useless variants!?

[Big Nev]

So… to summarise…

You can have light, medium, heavy & super-heavy tanks.

With research, you can “unlock” and build four variants of each tank which are tank-destroyer, SP-artillery, rocket-artillery and anti-aircraft.

With [expenditure of] experience [points] you can improve your basic model of tank (&, presumably, the variant types) with regard to armour, engine, gun (and/or other areas) as if these had been “researched ahead”.

But…

You can only build one “model” of medium tank (for example) at a time. You can only build one “model” of medium SP-Art at a time & only one “model” of medium TD. Any experience-based upgrades will either be applied to your current production or not.

You will not, for example, be able to build “Pz IIIs”, “Pz IVs” &/or “Panthers” at the same time as these are all “medium” tanks.

So, sorry steindroff, unless I’ve got it wrong (wouldn’t be the first time & won’t be the last) the entire question of prioritising “fast mediums” or “infantry support” variants to your different “exploitation” and “breakthrough” divisions is irrelevant because these variants won’t exist. You will have one model for each type & classification (L, M, H, SH) and that’s it.

I think this is a little sad. I’d much prefer a tech' tree where the infantry support & ant-tank branches were represented. Pz IIIs & IVs were both mediums which started-out at the same time but had different rolls. This won’t be represented unless you consider your German medium SPArt variants to be your Pz IVs.

But then the research dates are all wrong and both Britain & France need to be building HArm in 1938 (or even earlier) to represent Matildas & the S-35.

(OK, HArm with really crap SA for Britain and some other way to nerf early French HArm but still, these tanks were very well armoured)

 

[Wminus]

Too much micro/useless detail. PI should ignore stuff like this, as it will only serve to unbalance the game and make the AI's life harder.

 

[Axe99]

But…

You can only build one “model” of medium tank (for example) at a time. You can only build one “model” of medium SP-Art at a time & only one “model” of medium TD. Any experience-based upgrades will either be applied to your current production or not.

Do we know this for sure? I wasn't thinking about tanks so much (as I figured it'd all average out within the division, and there wouldn't be a huge benefit to having whole armoured divisions set up differently in different theatres) but for aircraft there'd be benefit in having different models of fighters of the same tech 'vintage', prioritised for different capabilities, and thought the variant system was in one way trying to cover for this (hence only one fighter type now, instead of int/multi-role/escort from HoI3)? I could be wrong too, just have a vague (very vague) memory of some discussion of the variant system with this in mind back in an earlier DD.

 

[steindroff]

You can only build one “model” of medium tank (for example) at a time. You can only build one “model” of medium SP-Art at a time & only one “model” of medium TD. Any experience-based upgrades will either be applied to your current production or not.

Where do you get this!? ... I thougt that we even will be able to produce a Panzer IIIA and a Panzer IIIH at the same time, for example, specially two different types of (medium) tanks (Panzer III and Panzer IV) :unsure: Like you build two different variants of a heavy cruiser at the same time ...

 

[tommylotto]

I am not sure you have it right Big Nev. it is my understanding that you have models you research, PzIII or PzIV, etc. There are also sub models that you can also research, TD, SPART, SPAA, SPRART. Once you unlock a model or a sub model, you can change the model by expending combat experience to change the model in up to four ways ( with tanks it is armor, gun, engine and reliability ). These changes will add bonuses and detriments. For example, up armoring will make a tank slower unless you up the engine also, which will make it less reliable unless you up the reliability too. So, you expend your combat experience however you want to customize your tank model however you want. When you get it just the way you want you save the customized tank as a variant -- a speedy PzIII variant or a slow tough PzIV variant -- or even a speedy IV and a slow tough IV. You can rename these variants like PzIVa or PzIVh, or fast PzIV and slow PzIV. Then when you start a production line you choose which variant from any of the variants you have saved that you want to build. You could be building fast PzIV variants and slow PzIV variants at the same time in different factories.

What we don't know is how these tanks, once built are assigned to battalions. I think tanks once built will be just light, medium, heavy or super heavy as far as the game is concerned. So, the game will not distinguish your variants within each class. You will not be able to get a battalion of all fast PzIV's or a battalion of all slow PzIV's. The game will just divvy up your medium tanks to your medium battalions. It would be nice to be able to designate a medium tank variant such as a speedy PzIII variant as a light tank to go to light tank battalions. Then the OP would get what he wants, tough PzIV's going to his medium battalions and speedy PzIII's going to his light battalions.

 

[Snow Wolf]

What we don't know is how these tanks, once built are assigned to battalions. I think tanks once built will be just light, medium, heavy or super heavy as far as the game is concerned. So, the game will not distinguish your variants within each class. You will not be able to get a battalion of all fast PzIV's or a battalion of all slow PzIV's. The game will just divvy up your medium tanks to your medium battalions. It would be nice to be able to designate a medium tank variant such as a speedy PzIII variant as a light tank to go to light tank battalions. Then the OP would get what he wants, tough PzIV's going to his medium battalions and speedy PzIII's going to his light battalions.

I think this could either be solved by using light tanks instead, or a simple drop down menu in the division designer/when you have the individual unit placed, which model you'd prefer assigned. Remember, PDS has not uncovered unit deployment yet, and since they have not, I'll suppose they are making a bigger deal about it..

also, let's arrange some words: 1939 tank = model, TD = subtech and better gun/armour = variant, yes?

 

[Wyrm]

You can only build one “model” of medium tank (for example) at a time. You can only build one “model” of medium SP-Art at a time & only one “model” of medium TD. Any experience-based upgrades will either be applied to your current production or not.

You will not, for example, be able to build “Pz IIIs”, “Pz IVs” &/or “Panthers” at the same time as these are all “medium” tanks.

Actually, in the DD regarding how production lines work they clearly stated that you WILL be able to produce both old and new models. This to force a player to make the choice if it is worth it to start building the new tank at a slower pace since it needs a new production line with no bonuses or keep producing the old tank at a higher rate since the production lines have kept going long enough for bonuses to accumulate. And ofcourse, you can do this by moving a few factories from the old line, keeping most of the bonuses for the older model and getting a few bonuses for the new one and slowly transfer industry over to the new model to have some sort of balance in the introduction of new models.

 

[eleetsdier]

So… to summarise…
But…

You can only build one “model” of medium tank (for example) at a time. You can only build one “model” of medium SP-Art at a time & only one “model” of medium TD. Any experience-based upgrades will either be applied to your current production or not.

You will not, for example, be able to build “Pz IIIs”, “Pz IVs” &/or “Panthers” at the same time as these are all “medium” tanks.

So, sorry steindroff, unless I’ve got it wrong (wouldn’t be the first time & won’t be the last) the entire question of prioritising “fast mediums” or “infantry support” variants to your different “exploitation” and “breakthrough” divisions is irrelevant because these variants won’t exist. You will have one model for each type & classification (L, M, H, SH) and that’s it.

I think this is a little sad. I’d much prefer a tech' tree where the infantry support & ant-tank branches were represented. Pz IIIs & IVs were both mediums which started-out at the same time but had different rolls. This won’t be represented unless you consider your German medium SPArt variants to be your Pz IVs.

But then the research dates are all wrong and both Britain & France need to be building HArm in 1938 (or even earlier) to represent Matildas & the S-35.

(OK, HArm with really crap SA for Britain and some other way to nerf early French HArm but still, these tanks were very well armoured)

You're wrong - you can produce different types of medium tank (or whatever) at the same time. This seemed pretty clearly laid out in the DDs on Production. There is a valid question as to how easily you can make good use of multiple versions of medium tanks. It's also worth asking if Tank Destroyers, Anti-Air, and SPGs are even broken down by weight class. Would a Marder II belong in a different type of battalion than a Stug III or a Sturmtiger?

I don't understand your complaint about HARM. Aren't the Matilda and the S-35 Infantry Support Tanks, which are modeled by the heavy armor available in 1934?

Specialist Panzer IIIs and Panzer IVs sounds great in a tactical game. At the divisional TOE level I don't think the two were thought of as a separate thing - every Armored Division would have both.

 

[Snow Wolf]

Specialist Panzer IIIs and Panzer IVs sounds great in a tactical game. At the divisional TOE level I don't think the two were thought of as a separate thing - every Armored Division would have both.

Agreed, they would lose to a light tank in speed anyway.

 

[GarfunkeL]

Too much micro/useless detail. PI should ignore stuff like this, as it will only serve to unbalance the game and make the AI's life harder.

Nonsense. The AI will be a pushover in any case. Paradox has neither the resources nor the skill to make a truly challenging AI for experienced players. If you want true AI challenge, go play Chess. Because that's where the argument about complexity and AI will always lead - you have to remove layers and layers of complexity for the AI to be competitive against a human, until you return to something like Chess.

Agreed, they would lose to a light tank in speed anyway.

Not necessarily. While Panzer I was faster than Panzer III, Panzer II was not. That's one reason that they worked so well together in tactical level, because they had identical speeds. Of course, on a divisional level the difference isn't that significant, because every division had a variety of tank models - Pz I's, Pz II's, Pz III's, Pz IV's, Pz 35(t)'s and Pz 38(t)'s.

Interestingly, I guess could make the case that Pz II was an up-gunned and -armoured "variant" of Pz I, thus losing the speed advantage.

 

[Axe99]

Nonsense. The AI will be a pushover in any case. Paradox has neither the resources nor the skill to make a truly challenging AI for experienced players. If you want true AI challenge, go play Chess. Because that's where the argument about complexity and AI will always lead - you have to remove layers and layers of complexity for the AI to be competitive against a human, until you return to something like Chess.

Or play Gal Civ II - it is possible to have challenging AI in modern grand strategy games .

 

[GarfunkeL]

Only for the first five games or so. Then you've learned the exploits and how the AI routine in general works and you can beat it. GalCiv2 gets a lot of hype for its AI but it's nothing spectacular.

 

[Wminus]

Nonsense. The AI will be a pushover in any case. Paradox has neither the resources nor the skill to make a truly challenging AI for experienced players. If you want true AI challenge, go play Chess. Because that's where the argument about complexity and AI will always lead - you have to remove layers and layers of complexity for the AI to be competitive against a human, until you return to something like Chess.

When did I say I wanted an AI that could go head to head with a human on even terms? Stop throwing straw men.

And I'm sorry, but there's a major difference between an AI than can beat the world's best players in a hugely popular game like chess, and an AI that can provide a fun challenge. Guess what, building an AI that can handle the basics of military strategy as well as respond in logical fashions is NOT impossible and certainly doesn't require a game of chess-level complexity. Just look at half a dozen different strategy games, like CIV4, Shogun 2 (which is far more complex than HoI3 for an AI), GalCiv2 etc. that certainly provide a challenge if the opposing computer player is given a boost (extra money and so on).
So yeah, cut the complexity and focus on making an AI which doesn't do stupid shit all time, that's what I wish for.

 

[Snow Wolf]

that certainly provide a challenge if the opposing computer player is given a boost (extra money and so on).

but i dont want anyone to have a boost
why would i give my enemy a boost

 

[Big Blue Blob]

When did I say I wanted an AI that could go head to head with a human on even terms? Stop throwing straw men.

And I'm sorry, but there's a major difference between an AI than can beat the world's best players in a hugely popular game like chess, and an AI that can provide a fun challenge. Guess what, building an AI that can handle the basics of military strategy as well as respond in logical fashions is NOT impossible and certainly doesn't require a game of chess-level complexity. Just look at half a dozen different strategy games, like CIV4, Shogun 2 (which is far more complex than HoI3 for an AI), GalCiv2 etc. that certainly provide a challenge if the opposing computer player is given a boost (extra money and so on).
So yeah, cut the complexity and focus on making an AI which doesn't do stupid shit all time, that's what I wish for.

Citing Shogun 2 as an example is a bit strange given how much of a wacky mess it was full of fantasy units and absence of tactics and formations, but I shall mostly ignore it.

I do not expect the AI to be on the level of the player, but I expect it to be good enough to play a reasonable game, unless it is controlling France, Italy, early USSR or a few others and is supposed to act stupidly. However, I see it as more important to reflect all aspects of the period, including those outside direct fighting like industry and mobilisation, than to make it so the AI can win. This also includes giving all nations their historical potential and OOBs, and not having AI bonus unless players choose it, unlike in EU4 where AI has non-optional bonuses like no naval attrition and spare diplomats.

 

[Axe99]

Only for the first five games or so. Then you've learned the exploits and how the AI routine in general works and you can beat it. GalCiv2 gets a lot of hype for its AI but it's nothing spectacular.

Well, if you play with exploits you're likely to never have a challenge. Playing against an AI and playing in 'the spirit of the game' rather than exploiting, provides a decent challenge over extended play. It's worth noting that these days, even PDS games provide a half-decent challenge if you're willing to forego exploits . That said, designing a game (and AI) that can get around human exploits as well as provide a decent challenge in the spirit of the game is a whole 'nother job entirely.

 

[Darkath]

I certainly look forward using my SH-SP-R-ART divisions

 

[GarfunkeL]

When did I say I wanted an AI that could go head to head with a human on even terms? Stop throwing straw men.

And I'm sorry, but there's a major difference between an AI than can beat the world's best players in a hugely popular game like chess, and an AI that can provide a fun challenge. Guess what, building an AI that can handle the basics of military strategy as well as respond in logical fashions is NOT impossible and certainly doesn't require a game of chess-level complexity. Just look at half a dozen different strategy games, like CIV4, Shogun 2 (which is far more complex than HoI3 for an AI), GalCiv2 etc. that certainly provide a challenge if the opposing computer player is given a boost (extra money and so on).
So yeah, cut the complexity and focus on making an AI which doesn't do stupid shit all time, that's what I wish for.

In your original post. Stop shifting goal posts, dude-bro.

The reason why Deep Blue beat Kasparov is that Chess is, at the end of the day, a very simple game. Sure, there are multiple strategies and many moves, but at any given time, there is only a very limited amount of moves that are not recipe for a quick suicide. Meaning that the AI "only" has to simulate how the game could proceed from that point in time onwards and then select the best move based on a relatively easy-to-program criterion.

Contrast that to practically any historical wargame, which are - more often than not - cluttered with different layers and require a human to use "fuzzy logic" to proceed. It's a practical impossibility to design an AI that can challenge an experienced player when you have multiple layers of complexity, especially when you also need to make the AI behave somewhat historically. As loud as the sandbox-players are before a game is released, they are always drowned out by the complaints after a game ships and wildly ahistorical events start happening. I remember Hoi3's launch well.

So the quest for a "challenging AI" is impossible and while pursuing it, the dev team can easily end up stripping layers of complexity off the game, layers that usually make experienced players actually enjoy the game far more. GalCiv2 is a very boring 4X game. Civ4 is not challenging at all as only challenge comes when the AI cheats like there's no tomorrow. Shogun 2 is not more complex than HoI3 AI-wise and it had a very shoddy AI.

Well, if you play with exploits you're likely to never have a challenge. Playing against an AI and playing in 'the spirit of the game' rather than exploiting, provides a decent challenge over extended play. It's worth noting that these days, even PDS games provide a half-decent challenge if you're willing to forego exploits . That said, designing a game (and AI) that can get around human exploits as well as provide a decent challenge in the spirit of the game is a whole 'nother job entirely.

Hah! But seriously, how do you think I can still keep playing HoI3? I've been playing it since original launch after all. Sure, mods help out but I always end up giving myself plenty of handicaps and I try to play as historically as possible, to get as much fun out of the game as possible - and I'm by no means an expert as I'm sure I would get utterly wrecked in multiplayer. I just wanted to point out that, like the quest for removing all exploits, a quest for "challenging" AI is ultimately futile. Sure, you don't want the game to be completely broken through an painfully obvious exploit or a completely inept AI but taking out features for those reasons is stupid and counter-productive.

 

[Axe99]

Hah! But seriously, how do you think I can still keep playing HoI3? I've been playing it since original launch after all. Sure, mods help out but I always end up giving myself plenty of handicaps and I try to play as historically as possible, to get as much fun out of the game as possible - and I'm by no means an expert as I'm sure I would get utterly wrecked in multiplayer. I just wanted to point out that, like the quest for removing all exploits, a quest for "challenging" AI is ultimately futile. Sure, you don't want the game to be completely broken through an painfully obvious exploit or a completely inept AI but taking out features for those reasons is stupid and counter-productive.

Oh aye, but HoI3 is very exploitable, from maximising your chain-of-command bonuses to effectively using terrain (yes, I hope that effectively using terrain won't be an exploit in HoI4!) to all of the movement 'dancing' and the like (and I'm sure there's a whole bunch that aren't coming to mind right away - instant boarding/disembarking, there's another one) that it's one of the most exploitable games out there - a bunch of features that really only human players use that the AI can't cope with (indeed, HoI3 had a bunch of features the AI pretty much never used at all). There's a reason I pulled up stumps from HoI3 once I heard HoI4 was coming and had read the first couple of DDs, because you have to remove so many features from the game to be on the same level as the AI.

Gal Civ II isn't like this (turn based, of course, helps with the dancing, but generally the AI will do everything an average player does - Gal Civ II had reactive force builds for example, something that PDS games still don't have), and EUIV is a long, long way forward (there's still a lot of dancing, and the instant embarking/disembarking is still there, but as long as I don't exploit either of these, I find I can usually have a good and still challenging game of EUIV - and the AI will generally do everything beyond these two things that the player does). I reckon that the initial release might have some teething troubles, but there's a fair chance that HoI4 will actually give us a game where we don't need to ignore features to be on a level playing field with the AI, and that'll go a long way forward to making it a decent challenge without needing to buff the AI/debuff humans. That said, by all means laugh at my preposterous claims a few months after release if it doesn't turn out that way .

 

[Wminus]

In your original post. Stop shifting goal posts, dude-bro.

The reason why Deep Blue beat Kasparov is that Chess is, at the end of the day, a very simple game. Sure, there are multiple strategies and many moves, but at any given time, there is only a very limited amount of moves that are not recipe for a quick suicide. Meaning that the AI "only" has to simulate how the game could proceed from that point in time onwards and then select the best move based on a relatively easy-to-program criterion.

Contrast that to practically any historical wargame, which are - more often than not - cluttered with different layers and require a human to use "fuzzy logic" to proceed. It's a practical impossibility to design an AI that can challenge an experienced player when you have multiple layers of complexity, especially when you also need to make the AI behave somewhat historically. As loud as the sandbox-players are before a game is released, they are always drowned out by the complaints after a game ships and wildly ahistorical events start happening. I remember Hoi3's launch well.

So the quest for a "challenging AI" is impossible and while pursuing it, the dev team can easily end up stripping layers of complexity off the game, layers that usually make experienced players actually enjoy the game far more. GalCiv2 is a very boring 4X game. Civ4 is not challenging at all as only challenge comes when the AI cheats like there's no tomorrow. Shogun 2 is not more complex than HoI3 AI-wise and it had a very shoddy AI.

Lol dude. Read my "original post" again then. I never expected the AI to be able to beat an experienced player head on.

And in the rest of your post you are just repeating yourself. The HoI3 AI was not good at all. Sometimes it would keep half its forces moving back and forth ad infinitum along the front line, launch attacks with only 1/3 of the divisions it could use, make idiotic amphibious assaults and so on. No way the STW2 ai was that bad. It did its job perfectly because the gameplay was designed around it, like it seems HoI4 is (look at new air warfare for instance - much easier for an AI to handle that).

Really, the fact that you deny there exist any good video game AIs at all just shows you are looking for excuses for PI to not cut micro from the game.