RHS vs AI

[kafka]

I'd like to try out one of the RHS scenarios, since due to time constraints I'll only able to play AI, which scenario should I take? Thanks

btw I plan to play Japan vs Allied AI

[el cid again]

RHS only has one scenario suitable for AI play at this time. It is RHS AIO (AI Oriented).

I am considering doing an RHS CVO variation for AI play (RHA CAIO???).

ALL OTHER RHS scenarios have problems. The biggest of these is they have interior river systems, and AI wastes most of the ships it control sending them to places they can never reach! Add to that that most RHS scenarios are Russian active, and these have Soviet Submarines - which AI cannot deal with (until after it is certain you have active Russians). AIO addresses both issues: it has no interior river systems and it has Russian subs delayed until summer, 1945.

AI is also weak - so AIO is based on RHS EOS - which is stronger than CVO or BBO family scenarios. It helps compensate for dumb AI if it has "more marbles to play with."

[kafka]

thanks sid

I read in the manual the RHSAIO scenario is only intended for the AI controlling Japan. If this is still true I think I'll simply have to wait

[Sardaukar]

You might have to wait long time, since inherently AI can only control Japan with any sort of resemblance of efficiency. Allied AI is step worse due to complexities of Amphibious warfare that AI just cannot deal well with.

[el cid again]

ORIGINAL: kafka

thanks sid

I read in the manual the RHSAIO scenario is only intended for the AI controlling Japan. If this is still true I think I'll simply have to wait

I don't think there is any point in even hoping. AI isn't AI - it is a misnamed thing. It is not intelligent. It is preprogrammed rules. It is not able to handle the Allied situation - fundamentally - at its core. And that from a Matrix programmer. Nothing to wait for: you can have a crude almost AI controlling Japan - or nothing. Not gonna change. To change it is to design a new system fundamentally. Ron is right.

[Sardaukar]

One thing that designers could do is to open the scripting of AI for Modders. Or allow use of AI modules properly (and explain them) in full campaign. It'd still be scripted PO (I prefer Norm Koger's "Programmable Opponent", his TOAW games allow very good PO scripting for scen designers, to "AI"), but would give certain unpredictability and "suspension of disbelief".

It seems that "AI" is too deeply coded that I doubt it can be done. I do have hopes, tho.

Joel Billings have explained it, in his very good posts, why AI is not better.

[kafka]

well thank you for your replies

I did not expect an AI equivalent to human intelligence, I'm just interested in the RHS scenarios and wanted to know if there is any of them AI capable in terms of the AI quality provided with the vanilla game. I certainly do not expect a scenario designer to improve the AI in this game, yet if the scenarios as such - due to an increased complexity level - lead to an even worse AI, then I'll have to pass

[Sardaukar]

I think best AI experiences I have had vs. AI were Allied games with Subchaser stock map mod (making it lot nicer looking while also correcting few errors) and using BigB mod.

Lets face it, AI can only play Japan decently and other than stock map, it tends to get confused.

I do play CHS vs AI for example on AB's excellent extended map..but AI cannot perform as well on that.

I think el cid had tweaked computer opponent quite a bit to allow Japan be competitive in his RHSAIO, but haven't tried that yet.

[TulliusDetritus]

El Cid Again, some months ago you said that [no matter what] the Japanese AI would economically die by september or october 1942. What about this AIEO scenario? Is this statement still true? I guess it is not, otherwise this AI oriented scenario would be a waste of time. So?

[el cid again]

This is a different matter - and once again - it is a function of hard code - beyond the reach of modders and players in the usual sense. The ONLY reason to CONSIDER an AI game of ANY sort (stock, CHS, RHS, name it) is that you do NOT intend to play past August 1944 - OR that you will consider the workaround I have posted. This problem is structural, endemic to WITP as such, and independent of mod or scenario.

The technial issue is kamakaze planes. For reasons wholly unclear to me - AI wants to convert EVERY Japanese squadron to Kamakaze - which of course is unreasonable, impossible, and wholly negates almost all the value of air forces: no longer can any air unit be a fighter unit, a bomber unit for land targets, a recon unit or even a transport unit: all must be naval suicide units that self destruct! By the end of November 100% of air units are this way - and new air units convert within a day or so. It is the end of Japan as a competator - even with AI controlling Japan - and all the Allies do is eat Japanese positions from then on - suffering occasional casualties to kamakaze attacks on ships (pretty rare as there is no proper recon).

The good news is that all this may be about to change. I don't know if this is one of the things addressed by the next patch or not? But air unit behaviors ARE addressed by it - and in fundamental ways - so I would not be surprised - in fact I expect it will be addressed. IF so ALL forms of WITP will no longer have this problem AFTER that patch.

That will not change that AI can ONLY do Japan, unfortunately.

For now, and in future if this problem is not fixed by the patch, IF you play a game past July 1944 (rare) AND IF you want to keep playing AND IF you don't want to see Japan collapse as a competitive player (to the small extent AI is ever competative) THEN you MUST reset the game to a HUMAN form - even if you don't play a human. You may pick either Face to Face (best) or PBEM - and you may either get a real human opponent (best) or you must do one of three things:

a) Just say "yes" to the Japanese turn - do NOTHING at all - look at nothing at all
b) Just pick a number (roll a die?) of air units - select them at random (roll two decimal dice?) - and convert them to kamakaze - then go to (a) above
c) Play the entire turn as Japan (and let what you learn represent the superb late war intel on the enemy the Allies had)

Sid

[el cid again]

ORIGINAL: Sardaukar

I think best AI experiences I have had vs. AI were Allied games with Subchaser stock map mod (making it lot nicer looking while also correcting few errors) and using BigB mod.

Lets face it, AI can only play Japan decently and other than stock map, it tends to get confused.


REPLY: Not exactly. AI is coded for location and, to some extent, situation. By changing the slot of a location, we can "reprogram" the AI - and it will use a different map as well as the stock map.


I do play CHS vs AI for example on AB's excellent extended map..but AI cannot perform as well on that.

I think el cid had tweaked computer opponent quite a bit to allow Japan be competitive in his RHSAIO, but haven't tried that yet.

REPLY: This is quite correct. It appears that you can "tell" AI "this slot is a place to send things to" by the initial value of supplies in that slot (90,000 plus does the trick). It is certain you can "tell" the AI "go to this hex on turn x" if you pick a major objective in stock - and use the slot of that location for your purposes. In this way you can change many things fairly successfully. There are other things. And I took out features of RHS which wholly mess up AI (Russian submarines - which do not exist in code even after the Russian navy got enabled by Matrix, or interior river ports).
To that add that we used a "Japan enhanced scenario" so AI has more marbles in the pot to play with - and bigger marbles in some cases.

But a CAIO scenario would not have the latter feature - it would be CVO type units exactly.

[Sardaukar]

I'd be very interested about CAIO, indeed. Due to time constraints, I play vs. AI only.

[WingedIncubus]

Hello,

I would also be interested in trying RHS against the Japanese AI, but is there a scenario playable against the AI, but without the Pearl Harbor Invasion in RHSAIO?

I don't want to be harsh or overly critical of why this... possibility is there for the only scenario adapted to the AI, but this "what-if" plan of an invasion of Pearl Harbor in December 1941 is IMO pure fantasy, like Germany invading Spain to reach Gibraltar or planning to invade the UK in 1942: technically feasable in theory, yes, but hardly historically feasable in any fashion. 

First, it wasn't part of the Japanese offensive plans at the time, as it was either Yamamoto's surprise attack against Pearl Harbor or a full concentration of force on a Southern Offensive to seize the DEI. Yamamoto's resignation threats ensured that the Naval Chief of Staff would go for a coup de main on Pearl Harbor, but they would have never approved a full invasion of Hawaii.

Second, no one was assured that Pearl Harbor would be the tactical success it is known now. Any plan of a possible seizure of Pearl Harbor was made a posteriori of Pearl Harbor, knowing its results. In fact, an invasion of Pearl Harbor would have been even harder to keep as a surprise since it would haved slowed down the fleet, decupling the chances of being inavertably spotted.

Third, it also goes against military logic, as capturing Pearl Harbor without holding Midway, Johnston and Palmyra would endanger any supply attempts, would overextend the supply lines beyond any Japanese capability, would need to keep KB permanently on the defence there, and would provoke the US into a full scale attempt to seize back Pearl Harbor at any cost. Also, what would the Japanese gain by it to reach the two main objectives of destroying the US carrier fleets and secure its oil supply lanes? I doubt both the Navy Chief of Staff or the Cabinet would have approved of such a plan, as it strives neither of these objectives and would strip Japan of its most important carrier task force.

Sorry if I sound a bit negative, but I'd like to try RHS and I am a bit disappointed to find that the RHSAIO involves something that goes beyond my logic. But if there are other scenarios that can be played in a satisfactory manner against the Japanese AI, I'd be glad to know it.

[Buck Beach]

ORIGINAL: Drakken

Hello,

I would also be interested in trying RHS against the Japanese AI, but is there a scenario playable against the AI, but without the Pearl Harbor Invasion in RHSAIO?

I don't want to be harsh or overly critical of why this... possibility is there for the only scenario adapted to the AI, but this "what-if" plan of an invasion of Pearl Harbor in December 1941 is IMO pure fantasy, like Germany invading Spain to reach Gibraltar or planning to invade the UK in 1942: technically feasable in theory, yes, but hardly historically feasable in any fashion. 

First, it wasn't part of the Japanese offensive plans at the time, as it was either Yamamoto's surprise attack against Pearl Harbor or a full concentration of force on a Southern Offensive to seize the DEI. Yamamoto's resignation threats ensured that the Naval Chief of Staff would go for a coup de main on Pearl Harbor, but they would have never approved a full invasion of Hawaii.

Second, no one was assured that Pearl Harbor would be the tactical success it is known now. Any plan of a possible seizure of Pearl Harbor was made a posteriori of Pearl Harbor, knowing its results. In fact, an invasion of Pearl Harbor would have been even harder to keep as a surprise since it would haved slowed down the fleet, decupling the chances of being inavertably spotted.

Third, it also goes against military logic, as capturing Pearl Harbor without holding Midway, Johnston and Palmyra would endanger any supply attempts, would overextend the supply lines beyond any Japanese capability, would need to keep KB permanently on the defence there, and would provoke the US into a full scale attempt to seize back Pearl Harbor at any cost. Also, what would the Japanese gain by it to reach the two main objectives of destroying the US carrier fleets and secure its oil supply lanes? I doubt both the Navy Chief of Staff or the Cabinet would have approved of such a plan, as it strives neither of these objectives and would strip Japan of its most important carrier task force.

Sorry if I sound a bit negative, but I'd like to try RHS and I am a bit disappointed to find that the RHSAIO involves something that goes beyond my logic. But if there are other scenarios that can be played in a satisfactory manner against the Japanese AI, I'd be glad to know it.

Just for FYI purposes, RHSAIO does capture Midway, Johnston, Hawaii (big island) and Maui at the begining. Maui becomes cannon fodder for the Pearl aircraft after recuperating. An occasional bombardment run from Pearl to Maui and Johnston further suppresses those bases but Maui Coastal Guns exact a toll. After two months of game time there has been no further support for Maui except additional aircraft. But the KB and Surface Forces do stick around for the first few days to a week and if you don't run or try to hit the peripherals you can scratch one or two carriers.

[el cid again]

ORIGINAL: Drakken

Hello,

I would also be interested in trying RHS against the Japanese AI, but is there a scenario playable against the AI, but without the Pearl Harbor Invasion in RHSAIO?

Not at this time. The AI is so weak we based AIO on the Japan Enhanced Scenario - and that has the IJN pinning the US Fleet to Pearl Harbor - to keep it out of the way of SRA operations. Also - Japan gets better quality units in various senses - for various reasons - essentially because it planned better in this hypothetical world than IRL.

I am considering doing a much weaker AIO - tentatively called CAIO - to communicate it is based on CVO. This would not work nearly as well - if one can ever say AI works well at all. I see it mainly as a training tool so players can get a sense of how CVO works before being humiliated by a human opponent.

RHS is NOT for AI play - it is focused on human players. For that matter, WITP is NOT for AI play either - in spite of the fact it has that option and most games are done that way. If it WERE for AI - AI would be better - and work on making it better would be onging - which so far it has not been. This is not entirely ideal - but the cost of doing anything else was and probably is prohibitive. Sorry.

[el cid again]

ORIGINAL: Drakken

Hello,


I don't want to be harsh or overly critical of why this... possibility is there for the only scenario adapted to the AI, but this "what-if" plan of an invasion of Pearl Harbor in December 1941 is IMO pure fantasy, like Germany invading Spain to reach Gibraltar or planning to invade the UK in 1942: technically feasable in theory, yes, but hardly historically feasable in any fashion. 

First, it wasn't part of the Japanese offensive plans at the time, as it was either Yamamoto's surprise attack against Pearl Harbor or a full concentration of force on a Southern Offensive to seize the DEI. Yamamoto's resignation threats ensured that the Naval Chief of Staff would go for a coup de main on Pearl Harbor, but they would have never approved a full invasion of Hawaii.

REPLY: In the sense that it was not the adopted plan, this is quite correct. In the sense that it had been studied - it is incorrect: Japanese planning to invade Hawaii dates to 1910. The first major naval study - headed by less senior Yamamoto no less - dates to 1928. A proposal was made as late as October 1941 - by two intel officers just back from Hawaii - to implement this planning. The proposal was not adopted - but it was considered - and could have been adopted (probably - if Yamamoto had threatened to resign if not adopted). Anyway - Yamamoto concluded two days after Pearl Harbor that a grave error was made in NOT accepting that proposal. This scenario set is based on the reverse case.

Second, no one was assured that Pearl Harbor would be the tactical success it is known now. Any plan of a possible seizure of Pearl Harbor was made a posteriori of Pearl Harbor, knowing its results. In fact, an invasion of Pearl Harbor would have been even harder to keep as a surprise since it would haved slowed down the fleet, decupling the chances of being inavertably spotted.

REPLY: We have no less an authority than Admiral Yamamoto himself, explicitly on this very subject, on Dec 10, 1941 (Japan time) that you have this perfectly correct that no one was "assured (such an attack) would be a tactical success" - although I prefer operational to tactical - since only an operation (like Malaya or Philippines or DEI) could take a major objective of this sort. The surprise issue is far less important - and indeed might (repeat might) not be even tried for. The goal - first of all - is to the the United States Fleet to fight. Getting it to fight at sea would be much better - where when a ship is sunk it stays sunk. Sun Tzu - popular in Japan - argues that it is better to cut of a few fingers than to hurt all of them. This is a very fine strategy - almost doomed to strategic success - even if it turns out to be expensive in operational terms - because Japan is fighting a battle it does not need to win - but the US cannot afford to let it win. It is my most favorite way to fight - and the biggest problems with it are logistical impacts and selling it in a world not yet convinced of the value of naval air power. IF attempted, it would be far more likely to succeed than it seems on paper. There is not any easy way to recreate the ignorance and arrogance of the Americans then, but they surely would have engaged - even single carriers were TRYING to engage - not like up with each other - and had the fleet been able to sortee - it not only would have done - it would have sailed AT the enemy - not run for Panama or the West Coast. Which is to say the prospects are that every major unit of the United States Fleet not on the West Coast or escorting convoys (a couple of CA were doing that) would have been lost within a few days. Then the problem becomes one of isolating Oahu and reducing it with air power - to the point it could be invaded. It actually does not matter in a strategic sense if that is ultimately successful or not - by the time it is clear the strategic objectives of the op will have been accomplished. Japan will have had no interference in the SRA worthy of note. And if it does succeed - which is probable - Japan has a war termination option like no other available - down the road. Ultimately - if we don't tire of the war enough to accept some sort of compromise over Hawaii - we offer the possibility of strategic domination of the Pacific to Japan, long term. Of course - that assumes that Japan uses the time it buys wisely. One can always lose a war - even one that one is winning. And I believe Japan insured its own defeat in many diverse ways. But it need not have done that. And we have learned since WWII that the US is not guaranteed to win every war it fights.

Third, it also goes against military logic, as capturing Pearl Harbor without holding Midway, Johnston and Palmyra would endanger any supply attempts, would overextend the supply lines beyond any Japanese capability, would need to keep KB permanently on the defence there, and would provoke the US into a full scale attempt to seize back Pearl Harbor at any cost. Also, what would the Japanese gain by it to reach the two main objectives of destroying the US carrier fleets and secure its oil supply lanes? I doubt both the Navy Chief of Staff or the Cabinet would have approved of such a plan, as it strives neither of these objectives and would strip Japan of its most important carrier task force.

REPLY: Once again, I am forced to agree with you. But in agreeing - it exposes you do not understand either the historical plans of Japan or the scenario in RHS. Johnston is captued on the first day of the war. Midway sometime in the first few days. Only Palmyra was not included IRL planning or this scenario - and that because it isn't required. But players have the forces to do it if they think it wise. Your reasoning is sound: the Japanese agreed; I agree. So that is not what was planned - nor what is simulated here. Indeed, the real life follow up in which this invasion WAS really attempted - we call it the Battle of Midway - DID include an op against Midway - which didn't work out - and in the event was probably not going to work out - even had all three US carriers been sunk that day. This follow up op was not facing the conditions of 1941, and didn't have the intel which was available in 1941 (because it was peacetime).


Sorry if I sound a bit negative, but I'd like to try RHS and I am a bit disappointed to find that the RHSAIO involves something that goes beyond my logic. But if there are other scenarios that can be played in a satisfactory manner against the Japanese AI, I'd be glad to know it.

The subject is esoteric. It took me 40 years to convince most the strategists I know it was a wise move - indeed the best possible opening move. It will grow on you as you think about it - and learn more about conditions in 1941. [NOT in 1942 or later - it is impossible then - or possible only with so much cost it loses the war even if it succeeds. Also - its strategic point is diluted - Japan needs its ships after capturing resources - and it needs the period during which they are captured free from the US Fleet - not some later time - and those ships cannot do two things at the same time either.] Anyway - strategy is immaterial: AI needs the EOS strength to do better - so dumb is it. And adopting the usual strategy will surprise no one. This way the US player is at least distracted for a while. Regretfully - the AI won't honor its assigned objectives more than a couple of days.

No - there is no other AI scenario at this time. And other scenarios won't work very well with AI. It hates interior rivers - that is it abuses them grossly - assigning obscure minor ports as home ports - sending ships to them - or from them - often with no cargo (and why send a ship with no mission????). It does not understand Soviet submarines - and it thinks active Soviets are at war (understandably enough). This latter rules out the majority of RHS scenarios, the others rule out all of them. Then there is the matter that Japan is badly outnumbered in most senses most of the war - and it is horribly unbalanced as a campaign. We cannot use a historical campaign as fought (CVO family) or as planned (BBO family) and hope for AI to be more than the "token opposition force" as in a US military exercise.

[WingedIncubus]

It's okey, I have already said that I wanted to play against the AI only for basic practice and understanding the game mechanics. It is just that RHS is quite attractive to me, and I wish I could have a bite on it before going for a human player. [:)]
 
Sorry if my criticism has seemed harsh, though. I realize now it is more by necessity than for historical sake - even though I would highly support a CAIO. [&o]

[el cid again]

Practice is the reason I may do a CAIO. It is a good idea to become familiar with the system - which is more complex in RHS than stock - by a long way.

But you can use AIO to practice on. If you can beat a stronger Japan - you might do OK against a more limited one.