-
We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.
Originally posted by The Federalist
Just wanted to mention:
The Germany-only 1939 scenario is now ready for download. Just click on the link in my signature
Is this for 1.03?
Latest patch? I patched 3R to 1.38 and there was no AI whatsoever.Originally posted by Munchkin
Nah, it isn´t. With the latest patch it´s one of the best.
(For example, once I Invaded Soviet Union via Turkey. AI didn´t react when I had the bulk of my panzers sitting on the border, so all it took was "empty hex exploitation" (not an exploit, mind You) and SU collapsed in one turn.)
So, if there is a newer patch somewhere, please let me know!
This has been a great thread with some excellent posts.
Those posted by Catholic Dog are, IMO, some of the best I have read.
I wanted to add some things that I have been covered in other threads, but I believe will help the AI respond.
#1- Movement rates need to be fixed. It is absolutely impossible for divisions to move that many miles a day. Movement rates should be lowered AND they should be tied to ORG rates. Lets say your max ORG rate is 70---if your ORG is at 70 (or 100%), then your division can move at 100% of its movement allowance. If your ORG is at 35 (or 50%), then your division can move at 50% of your movement allowance.
This will have the added effect of "slowing down" the game and making battles and movement more realistic.
#2- Supply. Supply rates for each unit should increase while in an enemy held territory-whether moving or in battle. If in battle, they should increase drastically. IMO, key provinces should be designated as "supply centers" and the supply rate should be calculated for each unit based upon its distance.
#3- ORG. ORG levels should affect movement (as described above) and should decrease based upon your rate of supplies. If German a unit is on the steppes of Russia, there is no way in hell that this unit should have 100% supply. If a unit doesnt have 100% supply, it shouldnt have 100% org.
ORG levels should also decrease through movement. When you move from one province and attack an enemy controlled province, your units should lose ORG as they move into attack. Also there should be an ORG loss for simple movement in your own controlled provinces. It doesnt make sense that a unit can move 2 provinces in a couple of weeks and have 100% of ORG Rate available.
Additionally, the rate of ORG recovery should be decreased. IF you have gotten beat bad or even won a victory, you should have to hold your unit back to regroup. Even in Barbarossa, the Germans had to hold entire corps back for a few weeks to reorg, reinforce and resupply.
#4-Garrison forces----You should have to maintain a garrison force in each conquered province to ensure supply lines (and hopefully in 1.04, prevent partisans). It absolutely absurd that you see posted screenshots of the eastern front with all the German forces in 4-6 provinces. In the conquered territory behind them, there are absolutely no garrisoned provinces. IF this doesnt happen, supply should be effected by 50% minimum. If you start garrison your provinces the further you go, you will see additional probs/speedbumps.
I think these 4 major points will go along way to "slowing" the game down for the AI. The game moves entirely too fast with very little penalties for outlandish behavior.
Hey if you want to take your light tanks or paras and exploit gamey tactics go ahead. However the game should have penalties for this. ORG rate loss for paras should be double or triple regular infantry. IF your light tanks get too far ahead, they will suffer from decreased oil and supplies----which will have a tangible effect on your ORG rate and combat effectiveness. Granted you may take that extra province, but when you have an ORG rate of 20% getting attacked by infantry with an ORG rate of 100%, you are gonna get whipped!
The most important thing, IMO, is that you should have the freedom to invade whoever, where-ever, whenever, etc. Hell you can even send you troops on half-baked offensives. So long as the game makes you "pay" the consequences is all that I ask.
Those posted by Catholic Dog are, IMO, some of the best I have read.
I wanted to add some things that I have been covered in other threads, but I believe will help the AI respond.
#1- Movement rates need to be fixed. It is absolutely impossible for divisions to move that many miles a day. Movement rates should be lowered AND they should be tied to ORG rates. Lets say your max ORG rate is 70---if your ORG is at 70 (or 100%), then your division can move at 100% of its movement allowance. If your ORG is at 35 (or 50%), then your division can move at 50% of your movement allowance.
This will have the added effect of "slowing down" the game and making battles and movement more realistic.
#2- Supply. Supply rates for each unit should increase while in an enemy held territory-whether moving or in battle. If in battle, they should increase drastically. IMO, key provinces should be designated as "supply centers" and the supply rate should be calculated for each unit based upon its distance.
#3- ORG. ORG levels should affect movement (as described above) and should decrease based upon your rate of supplies. If German a unit is on the steppes of Russia, there is no way in hell that this unit should have 100% supply. If a unit doesnt have 100% supply, it shouldnt have 100% org.
ORG levels should also decrease through movement. When you move from one province and attack an enemy controlled province, your units should lose ORG as they move into attack. Also there should be an ORG loss for simple movement in your own controlled provinces. It doesnt make sense that a unit can move 2 provinces in a couple of weeks and have 100% of ORG Rate available.
Additionally, the rate of ORG recovery should be decreased. IF you have gotten beat bad or even won a victory, you should have to hold your unit back to regroup. Even in Barbarossa, the Germans had to hold entire corps back for a few weeks to reorg, reinforce and resupply.
#4-Garrison forces----You should have to maintain a garrison force in each conquered province to ensure supply lines (and hopefully in 1.04, prevent partisans). It absolutely absurd that you see posted screenshots of the eastern front with all the German forces in 4-6 provinces. In the conquered territory behind them, there are absolutely no garrisoned provinces. IF this doesnt happen, supply should be effected by 50% minimum. If you start garrison your provinces the further you go, you will see additional probs/speedbumps.
I think these 4 major points will go along way to "slowing" the game down for the AI. The game moves entirely too fast with very little penalties for outlandish behavior.
Hey if you want to take your light tanks or paras and exploit gamey tactics go ahead. However the game should have penalties for this. ORG rate loss for paras should be double or triple regular infantry. IF your light tanks get too far ahead, they will suffer from decreased oil and supplies----which will have a tangible effect on your ORG rate and combat effectiveness. Granted you may take that extra province, but when you have an ORG rate of 20% getting attacked by infantry with an ORG rate of 100%, you are gonna get whipped!
The most important thing, IMO, is that you should have the freedom to invade whoever, where-ever, whenever, etc. Hell you can even send you troops on half-baked offensives. So long as the game makes you "pay" the consequences is all that I ask.
Slowing down the game will not help the AI, the computer can think faster then we can. If you don't believe it play the game on extremely fast without pause, and I am quite willing to bet you will lose, and badly.
Not true, you just need to re-think what a 'move' is.
Not true, you just need to re-think what a 'move' is.
Last edited:
Jack--
Sorry mate, you are wrong on this account. There is a difference between responding to information quickly and correctly. I have no doubt that on the FAST speed, the computer can make quicker decisions. Whether those decisions are correct or not still remains to be seen.
The changes I have recommended will drastically affect gameplay for humans. It will slow down the obscene and totally unrealistic games people have been playing. Granted this will effect the computer also, but all that requires is a code change to fix it. As for humans, you will no longer see people exploit paras, or sending light tanks on 500 km spearheads without stop for resupply/reinfor/reorg. The computer doesnt do this stuff anyway, so it wont be affected. It will affect the human player and balance the playing field. In that way it will help the AI.
Plus it makes the game a hell of alot more realistic. Dumb AI is DUMB AI. However even with bad decisions on the AI's part, I still believe it "impossible" to conquer Russia in 2 months. IRL, the German High Command---arrogant and all----estimated a minimum 10 weeks for a successful Barbarossa. That means if you as the German player invade on June 22, you are looking at least to fighting the Russians until the end of September. IMO, if you dont win in the Sept-Oct-Nov period, the game mechanics---i.e. weather, supply, org, (the friction of war) should be structured enough to prevent another effective large scale offensive until the next spring at earliest.
This should not be done in an attempt to create a historical outcome. Far from it, is what I suggest. What I want are "historical factors" affecting into your decisions. Realistically, if you invade Russia and do not conquer it by December, it is ahistorical for you to have the strength to launch a full scale attack in the dead of winter, with bad supply and org. That doesnt mean you cannot do it. It just means that your likelyhood of success is very low.
As an addition to my previous post, I would like to also suggest that weather have an effect on supply and org. It is not realistic that during the Russian winter that German units are supplied 100% and have an ORG rate of 100%. For that matter their strength rating shouldnt be at 100 either. There should be a constant drain on overall strength of the unit due to winter attrition.
Thats exactly my point. Increasing unit travel time (in proportion from unit to unit) should have zero effect on how the AI functions, or the decisions it makes. I was really only responding to your suggestion about slowing movement coupled with Org rot.
The unit speed change means that AI decisions are basicly more critical and a bad decision more costly in terms of time. Since the AI can react quickly, but not well this leaves it less able to correct its screw ups, AND makes each screw up more critical.
You suggest adding all kinds of causes of Org rot, but the AI is far less capable of dealing with that sort of thing then we are. Look at unit attrition suicide in EU2, hell look at low infrastructure unit attrition in HoI - its really silly - and thats all of a single factor -> province infra lower then 34 will cause units to fall apart - and the AI will kill itself on it. You are suggesting modelling supply lines more solidly then they are now. The result would be: more watching the AI kill itself.
The exception I do possibly agree with is the Garrison thing, but I think that might break combat in the east, and the european colonies. Or it might make it more historically accurate - I am not sure.
What is being referred to is the speed of conquests in this game. The game is over too quickly due to the fact that one can drive from Paris to Moscow in a matter of a couple of months game time. In reality, such an action should take years. For instance, it took Germany several months to get from the German/Bel-Lux border to Paris. In game time it could take a matter of a week (I've done it before...sooner if you just drop a paratroop on it). I feel that for the game to progress like an actual tactical/strategic simulation, movement values should be lowered as well as organizational penalties increased (someone mentioned this in this thread).
As for what a move is, in chess you can start the game by moving one of two different pieces: the pawns or the knights. The knights each only have two moves available to them as well as the pawns. Eight pawns times two moves plus two knights times two moves equals 20 maximum possible opening moves in chess. In HoI a move is simply defined as an action (like in chess) that you're allowed to do. Take the country of Germany...you've got nearly a hundred different research projects that you can do all in different combinations based on total ICs available. You've also got the ability to build from zero to several hundred units (not 100 at once, but rather different types up to 100 combinations). Then include all possible diplomatic moves (hundreds of countries that can be affected each with half a dozen different diplomatic actions to be attempted). Then you can move your forces in hundreds of different ways - not including the hundreds of different provinces you can move to. Don't forget the hundred plus of leaders Germany possesses that can be changed for each unit as well as a couple of new ministers (I'm assuming they're random as well as based on research and time). Lastly, just trading on the World Market could produce hundreds (if not thousands) of different possible combinations. So one does truly have millions of potential opening moves in HoI that the computer would have to take into account if one were to write code for the AI. Quite an impossible task...
Jack-
Ok now I understand what you are saying.
I do agree with you to a certain point.
At the end of the day you need good AI for a good game.
However, there are things you can do to realistically model the game, which has the effect of slowing down the player.
If the AI is bad, Id rather have it bad and me realistically face historical factors---such as decrease in supply, lower movements speeds, decrease in org/combat effectiveness.
I think some of this stuff can easily be modded or patched to create a realistic feel.
And you are correct to point out that these will be additional factors the AI will need to take into account. However this factors can be "quantified" in the AI---feints, suprise attacks and human "hunches" cannot.
Ok now I understand what you are saying.
I do agree with you to a certain point.
At the end of the day you need good AI for a good game.
However, there are things you can do to realistically model the game, which has the effect of slowing down the player.
If the AI is bad, Id rather have it bad and me realistically face historical factors---such as decrease in supply, lower movements speeds, decrease in org/combat effectiveness.
I think some of this stuff can easily be modded or patched to create a realistic feel.
And you are correct to point out that these will be additional factors the AI will need to take into account. However this factors can be "quantified" in the AI---feints, suprise attacks and human "hunches" cannot.
You can get it at www.the-underdogs.org for free.
Though the variants/partisans (what is it about partisans?) aren't implemented.
The AI isn't too hot (in one game I pushed the Russians back to the Urals in the 1944 scenario), and if you try anything original, like invading spain to get Gibraaaalta or invading Russia through Turkey, it can't handle it.
The best thing about the game (which HOI doesn't have) is that it has an Master level where the AI gets insane amounts of extra units and BRPs. That makes for a good game.
Agreed. But it will only make movement more realistic, from what i know of the current combat model, speed plays no part in determination of attack/defense/hit/damage.
Yes, like this idea. It'd be nice if you could create a supply dump in a province(that could be captured), which fed surrounding provinces(to a degree) with supply. I doubt this will happen in Hoi 1.
Agree with most points, except why couldnt a german unit be in full supply on the steppes? As far as org recovery rates, i don't think they're too bad now, but STR reinforce has to be redone. A division, army, corps shouldnt go from 3/3 to 100/0 with one click. A think clicking reinforce should start a gradual process, with the unit building str and org and a similar rate until full or it moves/gets attacked. One problem now is you beat a stack, it retreats, you pursue and what was 38/0 is now 100/0, and you end up chasing it ad infinitum instead of destroying it.
Disagree. I don't want to have to micromanage divisions to babysit provinces. However, if they could handle it abstractly (say by decreasing mp by 10 when you conquer a province to simulate police forces), that would be fine...
Inf attack/defend should be redone especially in regards to defending/attacking armor. IRL infantry was far more effective, particularly in urban terrain and against unsupported armor.
Man, that site is a nasty bit of pop-up work. Glad I have opera installed
Good stuff though, thanks...Lets not delude ourselves.
Yes.. because you can have set up trades on the world market using any number combination you want (up to what? the length of an unsigned int?) does not mean that realistically the world market will give you infinite play variation or infinite strategic variation.
An ideal world market handler is a formula that checks conversion rates against incomes against requirements, and results in ideal numbers of goods to trade. There is no depth, there is no shades of grey, no give, no take. Thats it, 10 lines of code.
Research prorities again realistically result in certain ideal paths through an overly convoluted (I'm sorry, but its needless, and its true) tech tree. The truth of the matter is there are certain landmark techs which need to be hit, there is a huge amount of waste, and if you sit down for 20min you can boil down tech tree climbs to perfect paths.
hundreds of countries that can be diplomatically effected? Again... I hate to say it, another formula will give you absolute perfect lists of who to influence first. IC production total, modified by distance from nearest enemy member of alliance - then you take into account current government leaning. Country with highest number should be influenced until a goal is reached. End.
I could address the rest. None of those are moves. Thats all side tasks that the really require nothing near complicated thought process, most have absolute best decisions which have no 'cost'. They are all seperate micromanagement tasks which the computer and AI excell at.
Again, its all what you want to consider a 'move'.
When I pick an opening in chess, I am not just moving a pawn, I am choosing a strategy, and a style of game.
Of valid chess openings there are many. Each have their own +'s and -'s.
I would just as much consider the first HoI 'opening' to be picking who to beat up, or deciding who will beat you up, in which case the number of real opening moves becomes far less. An offensive or defensive posture is taken, and in making these two decisions (who to target, and if you should take an offensive or defensive stance) about 90% of the decisions until you declare war or are declared war on are made for you.
(just because there is far more micromanagement in HoI then in chess does not mean the game is deeper...)
I don't really care about how unit speed relates to historical performance. I was just saying that slowing down the movement rates won't somehow make the AI smarter.
I know it seems like it should. "If I give the AI more time to react, it will react better" seems logical, its how people work. But it isn't how AI works. The number of game hours that pass before a unit gets from point A to point B (all things being equal) would affect AI performance about as much as changing the game speed.
Sorta simulates achieving total surprise, doesn´t it? Did you play above Normal? On the higher levels Ai gets so many extra units I can´t imagine any hex being empty.
A serious answer to the original poster's serious question: Yes.
I'm quite sure the reason you say "no" is because you're playing Germany. In both of my games so far (USSR, US) the AI Germans have proven to be invincible supermen who can neither be halted nor driven from their positions - even given equal numbers, which I have not been able to achieve.
In my current US game I have roughly 100 divisions in Spain (the Republicans won, joined the Comintern, and got stomped), with reputedly higher land tech than Germany, and they're being routed - even on the defensive in mountains.
I'm quite sure the reason you say "no" is because you're playing Germany. In both of my games so far (USSR, US) the AI Germans have proven to be invincible supermen who can neither be halted nor driven from their positions - even given equal numbers, which I have not been able to achieve.
In my current US game I have roughly 100 divisions in Spain (the Republicans won, joined the Comintern, and got stomped), with reputedly higher land tech than Germany, and they're being routed - even on the defensive in mountains.
But this is not as it should be.
It should be challenging for both sides, NOT give weird, ahistorical advantages to the side that lost.
Paradox should just include some AI-favoring scenarios OR incorporate those advantages for AI into Very Hard/Furious difficulty settings. In my opinion AI may cheat in hardest difficulties to give challenge. I'm not expecting some kind of Artificial Intelligence programming breakthrough discoveries. I love this game but as in most strategy games AI is beatable by seasoned players so lets face the facts and do something about it which does not demand gigantic re-programming effort.
Something like Federalists's hard German scenario would be cool, like Axis Very hard and Allies very hard etc. Yes, some might just urge people to use 3rd party mods to make the game more difficult, but offering official scenarios/settings would cut 90% of the talk about bad AI. Civilization AI cheats yet it is hugely popular regardless of that. It would simply make people buy the game more and wash away badmouths chanting about inept AI.
Self-invented restrictions and historical imitating just isn't satisfying alternative.
That is all, I rest my case.
Something like Federalists's hard German scenario would be cool, like Axis Very hard and Allies very hard etc. Yes, some might just urge people to use 3rd party mods to make the game more difficult, but offering official scenarios/settings would cut 90% of the talk about bad AI. Civilization AI cheats yet it is hugely popular regardless of that. It would simply make people buy the game more and wash away badmouths chanting about inept AI.
Self-invented restrictions and historical imitating just isn't satisfying alternative.
That is all, I rest my case.
Since Bolt has left, I may make some scenarios as he was planning to for 1.03 designed to make things more difficult for the human player. A lot more space to fill in the modding community.