Am I doing something horribly wrong here?

[f105d]

A little background: I'm playing 1900 USA start without historical resources, and with the Ai wars mod. And the year is 1916 a good 10 years in game after dreadnaughts became a legitimate naval force and not just being constructed. Along with most of the world in some bizarre 3 way world war between everyone and I'm only fighting the Russians.
So I went to check the almanac to see what amount is just in service since I had not checked since 1911 and I see for dreadnaughts across the board it seems pretty even world wide... then I see the Royal navy has 15 with another 5 in production and Imperial Germany has 11 with 4 in production, Battle cruisers 6 with 2 in production for GB and 9 with 1 in production.
I only have 7 dreadnaughts with 1 currently in production, 3 on hold along with 2 more classes waiting in the wings for the R&D funds. And Battlecruiser wise I only have 3 with 1 in production and 2 more waiting. But I'm updating my cruiser fleets and destroyers as well since it has for cruisers been a good 7-8 years since one was commissioned and well destroyers is around the same time frame.

So am I just doing my whole production wrong (not the usual build specific types for a given year in a cycle) or does the AI just have the ability to just build more ships?

[noshurviverse]

Are you placing ships in reserve or mothballs in times of peace? What about training, do you always have it active? Finally, what kind of Dreadnoughts are you building? If you're making, top-of-the-line behemoths while the AI is going for more traditional designs that could help explain the disparity.

[aeson]

Some potential contributing factors:

- Training. A 50% increase in fleet upkeep for Gunnery and (Night Fighting or Torpedo Warfare) with as many ships as the top powers have could easily be an extra capital ship or two every design generation.

- Research budget. I haven't checked, but I suspect that the computer normally uses between 8% and 10% research budget; if you use 12%, that's a difference that can add up over time.

- Fleet Readiness. Ships in the Reserve Fleet cost half and ships in mothballs cost one-fifth as much to maintain as ships in the Active Fleet in home waters; while it usually doesn't make that much of a difference right now since construction costs are generally much higher than upkeep on older ships - especially when there is a large capability gap, as between an early ~30,000t superdreadnought and a ~15,000t predreadnought - it does add up over time, just like with the research budget.

- Overseas service. AF and FS ships outside of a home sea zone (any sea zone containing a home possession - the North American East Coast, the Caribbean, and the North American West Coast for the USA) costs 20% more to maintain in time of peace than AF and FS ships inside of a home sea zone. Again, not much of an immediate impact, but it can add up over time.

- Refits. The computer only does fire control and basic maintenance (blank) refits and what look to be 'cheap' carrier conversions; if you shell out for engine rebuilds or gun/armor upgrades, you're spending considerably more on refits than the computer normally does. Major reconstructions can easily cost a third or more as much as building a comparable ship from the keel up.

- Coastal batteries. The computer usually builds a handful; how many do you build?

- Submarines. Usefully-large submarine fleets aren't cheap, and maintaining the scale of the fleet in the face of wartime losses can be very, very expensive - my ballpark estimate when using submarines is that I'll need to replace on average two submarines every turn of a war to cover losses. A medium-range submarine costs 3.2 million to build over 16 months, or 200 thousand per month. If I want to be able to replace two of them every month to cover wartime losses, that's 32 medium-range submarines building simultaneously - 6.4 million per month, or about the same as a 50,000t battleship. Doing the same with coastal submarines is 'only' 2.8 million per month, or about the same as a 25,000t battleship.

- Aviation. If you want early (primarily land-based) aviation to be useful in battle, you either need a very good idea of where you're going to be fighting, or you need to build airship bases and airfields all over the place. It might not be particularly expensive, but building and maintaining airship bases and airfields stocked with flying boats isn't exactly cheap, either.

Also, I'm not completely certain, but it looks to me like your capital ships are on average slightly larger than your opponents' capital ships, so they most likely cost a bit more.

Finally, remember that there is a degree of hysteresis with regards to fleet size and budget - the USA starts with a (relatively) small budget that grows rapidly whereas Britain and, on default (game) budget setting, Germany start with (relatively) large budgets that grow somewhat more slowly. This means that Britain and Germany can typically build larger fleets whereas the USA tends to have relatively newer fleets early on. The tipping point - when the USA's relatively fast budget growth finally gets it to the point of having about as many dreadnought-type warships as Britain or Germany - tends to come somewhere in the late-1910s to mid-1920s

[revillag]

That screen also does not count ships undergoing refit. You show 3 BC for 69,000 tons with no BC building but the total shows 156,000 tons. This means that you have at least one BC undergoing refit that is not being counted. As the previous poster stated, a major refit can easily cost 1/3 to half of a new ship. Also, I have found that the AI does not scrap old ships until they are extremely obsolete.

[christian]

most of the time the ai never has gunnery or torpedo training on which makes them a pushover in normal battles even when you are outnumbered because they cant hit

the ai can generally pump out an insane amount of ships but refuses to build ships which are quality over quantity

this usually leads to quality ships being able to whop their ass because a ship with 18 inch belt and 7 inch deck with 9 18 inch guns and 26 inch turret armor litteraly does not die in a battleship to battleship fight and it usually over 5 capital ships to run down and outgun a 90k supership purely due to the fact the 90k supership is usually immune above 10k yard range



also the ai is horrible as a fleet admiral and isent good at tactics
combine this with the fact they have no training their designs are bad and quantity focused
they suck


in short they mass produce shitty ships resulting in extreme tonnage numbers but with **** performance


the larger the fleet size you play on the more apparent it becomes


having played on fleet size 13 i got 10 90k superbattleships out and 8 supercarriers of 89500 tons (refit giving them 90k but with better torp prot) while usa and great britain had pumped out a total combined force of over 90 battleships/battlecruisers and 70 aircraft carriers

their tonnage is extremely overwhelming but slowly whittling down their numbers becomes quite easy when you realize their ships dont have guns bigger than 16 inches and they never have ships larger than 50k tons and their armor sucks







also the VP system heavily favors single ships of high quality large tonnage over multiple small tonnage low quality ships (also because a ship 2x the tonnage is usually more than 4x as effective due to pen thresholds)


also to be noted 90k ships are immune to all but torpedoes as long as you have a 7 inch deck (1400lb divebomber bombs fail to pen 7 inch decks and no shells can pen the deck at any range and as long as you have a 16-18 inch belt you are good)

[dorn]

The game is not built around such large fleet. It will be certainly unbalanced.

And actual issue with battle generator even highlight that.

[christian]

Sept 15, 2019 23:13:43 GMT -6 dorn said:

The game is not built around such large fleet. It will be certainly unbalanced.

And actual issue with battle generator even highlight that.

indeed but it dosent change the fact the ai produces an insane amount of battleships and other capital ships compared to the amount of cruisers they build 


but the ai still lacks the quality training and practise of your ships which in turn usually makes them fall short in actual battles despite you being outnumbered 2 to 1

[jwsmith26]

christian , I think you might be overestimating the value of training. When you examine the actual factors that impact rate of fire and accuracy, training is just one modifier among dozens. It has less impact than turning your ship while firing or smoke on the battlefield. My experience indicates that the cost to train is excessive for the advantage conferred. In my opinion, the AI's choice to not train is the correct one.

Now ship quality, that's a different story. The AI pretty much sucks at that. As far as the game overproducing battleships and capital ships, every RTW2 game is different. For instance, I have seen plenty of games where battleships are largely ignored by the AI in favor of heavy cruisers (which is weird but fun during battles). RTW2 is a long and complex game. When you push the boundaries of the game, such as by creating a large fleet of 90K ton ships, it can react in unusual ways. Again, that's part of the fun.

[aeson]

Personally, I'm more inclined to attribute the computer's relatively poor performance in battle to low crew quality due to keeping ships in lowered states of readiness in peacetime and sending ships haring off through sea zones without sufficient support capacity than to a lack of training. Gunnery training, after all, is only a +10 to hit chance - as jwsmith notes, a wide variety of other factors are easily as significant. Crew quality, though? That affects everything important, as far as I can tell, and Poor crew quality in particular is in my experience a serious detriment.

[christian]

Sept 16, 2019 11:13:40 GMT -6 jwsmith26 said:

christian , I think you might be overestimating the value of training. When you examine the actual factors that impact rate of fire and accuracy, training is just one modifier among dozens. It has less impact than turning your ship while firing or smoke on the battlefield. My experience indicates that the cost to train is excessive for the advantage conferred. In my opinion, the AI's choice to not train is the correct one.

Now ship quality, that's a different story. The AI pretty much sucks at that. As far as the game overproducing battleships and capital ships, every RTW2 game is different. For instance, I have seen plenty of games where battleships are largely ignored by the AI in favor of heavy cruisers (which is weird but fun during battles). RTW2 is a long and complex game. When you push the boundaries of the game, such as by creating a large fleet of 90K ton ships, it can react in unusual ways. Again, that's part of the fun.

what i personally find is that extra hitrate really pays off especially early game and early battle

the first one to get a hit in usually takes the advantage

ship battles the enemy usually cant bring in more than 2x your fleet size and thus when you already cant outnumber them crew quality and ship quality usually triumph

i usually always use gunnery training because 10% is quite a big deal to me

[dorn]

Sept 16, 2019 14:49:24 GMT -6 christian said:

Sept 16, 2019 11:13:40 GMT -6 jwsmith26 said:

christian , I think you might be overestimating the value of training. When you examine the actual factors that impact rate of fire and accuracy, training is just one modifier among dozens. It has less impact than turning your ship while firing or smoke on the battlefield. My experience indicates that the cost to train is excessive for the advantage conferred. In my opinion, the AI's choice to not train is the correct one.

Now ship quality, that's a different story. The AI pretty much sucks at that. As far as the game overproducing battleships and capital ships, every RTW2 game is different. For instance, I have seen plenty of games where battleships are largely ignored by the AI in favor of heavy cruisers (which is weird but fun during battles). RTW2 is a long and complex game. When you push the boundaries of the game, such as by creating a large fleet of 90K ton ships, it can react in unusual ways. Again, that's part of the fun.

what i personally find is that extra hitrate really pays off especially early game and early battle

the first one to get a hit in usually takes the advantage

ship battles the enemy usually cant bring in more than 2x your fleet size and thus when you already cant outnumber them crew quality and ship quality usually triumph

i usually always use gunnery training because 10% is quite a big deal to me

I completely agree with you. In this case it is mathematics. As to hit modifier from training is value added to base it means that it has higher % effect with lower base. And lowest base is at game start.

I find it quite useful in RTW1 in cruiser battles but not too important in RTW2 as cruiser battles are not so often as in RTW1.

[valikdu]

Sept 13, 2019 22:41:30 GMT -6 f105d said:

So am I just doing my whole production wrong (not the usual build specific types for a given year in a cycle) or does the AI just have the ability to just build more ships?

I think even the AI war mod can't sink ships yet, unless I misread? So, the problem is really that nothing is being sunk except in a war that you participate in. That was my impression.
 

[jwsmith26]

I don't believe that a +10 hit modifier translates directly to a 10% increase in hit chance. If I'm wrong please correct me.

[aeson]

Sept 16, 2019 15:43:23 GMT -6 jwsmith26 said:

I don't believe that a +10 hit modifier translates directly to a 10% increase in hit chance. If I'm wrong please correct me.

The below is a set of accuracy reports for a small cruiser firing at another small cruiser at various engagement ranges during a January 1900 fleet exercise:

I do not know how the modifiers are applied to get from the listed Basic Hit Chance to the listed Final Hit Chance, but it is very clearly not the case that it's [final hit chance] = product(1 + mod_n/100, n=1:N)*[basic hit chance], and it's also very clearly not the case that it's [final hit chance] = [basic hit chance] + sum(mod_n/100, n=1:N). That being so, I cannot see any straightforward way in which a +10 to hit chance could be interpreted as a direct 10% increase in final hit chance, unless positive modifiers are handled differently from negative modifiers.

[tortugapower]

Sept 16, 2019 15:04:18 GMT -6 valikdu said:

Sept 13, 2019 22:41:30 GMT -6 f105d said:

So am I just doing my whole production wrong (not the usual build specific types for a given year in a cycle) or does the AI just have the ability to just build more ships?

I think even the AI war mod can't sink ships yet, unless I misread? So, the problem is really that nothing is being sunk except in a war that you participate in. That was my impression.
 

I hope this doesn't derail the thread, but for clarification: the mod will not sink AI ships unless you quit the game and reload the Autosave after every month.

Correct operation for the AI Wars mod, if you want the game to update ships sunk properly, is:

- End turn to finish month
- Close game
- Load Autosave.bcs for current game

(No need to do this process for any turn where no ships are sunk.)