Japanese Merchant Convoy Composition

[Mike Solli]

This past weekend, I tried to figure out how I want to set up the cargo fleets in the Home Islands.  After much brain sweat, I think I figured out how to do it (last night in bed).  I have a list of what ships are in each port.  I know what the convoy routes are for each port outside of Honshu and how long for one lap.  I know what ships are currently available.  I'm going to total them by type, figure out what goes where, and then decide where each individual ship will go.  I figure I can do that in a couple of hours tops instead of a week (at the rate I was going). [:D]

[Graymane]

Not sure if this is the answer or not, but do you get a full 2 phases in the first turn? I think whenever I do a first turn as Allies, My ships only move 1/2 of what they should, so maybe yours aren't getting a full day of unload capacity?

[vonSchnitter]

Well Mike,

I do not know what you are smoking - joke - but I am quite sure your wet dreams on ideal tf composition solutions will collide with the game realities rather harshly. For two reasons:
As of yet, the system is not working as intended or/and
the system uses more parameters than the rule book reveals.

Whatever - both will have a rather uncomfortable go at you family yewels ... [:D]

Now - just to complicate matters:
How about invasion TFs ?

The bad news is: Any base with some modikum of CD capability will chew up your landing TFs in no time - even CAs or BBs in a landing TF are of limited use - they just chime in to surpress shore based guns after those have had a go at your transports - and the small fry (below CL) gets chewed up pretty badly.

As a recommendation: Use some of your time on cohabitation with the manual to find out about transports which may unload troops at the spot or the line of scremage ...
And places with MINES and CD are extremely unhealthy: A place like Balikpapan is as bad in AE like Singers used to be in WitP.

Cheers

[Mike Solli]

vonSchnitter, sure it may not be working as designed but it'll work.  For example, Shikoku needs to transfer 303k resources to Honshu each month.  I've figured out how to do it with a total of 20 xAKs.  There are 2 TFs each of 10 ships (10 Miyatis and 10 Adens) working out of two ports to two ports in Honshu.  I'm assuming only 80% efficiency (24 days a month loading, unloading or moving between ports).  They can haul about 402k resources a month.  That's 33% more than they need to haul.  So, if it works as intended, they'll complete their requirement in 18 days.  The 4 ports being used are not used for any other purpose.  I'm using only xAKs starting in the Home Islands and only 10-12 kt varieties.  If I can haul what's needed in less than a month, I can use the remainder to repair the ships.  What I probably will do in that case is to pull 1-2 ships out of the convoy at any one time to repair while the others work.
 
Invasions.  Ah yes, that's going to hurt the merchant fleet.  There's no way around it unfortunately.  That's what happened historically.  The way I am going to counter that is to use selective acceleration.  I plan on accelerating ships (all ship types, not only cargos) when they are 10-30x duration in days from being received.  That's the most cost efficient (in shipbuilding points) way to accelerate.  The easy way to see that is to try to accelerate a ship in the queue.  If you can only choose between accelerate and normal, that's when to accelerate.  If you can choose between accelerate, halt and normal, do not accelerate.

[Mike Solli]

Here's a shot of my newest spreadsheet. I'm not done with it yet (I'll finish it tonight). Lower left shows how I stand vs. required. Upper left shows all the classes with the important details (ship tonnage, cargo capacity, speed, fuel capacity and total number in the Home Islands). There are some hidden columns in there that break down how many are in each region. Then you can see Shikoku and the 2 convoys showing the start and end ports. When you add some ships, it'll calculate the TF tonnage and cargo capacity. Below that I show the max tonnage for the lowest size port. Start and end port sizes. Max load/unload per ship per phase and max daily load/unload per day. #/month is the number of round trips the TF can make in 24 days. Finally it shows how many tons per month will be transported to Honshu.

[Historiker]

I plan on accelerating ships (all ship types, not only cargos) when they are 10-30x duration in days from being received. That's the most cost efficient (in shipbuilding points) way to accelerate.

Will it be cheaper, then?

[Historiker]

bump!

[Mike Solli]

ORIGINAL: Historiker

I plan on accelerating ships (all ship types, not only cargos) when they are 10-30x duration in days from being received. That's the most cost efficient (in shipbuilding points) way to accelerate.

Will it be cheaper, then?

Yeah, it's the most cost effective way to accelerate. Here's the rule from the manual:

"13.4.1 JAPANESE SHIP PRODUCTION
All ships remove 1 day of delay when the delay is greater than:
»» 10 * Ship Durability
This automatic delay removal does not cost Naval or Merchant shipyard points. Those ships set
for normal construction with a delay less than 10 * Ship Durability require Naval or Merchant
shipyard points equal to their durability to remove 1 delay (each day). If set to accelerated
production, the ship will remove 2 days of delay (each day) for a cost equal to 3 times its
durability. A ship that has a delay over 10 * Ship Durability and less than 30 * Ship Durability
may be accelerated. It will remove 1 additional delay (other than the free 1 delay removal) each
day for a cost of appropriate shipyard points equal to the ship’s durability."

Here's the way I interpret the above:

>30x durability (days) = 0 cost for normal reduction of one day per day; no acceleration possible

10 to 30x durability (days) = 0 cost for normal reduction of one day per day; accelerate 1 additional day per day (for a total of two) at a cost of 1x ship's durability in shipbuilding points for each day accelerated

<10x durability (days) = 1x durability for normal reduction of one day per day; accelerate 1 additional day per day (for a total of two) at an additional cost of 2x durability in shipbuilding points for each day accelerated

Here's a chart:

[Historiker]

I see, thank you!

[vaned74]

I don't think this is entirely accurate on ship acceleration. The net yard cost is the same as best I can tell. The difference is on burden on daily shipyard points and whether you have the capacity to accelerate within your production.

An explanation:

Let's say you have a durability 10 ship that is 200 days until delivery.

Case 1 - No acceleration
Time to deliver - 200 days
Cost - 100 days at 10 durability = 1,000 yard points

Case 2 - Acceleration prior to "cost period" (ie delivery < 10x durability days)
Accelerate for 50 turns (means 100 days advance)
Time to deliver - 150 days
Cost - 50 turns of pre-cost accel at 1x durability = 500 yard points +
100 turns of normal cost = 1,000 yard points
Total Cost = 1,500 yard points

Case 3 - Acceleration in cost period100 days of no acceleration and no cost = 0 cost
50 days running at 3x durability cost
Cost = 50 x 3 x 10 = 1,500 points
Time to deliver - 150 days

So the total cost is the same whether you advance in the <10x durability period or in the >10x<30x durability period.

The difference lies in that if you wait to accelerate when your ships are all in the less than 10x durability days period (ie the normal cost days) then you put a lot of extra daily demand on your shipyard point production and if you have a lot of ships accelerated in this period and not a lot of surplus yard points stockpiled, you may not be able to afford it.

If you do no acceleration, the ship will

[Mike Solli]

I stated at the beginning that it was my interpretation.&nbsp; Here's the part that is vague (in my mind, at least).
&nbsp;
"A ship that has a delay over 10 * Ship Durability and less than 30 * Ship Durability
may be accelerated. It will remove 1 additional delay (other than the free 1 delay removal) each
day for a cost of appropriate shipyard points equal to the ship’s durability."
&nbsp;
The way I read it is that the cost is 1x the ship's durability to accelerate it during this period and no additional cost for the normal 1 day per day.

[vaned74]

Yes, I agree completely with your costing interpretation. I am just saying that I think (and emphasis on "I think") is that the total costs in yard points is the same regardless of whether you accelerate in the less than 10x durability days or in the pre-normal cost days. By normal cost I mean those days < 10 x durability where the only the ship advances is if you have yard points available.

Where things even out is that in the period 10x to 30x durability days - there is no cost for the ship to advance at the rate of 1 day per turn.

But, in the period less than 10x durability days it costs yard points equal to the durability of the ship to advance one day (there are no automatic, non costing advancements). Thus, the additional cost you have in accelerating (2x durability per day accelerated) is partly offset by the fact that there are less "normal costing days".

I could be wrong - but this seems to track with my spreadsheets on costing that I have done for this and Witp.

[Woos]

I completely agree with vaned (and tried to point it out to Mike in a previous thread already). IMHO the manual is quite clear:

Further out than 10*durability:
No acceleration:
2 days for 2 days delay
No cost

Acceleration:
1 day for 2 days delay
1 point cost
(so 1 day less for 1 point cost)

Less out than 10*durability:
No acceleration:
2 days for 2 days delay
2 points cost

Acceleration:
1 day for 2 days delay
3 points cost
(so 1 day less for 1 point cost)

So there is no gain in cost effectiveness independent on when you accelerate. Accelerating lots of ships which are more than 10 times out just means that you increase the number of ships which are less than 10 times out until they will use (even unaccelerated) more Nav/Merch Points than available.

Note that all costs are of course times durability.

[Mike Solli]

I see where you guys are coming from.&nbsp; Let me wrap my brain around it this weekend when I have some time.

[Mac Linehan]

ORIGINAL: pompack

Well Mike, I kinda approached it from a different direction. This example is for resources; oil is more complex because there are fewer size choices for tankers.

Assume little ships are less efficient (ton-miles move per dollop of fuel); note that every ship will be sailing every day (unless loading/unloading) but the short haul routes will spend proportionally more time loading/unloading thus that is where you want the small ships.
Assume that I will deliver to a port that is the same size or larger than the source port
Assume that escorts will be the limiting factor so strive to maximize the convoy size
Assume that all of the two smallest AKs are converted to PBs

So I find the closest port that is a source of goodies.
I take one escort (1941, I expect to have to add an additional escort every six months as the subs get better) and as many of the smallest AKs required to be just below the size of the source port.
I determine the transit time; divide capacity by twice transit time plus two days (not always true but this assumes one day load/unload) to give mean daily transfer rate (MDTR); divide daily resource production (or overland movement if port not resource center) by MDTR to determine ROUGHLY how many convoys are required to service the source.
Create the required number of convoys and start them moving. Note that they will bunch up at first but that doesn't matter; in the steady state they will NOT interfere unless you have created too many convoys.
Now repeat for each successively more distant source. Note that as you go further out, the ships in the individual convoys get larger simply because you run out of small ships.

Now I have found that the source ports are not fully utilized; a level 3 port is capable of handling 10k- 20k (big ship versus little ship) and that is more than can be made available in every case I have found (so far). The bottleneck can come at the receiving end with many convoys coming into a limited number of ports at irregular intervals. HOWEVER I have found that there are not really a limited number of ports. The trick is to insure that you direct (for as long as possible) single convoy STREAMS to a single port. When you have a case where the desination port is big enough to simultaneously take a convoy for each of two (or more) streams, then you can direct those multiple streams to a single port.

Now you will still run out of ports that way, so the next step is to monitor the loading of each destination port: a destination port can be considered as a one-convoy-port or a two-convoy-port, etc based upon the size of the convoys and the port. Each day simply note how many additional convoys of the largest size scheduled for that port could be accomidated there and note the most underutilized port on Honshu (this is actually more rigorous than necessary, I just eyeball it myself). When it is necessary to add the next convoy stream, route it to that port. As usual, there will be short-term conflicts but they will work themselves out as the convoys space out to to initial congestion (usually, there will always be recurring conflicts due to the beat frequecies generated by the varying convoy voyage times from different streams using the same destination port.)

The selection of a destination port is not as significant as it seems at first glance. When you use multiple destination ports, some convoys will have to steam further than others but it really doesn't matter that much as long as you avoid crossing routes (e.g. a southbound convoy from Sakahlin heading for Hiroshima meeting a northbond convoy from Luzon heading for Hirosaki would be a Bad Thing). By spreading the load among many destination ports you do tend to increase the sailing distance but you drastically reduce the congestion at the few big ports. Also note that you should be cautious about using a destination port that is adjacent to port on another island; that can reduce the amount of inter-island transfer by "osmosis" and require you to use precious shipping for inter-island transfers.

Now reading this, it sounds much more complicated than it really is. At a basic level it is use convoys that fit the source port and send them to as many destination ports as possible. They will conflict at first but then space themselves out automatically

pompack - was doing a forum search on the Japanese Economy and came across your entry. Could I trouble you give a more detailed explanation of your formula; I am a math wannabe but am not getting the whole picture.

Outstanding work; thank you for your help

His Imperial Japanese Majesties Economics Minister Macsan

[d0mbo]

Indeed an interesting find, this thread.

Did the latest patch make the task of running the convoys easier? I understand it made improvements to the CS command.

So, how do you Jap grunts go about the convoy side of the game nowadays?