|
Post by director on Jun 3, 2023 13:38:09 GMT -6
I do not think the AI is cheating in terms of resources, and I respect Fredrick and the team for striving to make a non-cheating AI. A competent AI is quite difficult to write, so game developers often short-cut the process by having the AI do things differently from the way humans do - the AI of the CIV series basically gets stuff gifted every so often, rather than actually figuring out what to build, for one example. I do not think the RtW AI cheats in that way. But whether or not Fredrick and the gang intend it or not, and whether or not the AI simply can take advantage of knowing what the human can only suspect, the AI torpedo 'superpower' gives every appearance of being a cheat. When ships with underwater tubes, who can only aim torpedoes by turning the entire ship, whose crews cannot see the enemy and must rely on relays of orders and which require the ship to slow down lest the torpedo be broken, can accurately calculate target range, speed and course to fire torpedoes at targets over five miles away, then something is broken - no navy prior to WW1 had the machinery or training to consistently do that and maybe none ever. When enemy ships can be seen to know where incoming torpedoes are (while the human cannot) and repeatedly take enough evasive action to avoid a hit, then something is broken. I have, several times, warned people that the AI is driven to maintain a position ahead of your beam, which allows it to fire torpedoes 'downhill', taking advantage of the combination of your speed and the torpedo's speed. The high closing speed makes it easier to get hits for several reasons, and that positioning makes it almost impossible for you to make a successful torpedo attack in return. That's just good AI - maddening, yes, but sharp practice that any human player would do well to remember. It vastly increases the percentage chance of hits, regardless of who does it and it is - not - cheating. 'Sniping' with torpedoes at the extreme limit of their range from underwater tubes is unhistorical. Until the Japanese in WW2 (and possibly some of the Royal Navy DD skippers) no-one had torpedo-men who could correctly get range, bearing, target course and speed correct - certainly not from fixed underwater tubes, certainly not in the period before WW1... Most destroyers (including the Japanese) couldn't even get a hit with a spread from above-water tubes. And almost never did a target ship see an enemy torpedo coming in time to execute just enough of an evasive maneuver to skirt around it. Spotting a torp could happen, and evasion could happen, but we remember them because they were rare and not because one navy had ballet-dancing captains who made their ships torpedo-immune at any speed above zero. This is not a game-breaking problem. It is likely a minor matter of checking on what the AI knows and when, and making some adjustments. But the issue has been here since RtW0, it has been complained about since RtW0, and it is still here. And I ask you - if the AI isn't exploiting the rules in some way (perhaps just by its knowledge inside the movement cycle) then... in Admiral or Rear-Admiral mode, where torpedo attacks are run by the same AI, why don't my ships achieve the same 'sniping' hits the AI does? Most of the time the struggle is to get them to launch torpedoes at all... (see the positioning factor, above). t3rm1dor - I've played this game countless hours in all three versions, and I can tell you that in the scenarios where one or more of my CLs take on an enemy CA, my gunnery will reliably fail to score hits while the CA seems to get a bonus. With twice as many guns, good crew quality and gunnery direction, and traveling at comparable or higher speeds, I should be landing at least as many hits as the enemy - and instead, I land about half as many. Armored and heavy cruisers are good, but they aren't supposed to be magical. In addition, you may have noticed that 'crossing the T' is not a good tactic. Somehow, you get fewer hits firing down the length of the ship than the enemy does firing half his guns at your broadside. Fredrick has confirmed that this is 'working as designed' - wrong, undoubtedly, but working as intended.
|
|
|
Post by wlbjork on Jun 3, 2023 14:21:02 GMT -6
CLs do have issues - being generally smaller than CAs - tend to be less stable gunnery platforms and more vulnerable to weather conditions. Same as if you're unfortunate enough to suffer the double whammy of a Fire Control hit and a bridge hit, your accuracy will go down the pan. Morale can also play a part, I think. As others have said though, open the ship window and look at the accuracy data so you know what's happening.
And sometimes, the RNG just hates you. In my most recent case, it actually hated Russia as their CA landed no hits on my CA which quite happily murderdeathkilled the Russian ship.
|
|
|
Post by Enderminion on Jun 3, 2023 15:20:49 GMT -6
I posted this over on Matrix not sure how many of the same people are here. "I'm 100% guessing, but it's possible that what's driving people nuts may be that both the "Destroyers launch torpedo" logic and the "oh dang that destroyer might launch" logic are making the exact same calculations. That leads to the AI knowing it only needs to evade when there's a high probability a torpedo is in the water, otherwise it can roll along fat dumb and happy. As soon as a destroyer enters launch parameters, that by default means the AI target knows it's in the same launch parameters... thus a perception that it's got some special knowledge and is avoiding the *actual* torpedoes." Basically, if both AI actions (launch and evasion) use the same algorithm, there's no conservatism (extra unneeded evasion) or risk taking and poor judgement (failure to evade) built into the AI's digital helmsmen. They will always evade only when they need to. Frederick and WilliamMiller have long stated (since RTW1 release or so... ) that there's nothing in the code giving the AI an inherent advantage, and given how open and engaged they are regarding just about everything I see no reason to believe otherwise. Exactly this, I've been trying to say that.
Regarding cruiser gunnery stuff, hits can compound pretty hard, if the 8" cruiser got lucky and scored a couple good hits early on, that advantage compounds as the 6" cruiser's accuracy was degraded.
|
|
|
Post by cormallen on Jun 3, 2023 15:25:52 GMT -6
Like everyone I've suffered from the AI's eldritch talents with the torpedo. I'm well aware of it's fixation with pulling lead on my squadrons, and I've even taken advantage of it tactically to tease their battle line apart.
I suspect the AI is just super aware of the launch parameters for torpedo attack and thus just extra on-the-ball both offensively and defensively. The former is quite annoying, but at least not entirely unfeasible, but the latter is (given the difficulty in reliably spotting individual torpedoes) decidedly unrealistic.
The player can even the odds by using Captain's mode and manually launching his own torpedoes but I find that a bit too much like an exploit in itself!
Ideally we should probably see more torpedoes launched optimistically during "Flotilla Attacks" producing the occasional hit after much expenditure perhaps? And, only the odd opportunist launch otherwise?
The magic of course involves a clever chap writing the code that makes this behaviour come to life, and that's WAY beyond my skills!
As it is we have the game we have, with its joys and wrinkles both.
|
|
|
Post by zederfflinger on Jun 5, 2023 20:49:07 GMT -6
The T-crossing thing seems rather odd to me. If that was the situation every admiral wanted to be in historically, why isn't this something you want to achieve in-game? Seems like a fairly major issue to me.
|
|
|
Post by wlbjork on Jun 5, 2023 23:06:36 GMT -6
The T-crossing thing seems rather odd to me. If that was the situation every admiral wanted to be in historically, why isn't this something you want to achieve in-game? Seems like a fairly major issue to me. 'Crossing the T' was great in the Age of Sail, when your weapons were hull mounted. Sail along in front, or better still behind, the enemy ship and unleash your broadside into their hull, where the cannonballs could fly relatively unimpeded for the length of the enemy ship. By the time of the pre-dreadnought battleships, things had changed drastically. The main guns were mounted in turrets above the deck and secondary guns were - if casemated - on the upper deck or two. Additionally, the citadel was completed by armoured bulkheads fore and aft and compartmentalisation meant that the shells were impeded. So the value of crossing the T had degraded to the point that the only real advantage was having fewer guns firing at you whilst you could still use your full broadside. I suspect the smaller profile of the ship from bows or stern on would also interfere with range-finding and other fire control equipment.
|
|
|
Post by zederfflinger on Jun 6, 2023 11:03:44 GMT -6
The T-crossing thing seems rather odd to me. If that was the situation every admiral wanted to be in historically, why isn't this something you want to achieve in-game? Seems like a fairly major issue to me. 'Crossing the T' was great in the Age of Sail, when your weapons were hull mounted. Sail along in front, or better still behind, the enemy ship and unleash your broadside into their hull, where the cannonballs could fly relatively unimpeded for the length of the enemy ship. By the time of the pre-dreadnought battleships, things had changed drastically. The main guns were mounted in turrets above the deck and secondary guns were - if casemated - on the upper deck or two. Additionally, the citadel was completed by armoured bulkheads fore and aft and compartmentalisation meant that the shells were impeded. So the value of crossing the T had degraded to the point that the only real advantage was having fewer guns firing at you whilst you could still use your full broadside. I suspect the smaller profile of the ship from bows or stern on would also interfere with range-finding and other fire control equipment. You would be more likely to score hits though? If your battleship's shells land in a horizontal spread, it seems that you would score hits more frequently if the opponent is bow-in. You wouldn't get the super accurate salvos where you score 4+ hits, but that doesn't really happen in RTW anyways.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jun 6, 2023 11:31:35 GMT -6
I posted this over on Matrix not sure how many of the same people are here. "I'm 100% guessing, but it's possible that what's driving people nuts may be that both the "Destroyers launch torpedo" logic and the "oh dang that destroyer might launch" logic are making the exact same calculations. That leads to the AI knowing it only needs to evade when there's a high probability a torpedo is in the water, otherwise it can roll along fat dumb and happy. As soon as a destroyer enters launch parameters, that by default means the AI target knows it's in the same launch parameters... thus a perception that it's got some special knowledge and is avoiding the *actual* torpedoes." Basically, if both AI actions (launch and evasion) use the same algorithm, there's no conservatism (extra unneeded evasion) or risk taking and poor judgement (failure to evade) built into the AI's digital helmsmen. They will always evade only when they need to. Frederick and WilliamMiller have long stated (since RTW1 release or so... ) that there's nothing in the code giving the AI an inherent advantage, and given how open and engaged they are regarding just about everything I see no reason to believe otherwise. No one says they programmed the AI to cheat on purpose. But the fact is that the amount of BS hitscans is just mindblowing. I mean, my DD flotilla can perform an attack from an advantageous firing position against enemy dreadnoughts, only to leave with 0 torpedo hits, and 2 of my DDs getting torped by freakin dreadnoughts. If that happened once per save file, or once per war, or once per year, I would not really care. But it happens in at least HALF of all battles. My BC gets closer than 10 000 yards to enemy BB? Torpedo hit. No matter how fast I go or how much I maneuver. My CL gets 6000yds away from an enemy CL? Torpedo hit. My DD flotilla swarms an enemy DD flotilla resulting in a very close range combat? 5 torpedo hits for the enemy, 0 for me. My flotilla sends a wave of 20 torpedoes at 5000yds at night into enemy battle line? 0 hits, at least on the enemy side, because one of my DDs got hit by a freakin underwater launcher on an early dreadnought. Right now I just don´t play the game, as it is generally just infuriating cycle of try, get bullshitted, restart the game, try again, get bullshitted again in a slightly different way...
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jun 6, 2023 11:42:24 GMT -6
CLs do have issues - being generally smaller than CAs - tend to be less stable gunnery platforms and more vulnerable to weather conditions. Same as if you're unfortunate enough to suffer the double whammy of a Fire Control hit and a bridge hit, your accuracy will go down the pan. Morale can also play a part, I think. As others have said though, open the ship window and look at the accuracy data so you know what's happening. And sometimes, the RNG just hates you. In my most recent case, it actually hated Russia as their CA landed no hits on my CA which quite happily murderdeathkilled the Russian ship. IIRC the game has a modifier which makes your ship less likely to hit with gunfire if its currently being hit by enemy gunfire, so the first one to start scoring hits (usually the one that has either better RNG, better FC, or more guns capable of firing at that target at the moment) is more likely to just chew the enemy and spit out a burning wreck. I do not have a problem with that, although it makes the pre-dread hunt with my super CA´s during the mid 1900s a bit too easy, as having a broadside of 6 or 8 8in guns is generally much better than having 4x12in (-2) guns, as the fire risk in this game is just absolutely devastating to the point that Im sinking battleship after battleship without penetrating their citadel just once. But OK, thats lets say, plausible. I can accept that. And it sometimes happen even to my ships, when I build an early cross-deck fire (Von der Tann-ish) BC and the enemy has a hexagonal BB that can fire 8 gun broadside a bit better than my ships. Also, Russians have the retarded crew debuff which makes them less likely to do anything meaningful even if their ships arent a total pieces of junk, so thats also fine. What makes me really mad is when Im shelling the enemy BB battleline from "upstream" (being slightly ahead of them) with my BCs going 27kts and the freakin dreadnoughts start hitting 1912-era torpedoes upstream at my BCs at 8-12k yds. And when that happens twice in an hour, I really feel properly infuriated and my gaming mood is ruined.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jun 6, 2023 12:06:31 GMT -6
The T-crossing thing seems rather odd to me. If that was the situation every admiral wanted to be in historically, why isn't this something you want to achieve in-game? Seems like a fairly major issue to me. Crossing the T may be useful if your side armor can take a hit (for example a BB, or CA if you only face CLs and some crappy CAs). Especially if you have turrets that cannot fire forwards or backwards. Such turrets cannot be used in a chase fight anyway, and if the enemy has for example just a single forward turret, and you have a forward turret, aft turret and two midship turrets (like for example Gangut class), you have about 4 times the firepower. And even though the angle of enemy belt armor basically prevents a clear penetration, you are very likely to just demolish enemy bow section, superstructure and likely also the front turret, softening the target for the following broadside combat so that once the enemy turns its broadside, its already halfway to the bottom. I have found out that if you do that with a ship that has really big guns (for example if your 1st gen BB has 14in guns for some reason) you can crush your enemy in this process so badly that instead of turning broadside it immediately starts running (more like "crawling" away due to heavy damage). I think that there may be some condition under which a bow/aft hit may count as a belt hit or engine room hit, because such huge hits into the bow section can rapidly slow down the enemy ship without hitting its engine room through the belt armor. And no, it isnt caused by the flooding as flooding takes time, but these hits can slow down the enemy instantly. On the other hand, I cant think of a single scenario when attacking enemy broadside head-on would produce a favorable results. In RtW3, it looks like even a 4 gun pre-dreadnought gains a massive hit-chance bonus when it can use all of its 4 main guns.
|
|
|
Post by t3rm1dor on Jun 6, 2023 12:32:27 GMT -6
t3rm1dor - I've played this game countless hours in all three versions, and I can tell you that in the scenarios where one or more of my CLs take on an enemy CA, my gunnery will reliably fail to score hits while the CA seems to get a bonus. With twice as many guns, good crew quality and gunnery direction, and traveling at comparable or higher speeds, I should be landing at least as many hits as the enemy - and instead, I land about half as many. Armored and heavy cruisers are good, but they aren't supposed to be magical. In addition, you may have noticed that 'crossing the T' is not a good tactic. Somehow, you get fewer hits firing down the length of the ship than the enemy does firing half his guns at your broadside. Fredrick has confirmed that this is 'working as designed' - wrong, undoubtedly, but working as intended. Crossing the T is useful deepending on the situation - a semi spherical crossing for instance do allow concentration of firepower on enemy lead ships. Also that crossing the T is extremely useful for cutting any enemy force ability to manuvre. Regarding CA the two main advantages I can see working for then is range and that at some points are able to mount better fc in regards to accuracy. Another thing entirely is armor and damage per hit, I suspect that what you are experiencing is a faster degradation of cl performance compare to CA when in combat conditions. But in close ranges nigth figthing for example, an auto 6 inch CL can make very short work of an enemy CA.
|
|
|
Post by dorn on Jun 7, 2023 9:26:14 GMT -6
I posted this over on Matrix not sure how many of the same people are here. "I'm 100% guessing, but it's possible that what's driving people nuts may be that both the "Destroyers launch torpedo" logic and the "oh dang that destroyer might launch" logic are making the exact same calculations. That leads to the AI knowing it only needs to evade when there's a high probability a torpedo is in the water, otherwise it can roll along fat dumb and happy. As soon as a destroyer enters launch parameters, that by default means the AI target knows it's in the same launch parameters... thus a perception that it's got some special knowledge and is avoiding the *actual* torpedoes." Basically, if both AI actions (launch and evasion) use the same algorithm, there's no conservatism (extra unneeded evasion) or risk taking and poor judgement (failure to evade) built into the AI's digital helmsmen. They will always evade only when they need to. Frederick and WilliamMiller have long stated (since RTW1 release or so... ) that there's nothing in the code giving the AI an inherent advantage, and given how open and engaged they are regarding just about everything I see no reason to believe otherwise. No one says they programmed the AI to cheat on purpose. But the fact is that the amount of BS hitscans is just mindblowing. I mean, my DD flotilla can perform an attack from an advantageous firing position against enemy dreadnoughts, only to leave with 0 torpedo hits, and 2 of my DDs getting torped by freakin dreadnoughts. If that happened once per save file, or once per war, or once per year, I would not really care. But it happens in at least HALF of all battles. My BC gets closer than 10 000 yards to enemy BB? Torpedo hit. No matter how fast I go or how much I maneuver. My CL gets 6000yds away from an enemy CL? Torpedo hit. My DD flotilla swarms an enemy DD flotilla resulting in a very close range combat? 5 torpedo hits for the enemy, 0 for me. My flotilla sends a wave of 20 torpedoes at 5000yds at night into enemy battle line? 0 hits, at least on the enemy side, because one of my DDs got hit by a freakin underwater launcher on an early dreadnought. Right now I just don´t play the game, as it is generally just infuriating cycle of try, get bullshitted, restart the game, try again, get bullshitted again in a slightly different way... If your DDs are hit from CLs, you are clearly not advategous position as the difference in speed between DD and torpedo is not high. It have never happened to me to have DDs in advantegous position and get torpedoed.
If you really see same cases, please provide save ....
I get quite regular torpedo hits during night without any problem but bear in mind that 3 % hit probability and even less is nothing wrong. AI can have higher hit rates than you if you play more offensively/agreesively than cautions AI, which most players including me do.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jun 9, 2023 15:03:12 GMT -6
OK, AI 1915 DDs now not only auto hit any ship, but they are also able to reload their torpedoes to hit again...
|
|
|
Post by Enderminion on Jun 9, 2023 15:21:02 GMT -6
OK, AI 1915 DDs now not only auto hit any ship, but they are also able to reload their torpedoes to hit again... It's entirely possible for them to have had additional torpedo launchers on the other side of the destroyer, or to not have fired all of their torpedoes the first time around.
|
|
|
Post by christian on Jun 10, 2023 6:54:30 GMT -6
If I remember well in WW2 Japanese long lances has about 3 % hits, so for every 100 torpedoes fired only 3 hit target so RTW is quite realistic in that way.
Torpedoes in WW2 was mainly for night attacks when they can be deadly or used for harassing enemy or force them away at least temporary.
actual hitrates were 6,71% which all things considered is very good. www.navweaps.com/index_tech/tech-067.php
|
|