originally set up weren't well balanced to allow a survivability emphasis. If the unit was roughly as likely to die taking drill or barrage, it was always better to take barrage and at least take more of the bastards down with you
This sounds sensible. I defer then to your superior understanding of the details of the combat mechanics.

The one think I want to keep away from though is making the siege units too good. In general, I think that siege units in Dune Wars should be weaker than they are in vanilla. Siege units are very un-thematic and uncanon in the Dune Universe. In the real Duneiverse, they are obsolete because of shields, and the only time they're used, they're really used against a geographic feature (to collapse cave entrances) rather than against the enemy. Whereas in vanilla Civ4/BTS, siege units are really key to stack combat.

So, we leave siege units in the mod there because we need the mechanic of reducing city defenses, and because we leave the collateral because we want them to be at least somewhat useful, but the overall design goal should have them as less useful than vanilla, and that the main city attack unit should be melee class, not siege.

We also don't have the mounted unit/flank attack mechanic going on, so we don't have a direct siege unit "counter".

Just as a side note, I'm thinking the sudden jump in bombardment damage between the Tier 2 and 3 siege weapons is probably justified because that's about when ablative shields will start to show up.
Agreed, but I think that force shields *should* be fairly hard to knock down even with same era siege units. It should sometimes be worth just swarming the defenders with melee (who ignore the shield) rather than besieging to take the defenses down.
 
Also wanted to point out that the amount of damage inflicted by the siege unit on the actual 'defending' unit has no impact on the collateral damage it inflicts on the other units in the stack. So a unit with first strikes that does more damage on average to the 'defender' doesn't do more collateral as a result.
Just to clarify here: my impression wasn't that the amount of damage mattered, but the number of combat rounds you won did matter.

So for example; suppose in normal fighting, no first strikes, that my unit with collateral damage win 2 combat rounds on average. Suppose that we add a first strike round, which I have a 40% chance of winning. Then the expected number of rounds I win is 2.4. Doesn't this mean I do more damage, on average, to units that are not the primary defender?

Or am I misunderstanding collateral damage again?
 
Huh, ok, I misunderstood the combat mechanics then. My bad.

It's OK, I'm pretty sure about 99.5% of players lack any idea of how the combat is calculated, and thats OK as long as the odds shown in game are accurate (which they are). The only reason I know all this is because I've studied and modified the combat code in some detail before.

So, we leave siege units in the mod there because we need the mechanic of reducing city defenses, and because we leave the collateral because we want them to be at least somewhat useful, but the overall design goal should have them as less useful than vanilla, and that the main city attack unit should be melee class, not siege.

I'm all for that too, and it works well as such. The key is that melee units are more effective at city attack even against well defended cities in DuneWars so that you can win battles without lowering the defenses with well promoted units unlike BTS where in most cases lowering the defenses is a little more mandatory. Though Siege units do afford an opportunity for a different play style/strategy which is good.

Agreed, but I think that force shields *should* be fairly hard to knock down even with same era siege units. It should sometimes be worth just swarming the defenders with melee (who ignore the shield) rather than besieging to take the defenses down.

They are (I haven't changed any thing there by the way), I've done it before with multiple siege units with accuracy promotions and it still takes several turns on epic. Which means if you want to take down the shields in a decent amount of time, you can't really take the drill promos, specialization at it's finest :)

Just to clarify here: my impression wasn't that the amount of damage mattered, but the number of combat rounds you won did matter.

Nope, not how it works. You don't have to do ANY damage to the actual 'defender' unit to inflict collateral damage, which also is completely independent of the battle result entirely. Instead the damage inflicted involves a direct comparison between the (BaseStrength * CollateralDamagePercent) of the attacker to the (BaseStrength) of the defender modified by the GlobalDefine value for Collateral Combat Damage. It's even applied in the code before the first round of the actual battle.

Since it uses "BaseStrength" instead of "MaxStrength" or "CurrStrength", it also ignores any modifiers or current health in the calculated damage, something I've changed in personal mods in the past. A catapult that has 1hp as it attacks a stack will indeed do just as much collateral damage as a full strength catapult in BTS, kinda hooky, damned lazy programmers. I've actually added some code to make the damage and number of units hit a little more variable (no real impact on gameplay it's so minor, just makes the numbers a little more varied for better immersion), but I hadn't planned any other changes from vanilla behavior without any demand for it.
 
By the way, I know Deliverator in some older thread talked about using an FfH style building based mechanism for removing certain negative promos like 'plagued' and 'poorly maintained', but nothing ever came of it. For now I've got no problems with the way plague works, but I do think that poorly maintained should disappear as you get the technology to repair the equipment. It doesn't make good sense to a player that I can build new ones, but can't fix these things I found in the desert up.

This doesn't need to be complicated at all. If you get a 'poorly maintained' thopter, then after researching 'Combat Ornithopers', spending a turn in one of your cities should remove it. (I don't think it should require any buildings for the record, you don't need a factory to build a new ornithopter, you sure don't need one to repair one)

Any objections?
 
Huh. So a heavily damaged collateral unit that never wins a combat round will deal the same damage, every time, to units in the rest of the stack as a the same unit would if it were full health and won the fight?

How bizarre. Thanks for explaining.

It sounds like we're in agreement on design.
 
By the way, I know Deliverator in some older thread talked about using an FfH style building based mechanism for removing certain negative promos like 'plagued' and 'poorly maintained', but nothing ever came of it. For now I've got no problems with the way plague works, but I do think that poorly maintained should disappear as you get the technology to repair the equipment. It doesn't make good sense to a player that I can build new ones, but can't fix these things I found in the desert up.

This doesn't need to be complicated at all. If you get a 'poorly maintained' thopter, then after researching 'Combat Ornithopers', spending a turn in one of your cities should remove it. (I don't think it should require any buildings for the record, you don't need a factory to build a new ornithopter, you sure don't need one to repair one)

From memory, the current design is that the hospital removes plagued and the factory removes poorly maintained.

I don't mind dropping the factory requirement for poorly maintained and moving it to tech, but I'd prefer to keep the hospital requirement. Tleilaxu plague isn't nearly as bad as it was, since we stopped it from affecting vehicles and aircraft. I'm not sure if its balanced yet though, it might still spread a bit too fast.
 
That's how I remember the discussion, but I just checked over the code to see what was done and not done. The clinic and suk hospitals do remove plague, but there is currently nothing coded concerning 'poorly maintained'. Like I said, the plague promotion works fine to me, though I haven't payed close attention to it's spread rate yet. Unless I come across any reason to, I'm not rushing to make any changes there.

I'm about to test some new code for removing the 'poorly maintained' promo.
 
The clinic and suk hospitals do remove plague
I couldn't remember that exactly in my last game - so built all the health buildings to get rid of the dreaded TP. It is a small detail, but could that info be added to the Dune Encyclopedia at some point? It would help newer players (and older drunken ones too). :)
 
I couldn't remember that exactly in my last game - so built all the health buildings to get rid of the dreaded TP. It is a small detail, but could that info be added to the Dune Encyclopedia at some point? It would help newer players (and older drunken ones too).

Yes it can, especially for the druken ones, I never know when that might be me :lol:.
I'll have to write some text for some of the new concepts being/already added anyway for the next patch.
 
Just a quick status report, I've got the code to remove poorly maintaned promo working as proposed, including a message to let you know when it's removed. Just to recap, if you got a quad/trike from a goody hut, as soon as you research that unit's prereq tech, so light manufacturing in this case, spending a turn in any of your cities will remove the promo.

While I was working on that I updated the message given when sandstorms destroy improvements and units. They had no sound or icon shown on the map before and I found myself never knowing it'd happened 90% of the time as the message got lost in other messages. It'll now give sound and show an icon over the location so you've got a better heads up. Same thing when worms destroy spice harvesters.

I've also got the Ordos ability to steal some technology when capturing cities working. It can only steal from techs the enemy has that you can currently research (as in, you need all the prereqs), and can steal anywhere from about 10-30% of the cost of a tech based on city size and, more loosely, research capacity, and can steal research in 1-3 techs. How many techs is of course limited by how many they have that you can actually research, but if they do have multiple techs to steal from, the likelihood of stealing from more than one is modified by city size, so major cities are both more likely to give more techs and to give more research in the tech. I think that makes good sense.

Obviously, this means you benefit more from attacking more advanced enemies :).
 
I am pretty worried that these neat new effects will widen the gap between the AI players and the human player. Although a human player may eventually realize that moving a poorly maintained unit to a city is a good idea, the AI never will. This has driven a lot of the previous decisions about designing features. I was never able to understand general unit AI well enough to make this type of modification. Certain special units, such as worms, have unique AI, which is much easier to implement.

Do you think it is important for the AI to understand these features, or do you understand the unit AI well enough to have the AI move poorly maintained units to cities?
 
There are so few 'poorly maintained' units in the game that the effect on the AI is almost nil, but the ai will almost certainly by sheer accident have the unit in a city at least one turn, even if just passing through it's territory or wandering around waiting for something to happen, so that's not even a worry. And if they don't, that one unit in the hundreds they will build won't matter.

For the human, it may be one unit, but the player can sit there and think, 'Why the hell am I stuck with this promotion, there oughta be some condition it goes away at, right - <click>World Builder</click>'. It's more of a psychological issue than a game play one as such. In such a case it's much more important, and good design, to give the player a logical way to deal with something where common sense says there should be, so long as it doesn't hamstring the AI.

I do pretty well always consider how the AI will use something, assuming it has a large enough gameplay impact for it to be important for the AI to use it (intentionally), when designing and thinking out new/altered gameplay elements. For example, right now I'm playing with a couple of changes in my working copy balancing out the siege promotions/units a little better, but knowing how the AI will use siege units, I'm also balancing them to play to it's playstyle. Additionally, I'm also double checking the potential AI values assigned to each of the possible promotions for both siege unit_AI's (COLLATERAL and ATTACK_CITY) to make sure it'll pick the appropriate promotions given the new values and for the correct unit_AI.

I'm of course not saying you shouldn't point out these kinds of things, I can't think of everything, and sometimes I miss something I should have thought of, but I do want you to know it's always part of what I'm thinking about.

or do you understand the unit AI well enough to have the AI move poorly maintained units to cities?

I could do it with a little more study of existing AI code, and a few hours more of tinkering and testing, but it'd be way more of a pain than what it'd actually be worth. I've got a fair certainty the AI will do it unintentionally on it's own in a vast majority of cases. The actual easier and more reliable approach is to define that the AI only needs the unit to be in it's territory. (Come to think of it, it wouldn't hurt anything to let the player just need to be in his territory instead of a city too, saves the micromanagement for a minor and logical effect)

As a side note, there are two particular unit AI issues I would like to tackle for upcoming patches. One is the well know all-terrain transport issue of not dropping the troops off when in danger. The other thing I'd like to do is a general rewriting of the espionage unit AI. Both of these if fixed/improved could vastly improve the game.
 
As a side note, there are two particular unit AI issues I would like to tackle for upcoming patches. One is the well know all-terrain transport issue of not dropping the troops off when in danger. The other thing I'd like to do is a general rewriting of the espionage unit AI. Both of these if fixed/improved could vastly improve the game.

On the drop off problem, you can probably find the post from Sephi last year where he proposed some code, and you can also see some comments in the sdk files where I added this then removed it because sometimes it caused a hang. It doesn't "seem" that difficult, but there must be something we don't quite understand.

On the espionage AI, I completely agree. I think that looking through the weights for the missions and adjusting them "should be" enough. I never added a weight for the convert city mission, but it was on my to-do list. It will be a success for this feature, the first time a player writes an angry message here to complain that the AI converted one of their cities. I don't think this has ever happened yet.
 
I've also got the Ordos ability to steal some technology when capturing cities working. It can only steal from techs the enemy has that you can currently research (as in, you need all the prereqs), and can steal anywhere from about 10-30% of the cost of a tech based on city size and, more loosely, research capacity, and can steal research in 1-3 techs
I think I would change this so that it only steals for one tech, and I would increase the upper bound of the amount stolen (and make it very more on city size). If you get a very large city (eg size 20), it should give you a free tech, or almost. So maybe 10-90%.

As to multiple techs; this could get annoying and confusing. I think a more focused effect (lots of research on a single tech) would feel more significant to the player than a diffuse effect (little bit of research on multiple techs).
Also; does Civ4 have tech progress degradation (where progress in techs you aren't currently researching just declines) or is that just in city production? I forget.

So, I would just give them X% of the most expensive tech they have that you have all the pre-reqs for. So there is still incentive to go against someone more advanced than you, but no particular issue about whether they have 1 tech more than you or 3 techs more than you.

I am pretty worried that these neat new effects will widen the gap between the AI players and the human player. Although a human player may eventually realize that moving a poorly maintained unit to a city is a good idea, the AI never will.
I think the AI is highly likely to eventually the unit back to a city. So there is actually a lower AI-human gap where all one has to do is return to a city, as opposed to returning to a city with a factory, which was the original design goal.
So I agree with Chris here.

There are so few 'poorly maintained' units in the game that the effect on the AI is almost nil
We did have player feedback that sometimes a game was too easy and was ruined because they popped a vehicle early on and used that to annihilate an enemy. So its rare, but early in the game a single strength 5 unit can have a big impact.

As a side note, there are two particular unit AI issues I would like to tackle for upcoming patches. One is the well know all-terrain transport issue of not dropping the troops off when in danger. The other thing I'd like to do is a general rewriting of the espionage unit AI. Both of these if fixed/improved could vastly improve the game.
This would be fantastic.
 
On the drop off problem, you can probably find the post from Sephi last year where he proposed some code, and you can also see some comments in the sdk files where I added this then removed it because sometimes it caused a hang. It doesn't "seem" that difficult, but there must be something we don't quite understand.

I know from reading around the forums that a large majority of infinite loops are the result of grouping/ungrouping actions and of course... cargo. If nothing else we can add in a safety valve of sorts, an incremented variable that if it hits it's cap exits from the loop, but I'd like to really find the root of the problem. I didn't find any comments in CvUnitAI from you or sephi related to it though, so it might have gotten shuffled out in the BetterBugAI version we're built on now? I've got an old 1.9 installer too so I'll check that, any idea what function I'm looking for?

It will be a success for this feature, the first time a player writes an angry message here to complain that the AI converted one of their cities. I don't think this has ever happened yet.

It won't happen without any AI value added to it, but I'm still on the fence wether this should be used against the player. Most espionage missions are all fun and games, but this... I don't know. I might just only allow it AI vs AI.

I think I would change this so that it only steals for one tech, and I would increase the upper bound of the amount stolen (and make it very more on city size). If you get a very large city (eg size 20), it should give you a free tech, or almost. So maybe 10-90%.

As to multiple techs; this could get annoying and confusing. I think a more focused effect (lots of research on a single tech) would feel more significant to the player than a diffuse effect (little bit of research on multiple techs).
Also; does Civ4 have tech progress degradation (where progress in techs you aren't currently researching just declines) or is that just in city production? I forget.

So, I would just give them X% of the most expensive tech they have that you have all the pre-reqs for. So there is still incentive to go against someone more advanced than you, but no particular issue about whether they have 1 tech more than you or 3 techs more than you.

I agree, and already done. Though for now it's still a random tech from the available ones, I'm not sure if it's a good idea to always give the most expensive, might just weight the cost into the random assignment.

We did have player feedback that sometimes a game was too easy and was ruined because they popped a vehicle early on and used that to annihilate an enemy. So its rare, but early in the game a single strength 5 unit can have a big impact.

I know all about that. With the system in place, you'll still have to rearrange your research priorities if you want to get a popped unit up to full strength early on, and bring it back to your territory, so I'm not expecting any balance issues.
 
The AI's getting smarter...

He he, I'm actually a little scared of what I accomplished today. I against better judgement spent the better part of the day pouring over and getting familiar with parts of the AI code and playing several test games in debug mode noting AI behaviors to fix and improve. And I didn't even wind up touching either of the items I mentioned the other day, instead my recent attempts to try out the new Ordos technology by conquest ability really highlighted to me just how bad the AI was at keeping up, so I started studying why. They often would utterly destroy their economy early on and then never recover.

I won't go into too many specific details, but I think players will be very surprised when they play the next patch how much better the AI keeps up technologically. The primary problem really was that the AI was over expanding without properly considering it's economic position, and I've helped them out with that with new checks on when to build settlers and what to emphasize when the budget gets tight.

What I'm maybe most excited about is that I've fixed several important flaws in the worker AI. The old code contained logic that was a hangover from the vanilla idea that land and sea workers are different things altogether so that a worker had to first almost accidentally end up in a desert tile before even considering working spice tiles. They were also using the default logic that any extra spice resources after the first one was hooked up were worth less. What I've done is tied the value of spice resources to the economic position of the civ (so the weaker their economy, the more workers seek out spice tiles) and they now aggressively develop their spice fields (if they have them) even in the early game to help support their expansion and research levels.

I've also rebalanced how several technologies are valued. If you've played very many games, you should know that the AI went after Dune Topography like a pack of rabid wolves, neglecting most other technologies in the process, while Spice Extraction was often an after thought to them. Again this was due to various facets of the old vanilla logic. Dune Topography is now much better valued, and the AI will specifically look at how many spice resources it has in it's borders when considering Spice Extraction.

I'm also doing some additional testing on some changes to how much of their budget AI's will spend on units. The general idea is I'm scaling back the unit costs slightly in the handicap infos(for the higher difficulties) while at the same time reducing what percentage of their budget they're willing to spend on units, so they should end up with similar sized armies as now, but without cutting into their economy near as much.

I've also addressed several smaller but still important issues I found in the code. On example is some of the BBAI code that siphoned off research% for espionage%. It actually does a great job of evaluating how much is needed for spending against opponents, but it didn't originally look at the economic position, so AIs that were already at very low research levels in the early game(due to some of the issues mentioned above) were also shifting what little research they were doing into espionage.

Anyway, long story short, you might really need to drop a difficulty level with the next patch, or play harder, you're choice :D.
 
It won't happen without any AI value added to it, but I'm still on the fence wether this should be used against the player. Most espionage missions are all fun and games, but this... I don't know.
I agree this is a risk, but I think I'd try including it. The mission is very expensive, I doubt the AI will ever really use it, and if they do, then it will be because you haven't been investing espionage against them.

I'm not sure if it's a good idea to always give the most expensive, might just weight the cost into the random assignment.
My worry about not doing this is that potentially the fact that they have more extra techs than you would decrease the value of what you get from them.

Suppose you have a system where it picks a random tech.
Suppose they have two techs you don't have, with costs of 500 and 800.
You capture a city that gives 50% of a tech; the expected beaker yield is 325.

Now suppose that they have three techs you don't have, with costs of 100, 500, and 800.
You capture a city that gives 50% of a tech; the expected beaker yield is 231. So, by adding another tech, you are now worse off.

The primary problem really was that the AI was over expanding without properly considering it's economic position, and I've helped them out with that with new checks on when to build settlers and what to emphasize when the budget gets tight.
Interesting. I have generally found massive settler spam to be a reasonable strat in Dune Wars, at least relative to vanilla. Getting the good city spots is important.

What I'm maybe most excited about is that I've fixed several important flaws in the worker AI.
This sounds very promising. Do they actively build more workers, or just send the workers towards spice tiles?

I've also rebalanced how several technologies are valued.
This sounds promising too.

The general idea is I'm scaling back the unit costs slightly in the handicap infos(for the higher difficulties) while at the same time reducing what percentage of their budget they're willing to spend on units
Interesting. I don't mind making the game a bit tougher, its pretty easy on all by the very highest difficulties.

Great job. I have huge respect for anyone improving the AI code (I am no coder.....).
 
Top Bottom