|
Post by anthropoid on Jan 11, 2020 9:30:37 GMT -6
Ha! Well there you go! Conspiracy theory about sadistic artificial intelligence debunked!
|
|
|
Post by anthropoid on Jan 11, 2020 9:48:26 GMT -6
I would say that if you have 2 different caliber on ship, it is better to have them for different target. Having 9" and 10" battery is not so good because both batteries will probably target same ships making both batteries less accurate and having one single battery is much better. If you use 9" with 6" battery it has much more sence as both batteries are primary against different targets. This means that at long range only one battery is firing and at shorter ranges smaller caliber battery can still have effects on unarmoured or less armoured part of ships if both batteries fire at one target.
Try to destroy CL target (scout) with main battery of heavy guns. You will need a lot of ammo to do it at long range. It is much pratical do it with smaller caliber guns as enemy CL cannot jeopardize heavily armoured battleship by guns, only by torpedoes.
Ok, question about this. Lets say we are talking about a CL with 6" mains and perhaps a LOT of them (one of the early default auto designs has about 12 or maybe even 14 I think). Lets say we are talking about my most recently designed B, the Alnair Class (did another reboot as Germany) with 4x12"(-1) and armor along the lines of B8 BE3.5 D2 DE1.5 CT8.5 T8 TT2.5 S7 and maybe . . . 8x6"(0). I might be a bit off on that armor scheme but it is a 16k ton boat built by France with medium range and normal engine characteristics. No Torps on the B at all. Now according to the "Gun Data" table, this Alnair Class B is "immune" to gun 12"(-1) gunfire at all the way down to 5000 yards. However, I have never taken that literally, as I'm aware of how much difference angle of impact and various other factors can make in whether or not a bullet penetrates a surface. Might it be better to simply replace all instances of the word "immune" with "resistant to?" Referring back to this CL design . . . I've had several engagements between ships at this relative extremes of armoring and firepower and my experience has been: no, a relatively heavy armor layout which the Gun Data table says offers "immunity" to heavy guns at ~>5000 yards does NOT mean that a smaller caliber "cannot jeopardize" it. It may only cause "light" damage, but if it scores a lucky hit, it seems it can also cause pretty serious damage, though sinking may not be much of a real threat. I see several of you folks who know more about the era and the game saying things like "lots of smaller guns isn't as good as a good amount of big guns . . . but lots of smaller guns can still get the job done." For example, one guy saying that he uses CLs with tons of 6" guns to "buzz saw" computer opponent capital ships. So in effect, I think the confusion to me is: I take words like "immune" and "cannot jeopardize" quite literally. If I'm in a tank and my opponent has nothing more than a sling shot and stones then yeah! I'm "immune" and he "cannot jeopardize" me! But if he has a .38 revolver . . . hmmm, he MIGHT be able to figure out some way to get on top of my tank and stick that gun muzzle in a port and cause havoc. It IS only a six shooter revolver so he'll have to reload and by then maybe one of our crewman can jump out and take him out . . . but, even with just that small step up in firepower, I'd say there is an increase in the chance to be "put in jeopardy." If he has an even more potent firearm then that risk of harm might well tend to go up! Now obviously 6"(0) shells at Tech level 0 or 1 are not quite the relative equivalent of a terrifying infantry manned anti-tank rifle . . . BUT maybe a 6"(2) battery fired from a thin-skinned CL with AP Tech in the 4 or 5 range IS the relative equivalent of a seriously jeopardizing egghshell armed with sledge hammer? That is the sense I have for how this stuff actually works and I'm gathering that the game with its params that represent gun quality and AP or explosive shell tech levels is also trying to model these kinds of dynamics?
|
|
|
Post by anthropoid on Jan 11, 2020 13:19:11 GMT -6
|
|
|
Post by trenton59 on Jan 11, 2020 16:56:34 GMT -6
For "immunity" against shell fire, it is best to remember that there is really no such thing, even a 90,000 ton battleship with maxed out late game armor can be sunk with enough 2" shells. Even if they do not penetrate the ships armor, every shell in the game has at least some high explosive filler, and will do some damage to the ships superstructure and hull, and can start fires, which is a common way to sink battleships early on, just pelting them with enough HE to burn them to a cinder.
As for HMS Majestic, that is the maximum, but below decks the barbettes had only 7" of armor, while the actual gunhouses were at most 10", this is also a fairly early armor steel, so it is thicker than later ships would use for the same level of protection. It averages to ~12.6" thick.
|
|
|
Post by rodentnavy on Jan 12, 2020 4:15:12 GMT -6
I hear ya, and I'm already seeing what you are talking about. The sad truth is: the player should NOT feel motivated to NOT have older ships on hand simply because the RNG prefers to put weaker entities into the battles. On the other hand, based on my own experiences programming (limited and sporadic) I have considerable empathy for Mr. Miller. This stuff is NOT easy to create. One thing "we" (I was only superficially and briefly involved but I was involved so . . .) on a team that worked on a different game I can neither confirm or deny the name of (NDA and all) but which covered a similar time frame and also covered lots of naval combat, were doing (not my idea): there was an "A.I." guy (I hate to use that term but I will for brevity) who basically was a scripter and was very well read in the nitty-gritty history of the war. He would write up pseudo code description that described specific operations for the bad guy nation (the one most players tended to play against in singleplayer) and these would include date frames (earliest, middle, latest date), explicit objectives (10,000 supply points in Guam, at least 6 DDs, 2 CAs, and 1 CVL in Rabaul, that sort of thing), and a sequence of preparatory steps, ready states and execution sequences. As you can see, this is already a bit more complicated than the operational considerations in this game. Anyway, these combined with a reasonable amount of "cheating" for the A.I. meant that the thing was pretty passable as an opponent. Just like everything in programming, if you break it down into small pieces and then get each cog in the machine to work with the one next to it you're off to a good start and you just keep building up from there. What might this sort of approach mean for a game like this? Well, just spit-balling . . . each position might have two or more "Strategic Paths" it could "choose" early in the game, and each of those paths might have splits in them that were at least in part decided by player actions. Each of those splits might have a set of operational objectives that would then have subsidiary build priorities, diplomatic actions or whatever. These then would also relate to actual pre-war dispositions and likely operational actions during a war. That is a ton of work for one guy though, and given the scope and production resources of the game, the current system is pretty damn good. ADDIT: I just reread what you wrote and what I wrote and it probably seems like the "operational" and even strategic level stuff I'm talking about doesn't make sense for how battle tactics would play out. I guess what I'm getting at is: tactics should be the sensible downstream result of strategy with operations as the intermediary that sets up a force to carry out strategy through tactics. At present, it is not clear that the computer controlled positions have much in the way of distinctive strategies either at the level of national peace time policy and action, or at the level of "diplomacy by other means." IF some semblance of "A.I. Strategy" and in particular at least some degree of national distinctiveness in strategy (a war against Russia should feel a bit distinctive to a war against the U.S. eh?) then the terms for programming differential tactics might also become more clear. I don't see an inherent problem with the "A.I." gaming points, just that it shouldn't feel like the ONLY trick it has up its sleeve? Probably just a bunch of rambling of no use, but maybe of some . . . The battle generator has a bias towards selecting newer ships for battles. I appreciate that you do not feel that way, but that is the way it is coded. There is of course always a chance/risk that if you keep older ships in active status, they will get selected in a battle. If you don't want them to be engaged, scrap them or put them in reserve.
As for the tactical AI counting VP:s. No, it doesn't do that. What it does is that it constantly evaluates several parameters like own strength, enemy known strength, ammo remaining, damage etc. The AI will generally disengage when it has a large proportion of its ships damaged , is low on ammo or feels outnumbered. It will not disengage based on VP status.
Well funnily enough I had a chance to test both halves of the statement shortly after this. My Mediterranean destroyer group decides to hit a convoy escorted by a lone light cruiser (representing the entirety of the enemy force in the region so not unexpected). Except the only ships missing from the force are the four modern (its the 20s) 1100 ton destroyers. That part of the program I think is broken and has been for some time. However the old destroyers do manage to tickle the modern 1920s era CL with their 3 and 2 inch guns and it withdraws which makes sense in light of your efforts to simulate psychology so that part is likely working as designed. Obviously being shot as is scary and deafening, actually shooting your own guns especially warship size ones can be a somewhat battering experience, something they skip over in the movies and it is a triumph that you can somewhat have the AI simulate it. However the force selection mechanism is broken and has been so for sometime. It is not that I resent having my old ships caught up in combat, that feels realistic*, what I do resent is knowing that it is my old ships that will be selected for first and all too often only. Because I believe you when you say the coding was meant to select for modernity but currently it is not working. The aim of airing this sort of thing here is to pool experience and highlight issues. *it can even lead to some epic engagements like the old protected cruiser leading German pocket battleship style BCs away from a convoy and thereby scoring a win despite getting sunk.
|
|