MYTH: Japan has a Merchant Shipyard Surplus: BUSTED

[erstad]

Japan starts with a merchant shipyard capacity that is around double the day 1 needs. This has led several folks to decide to turn off merchant shipyards and/or convert to naval.

However, WitpStaff says that in a year, demand will outstrip supply. Is that a bubble or sustained?

A sufficiently clever and motivated person would figure out how to take the CSVs from WitpStaff and generate a complete profile. Not being sufficiently motivated (which means we'll never find out if I'm sufficiently clever), I just ran an AI game in continuous mode. The assumption is the AI won't expand merchant yards (verified, end value was same as beginning), doesn't shut down yards, and doesn't halt/accelerate ships (at a couple of the stopping points, I looked at the queue and generally nothing was halted/accelerated except for maybe a dozen MTB that are probably a special case.

So here's the accumulated merchant ship points:
03/19/42: 44,308
08/02/42: 96,335
10/16/42: 107,938
03/03/43: 99,740
06/26/43: 91,996
09/06/43: 75,274
12/16/43: 19,788
01/17/44: 0!

So, even keeping the initial merchant shipyards running consistently, you will run out of merchant shipyard points by the end of 43/early 44.

This assumes you want everything they produce, of course.

[n01487477]

AI games are problematic for testing in that the normal rules to shipbuilding don't apply ... I've tested this. My calculations have it at an average of 1082 Merchant points needed for the course of the war to build everything.

[Mike Solli]

Damian, is that without accelerating anything?

[n01487477]

Yeah Mike,
no acceleration, the calculation is a simple extrapolation iirc of durability of merchant ship list (which are delayed) * delay = Total Merchant Points needed

Then calculating, Total Merchant Points needed / Total no. of days in the war (remaining).

Obviously this doesn't account for spikes ... and is not the best way to calculate, but this is a simple beginning.

Cheers
Damian

[tigercub]

Thanks noted Damian,

I read all your input and love your work mate!
[&o]


Tiger!

[tigercub]

So knowing what has been said now do i change at start of campaign or wait until later in the war?


Tiger!

[erstad]

ORIGINAL: n01487477

AI games are problematic for testing in that the normal rules to shipbuilding don't apply ... I've tested this. My calculations have it at an average of 1082 Merchant points needed for the course of the war to build everything.

All the more reason we need to advocate getting "continuous" mode for head-to-head play!

I didn't know the AI shipbuilding was different, but your conclusion is the same - the initial merchant shipyards are insufficient for the total buildout. Can I get a text or excel file of the usage by day or whatever? I'd like to compare it to the AI game.

[Shark7]

Given that I like to accelerate many of my merchant type ships, I was already planning on increasing it to between 1200-1300 on the merchant shipyards. The shipyards are actually the cheaper HI users, and once you are done with them, you can just turn them off.

[Woos]

A sufficiently clever and motivated person would figure out how to take the CSVs from WitpStaff and generate a complete profile.

Actually one only needs to be sufficiently proficient in SQL, connect to the WitpStaff database mentioned in the Tech Support Thread and give this query to the database

Code: Select all

SELECT SUM(CASE WHEN "turnsTillAvail"-365<=a."durability"*10 THEN a."durability"*("buildspeed"*2-1) ELSE a."durability" END) 
 FROM "ShipBuildqueue" AS a 
 JOIN "Ships" AS b ON a."shipID"=b."shipID" 
 JOIN "ShipClasses" AS c ON "shipclass" = "shipclassID" 
 JOIN "ShipTypes" AS d on c."shiptype"=d."shiptypeID"
 WHERE "turnsTillAvail">365 AND "buildspeed">0 
 AND "turnsTillAvail"-365<=a."durability"*10*("buildspeed"*2-1) 
 AND NOT d."navalProduct"

You need to change the 365 to whatever number of turns you want to look into the future. This will give you the amount of Merchant build points cosumed at that point in time assuming nothing is accelerated, stopped, ... .
When I did this a year ago, there was a peek end of 1943 with a requirement of ca. 1600 buildpoints. Probably changed in the mean time.

If you want to sum build cost, the following query helps:

Code: Select all

SELECT SUM(CASE WHEN "turnsTillAvail"<= 1272 THEN CASE WHEN "turnsTillAvail"-a."durability"*10<0 THEN "turnsTillAvail"* a."durability" ELSE a."durability"*a."durability"*10 END
 ELSE CASE WHEN "turnsTillAvail"-a."durability"*10<0 THEN 1272* a."durability" ELSE a."durability"*(1272-("turnsTillAvail"-a."durability"*10)) END END)
 FROM "ShipBuildqueue" AS a
 JOIN "Ships" AS b ON a."shipID"=b."shipID"
 JOIN "ShipClasses" AS c ON "shipclass" = "shipclassID"
 JOIN "ShipTypes" AS d on c."shiptype"=d."shiptypeID"
 WHERE "turnsTillAvail">0 AND "buildspeed">0
 AND "turnsTillAvail"-a."durability"*10<=1272
 AND NOT d."navalProduct"

Depending on what you replace 1272 with it gives you the overall amount of buildpoints needed up to that date.

Reading through the old thread on the development forum shows that the above analysis has seemingly caused a reduction of IJ merchant yards *vanishes silently from the forum before the IJN fanboys catch up to this*.

[n01487477]

ORIGINAL: Woos

A sufficiently clever and motivated person would figure out how to take the CSVs from WitpStaff and generate a complete profile.

Still testing - sufficiently motivated,& maybe too not dumb - I believe[;)]

[n01487477]

Oops posted the wrong one ... obviously not smart enough [;)]

*** edit note: my previous calculations were also wrong about the Optimum Merchant HI needed too, revised to 887, you only need to build more if you want to deal with the peak...
AND the same goes for Naval Build points which have now been revised down to 1236

BIG SORRY guys!!! I was using an old algorithm, that dealt with the wrong number of turns in the GC.

[tigercub]

They start with 1386 naval yards so i have some up my sleeve, Merch needs a little expanding,hey you guys think HI needs to be expanded or is this still unclear, my aim as a Japanese player is to expand as little a possible but still get what i need to win with out killing my fuel supply's.

Tiger!

[n01487477]

Mate,
I wouldn't be converting anything to Merchant build, Japan has enough AK /AP's IMO. Having to delay some of the merchant build past its due date is not a biggy.

The Naval build is far more important and if you look at the peaks around day 450 ~ 1100, you're looking at an average of 1500 Naval Pts / day, maybe these points would be accumulated prior to this, but given my desire to accelerate, I'm sure it probably won't be.

I think there is more analysis to do here, but given the problem with supplies, I'd be holding off on converting or building too much.

Cheers
--Damian--

[Bliztk]

One question, what classes use Merchant points, and what classes use Naval shipyard points ?

I`m preparing an excel to make proyections, and need to be 100% sure

[Historiker]

ORIGINAL: Bliztk

One question, what classes use Merchant points, and what classes use Naval shipyard points ?

I`m preparing an excel to make proyections, and need to be 100% sure

http://hc-strategy.com/ae/wiki/index.ph ... _type_list

[Historiker]

ORIGINAL: n01487477

Oops posted the wrong one ... obviously not smart enough [;)]

*** edit note: my previous calculations were also wrong about the Optimum Merchant HI needed too, revised to 887, you only need to build more if you want to deal with the peak...
AND the same goes for Naval Build points which have now been revised down to 1236

BIG SORRY guys!!! I was using an old algorithm, that dealt with the wrong number of turns in the GC.

This looks like you have to have around 1200 merchant points to build a pool big enough to provide the necessary points for the peak. After crossing it, one can shut down more and more merchant shipyards.

[Bliztk]

Here is my take on Merchant Shipyard.

Average from 41/12/7 to 43/12/31 is 909 shipyard points without accelerating anything

[n01487477]

Blitzk,
glad to see my data and your data looks almost the same, finally must have done something right !
 
Cheers
Damian

[Bliztk]

Yes, but I have done it manually !

If I could programme some macros, my life would be easier.

I`m now tackling with accelerating ships, what ships and at what cost [X(]

[bklooste]

It makes sense Japan massively expanded there merchant building in 44 ( eg 44 was almost 8 times 41-42 and almost 2.7* 43) ) when under strat bombing , kind of makes you wonder why they didnt do such increases in 40...

Warships Merchant Warplanes
1939 ? 320,466 ?
1940 94,700* 293,600 4800
1941 225,000 210,400 5100
1942 254,000 260,100 8900
1943 230,000 769,100 16,700
1944 468,000 1,699,200 28,200
* reduced due to Yamato and carriers

Only way you can replicate this is grow your merchant yards 8 fold in 43 , supplies anyone ?

Personally I believe Yamato and Musashi were accelerated historically and shipyards should reflect this capability.

Ben.