• 2021 '20 '19 '18 '17 '16


    This is a technical piece written in an academic style. If you’re new to Axis & Allies, or if you don’t like math, or if you want an exciting post, please read something else! On the other hand, if you love puzzles and abstract analysis, read on!


    I’ve been thinking a lot lately about Japan’s logistics: the puzzle of how to quickly get as many troops as possible off of the Japanese home islands (“Tokyo”) and onto the Eurasian mainland (“China”). The puzzle is interesting because Japan can only build 8 units a turn in Tokyo, but Japan’s economy is usually too big to limit itself to just 8 ground units. For example, a reasonable mix of 8 ground units is 5 infantry, 2 artillery, and 1 tank, which together cost 29 IPCs – but by the end of turn 2, Japan should already be collecting around 36 IPCs. How do you spend the extra seven bucks?

    If you throw the money into tanks, you wind up with a very tank-heavy army, which is of limited use when you’re slogging through stiff resistance on your way to India and the Caucasus.

    If instead you build a plane, you run into a different problem: now you have an odd number of ground units. For example, you could build 4 inf, 2 art, 1 tnk, 1 ftr, for a neat total of 36 IPCs…but then one of your transports is only carrying a single unit, which means that transport is probably giving you a lousy return on your investment. If you try to compensate by building two planes a turn, so that you can fully load three transports, then you run out of money! 4 inf, 1 art, 1 tnk, 2 ftr, for example, would cost you 42 IPCs, which is more than you’re likely to have for the first four turns or so.

    You could try building a factory in, e.g., Manchuria – but, again, there’s a tendency for Japan to run out of money; if you try to build eleven units in a reasonable mix, e.g., 7 inf, 3 art, 2 tnk, then that would cost 45 IPCs, which you’re just not going to have in the opening.

    What fascinates me is that all of these problems crop up even before you start to keep track of what America is doing to harass you in the Pacific, or the exact turn on which you hope to capture India, or anything like that. The problem comes straight from the setup conditions, and I’m going to try to solve it. There will be some math here, but if you stick with me, I’ll give you some concrete recommendations in this post about how many transports to buy as Japan, and in my next post, I’ll try to tackle Japanese factory construction.

    The Economic Discount Rate

    In economics, the ‘discount rate’ means the speed at which money loses its value. For example, if I offer you a choice between $6,000 today and $6,001 next year, you’ll probably pick the $6,000 today – it’s just not worth waiting a whole year to collect one extra dollar. If you had the money today, you could use it to invest in new resources and solve urgent problems, which is worth more than just one measly dollar. Maybe you can use the $6,000 to replace your broken-down van with a more reliable car, and save big bucks on auto repair payments.

    OK, so – would you rather have $6,000 today, or $7,000 next year? How about $9,000? $30,000? At some point, unless you’ve only got a few months to live or something tragic like that, you’ll think it’s worthwhile to wait and take the larger payoff further in the future.

    Let’s say your break-even point is $9,000. You figure that between the interest on your credit card, the money wasted at the mechanic, the hassle of having to take off work to go to the repair shop, etc., it really doesn’t matter which option you choose; you come out about the same if you take $6,000 now or if you take $9,000 next year. That means your “discount rate” is 33%, because money next year is worth 33% less to you than money this year. $9,000 * (1 - 33%) = $6,000.

    I think this is about the situation on the Axis and Allies board: 6 IPCs right now are worth about the same amount as 9 IPCs next turn, or 13 IPCs two turns from now. If you give me 6 IPCs right now, I can use them to buy 2 infantry that will let me hold India for another turn, collecting India’s 3 IPCs. On the other hand, if you give me 6 IPCs next turn, I’ll lose India, and then I’ll only have the 6 IPCs. That’s why you have to pay me 9 IPCs if you wait until next turn – the IPCs aren’t worth as much if I have to wait for them. If I’m going to choose a strategy that delays when my troops get to the front lines, then that strategy better have some other payoff that compensates for the lost income from not being able to hold on to (or rapidly conquer) the territory on the front lines.

    You could make a plausible case that the “real” discount rate for Axis & Allies is anywhere from 15% to 60%. Part of it depends on what kind of strategies people are using. If you’re blitzing for Moscow with an army of all tanks and bombers in a low-luck game, the discount rate will be very high: I need that money right now, before Moscow falls, and every unit I add has a big chance of swinging that battle. On the other hand, if you’re slowly teasing each other with destroyers in a full-luck game, then the discount rate will be very low: as long as I get my money eventually, I can afford to wait a few turns without losing anything more than a couple of sea zones.

    Adjusting for the Cost of Delivery

    If you look closely at the stats and prices of A&A units, you can see that quite a lot of the cost of a high-priced unit goes into paying for its higher mobility. For example, a bomber has 4 Offense, 1 Defense, 6 Movement, 1 Hit Point, and costs 12 IPCs. But for those same 12 IPCs, you could buy two infantry and a tank, which would have 5 Offense, 7 Defense, 1 Movement, and 3 Hit Points. If you’re not concerned about the movement rate, the 2 inf + 1 tnk is a much, much better buy.

    When it comes to evaluating Japanese purchasing plans, I think we’re not much concerned about the movement rate. The hard part is getting units to China via factories or transports. Once they’re in China, they don’t really need to move more than one territory per turn. For example, Yunnan is three moves from Moscow and two moves from India. Unless your opponent’s an idiot, all of those territories will be defended, often staunchly enough that you can’t afford to advance with just your tanks unless you want those tanks smashed to bits. And if you can only move one territory per turn, your tanks aren’t worth the 6 IPCs you paid for them, nor are the bombers worth the 12 IPCs you paid for them. They’re worth much less.

    This might be controversial, but I’d suggest the following values for units that are intended for a slow ground assault, once delivery is accounted for. The values are in an imaginary currency called “siege points,” which represents the value of a unit that’s part of a slow, ground-based campaign of attack, like the ones that Japan can expect to fight in southeast Asia.

    Infantry – 3 siege points
    Artillery – 4 siege points
    Tanks – 5 siege points (even though they cost 6 IPCs)
    Fighters – 6 siege points (even though they cost 10 IPCs)
    Bombers – 7 siege points (even though they cost 12 IPCs).

    To double-check these values, imagine attacking with 2 inf, 2 art vs. attacking with 1 inf, 1 art, 1 bomber. The all-ground forces have 8 points of offense and 4 hit points, and if they win they hold the territory with 8 points of defense. The mixed forces also have 8 points of offense, but only 3 hit points, and if they win they only hold the territory with 4 points of defense (the bomber can’t land in the newly captured territory). If anything, the bomber is less effective than 1 inf, 1 art would be. I’m crediting the bomber with 7 imaginary ‘points’ of siege value in part because the bomber is flexible, and no siege is so static that it’s never worthwhile to have some extra mobility. For example, the bomber helps deter enemy naval reinforcements, and gives you the option of a strategic bombing run. In a straight-up attack, though, the bomber isn’t even as good as two ground units, let alone four ground units.

    Don’t get me wrong: a bomber is probably worth the 12 IPCs you pay for it when it’s sitting in Tokyo. The fact that the bomber can take off from Tokyo and get its fat behind over to the front lines without any help from a transport or a new factory is a huge asset. But once you separate out the cost of delivering the bomber to the front lines, as I’m about to do in the next section, then the bomber’s extra mobility is not very valuable. It’s much better to have a bomber in Tokyo that can move to the front lines than to have three ground units stuck in Tokyo sitting on their thumbs – but it’s usually much better for Japan to have three ground units on the front lines than to have one bomber on the front lines.

    Assumptions and Scoring

    So, with all that theory in mind, here are my assumptions for the thought experiment:
    (1) Japan is ferrying all available troops to Yunnan in southeast China.
    (2) Cargo has to arrive in Yunnan by turn 6 at the latest, or it is worth no points.
    (3) Japan earns 30 IPCs on turn 1 and 38 IPCs on turns 2, 3, 4, 5, and 6.
    (4) Japan builds no factories.
    (5) Japan never has to devote any resources to defending its transports (other than its starting navy and air force).
    (6) Japan has a discount rate of 33%, meaning that a soldier delivered on turn 2 is worth 1.5 times as many points as a soldier delivered on turn 3.
    (7) Japan is willing to evacuate any/all of its island garrisons in order to support the blitz through Yunnan to India and Moscow.
    (8 ) Japan will try to fill any transports it has as best it can, and then spend any remaining money on planes.
    (9) The UK sinks one of Japan’s two starting transports on the UK’s first turn. Other than that, the UK navy / air force in the Pacific is immediately sunk or evacuated and poses no further problems until after turn 6.

    After six turns, Japan gets a ‘siege score’ based on how many siege points’ worth of soldiers it delivered to Yunnan, and how quickly the soldiers got there. Turn 1 is not scored, because the delivery is totally based on the initial setup. Soldiers delivered on turn 2 are worth full credit in siege points, and then soldiers delivered on turns 3, 4, 5, or 6 are discounted by 33% per turn, compounded each turn. The strategy that gets the highest adjusted total of siege points is considered to have ‘won’ the thought experiment.

    How Many Transports Should You Build as Japan?

    So: given these assumptions, how many transports should you build if you want to optimize your discounted siege score? Let’s take a look at four of the major options.

    Build Zero Transports (Stick with only one Transport)

    With only one transport, you can ship 2 ground units each turn, and the rest of your cash is going to have to go into planes. Importantly, you start with six ground units in Tokyo from the initial setup (4 inf, 1 art, 1 tnk) and two ground units (1 inf, 1 art) in Manila that are very easy to deliver to Tokyo (you can pick them up on your way without losing a turn). That means that for the first four turns, you can fill your transport with pre-existing ground units, so you don’t have to build any ground units on turns 1, 2, or 3. Starting on turn 4, you need to build two ground units each turn so you can keep your transport full.

    That means on turn 1 you can build 3 fighters, on turns 2 and 3 you can build 3 bombers, and on turns 4 and beyond you can build 1 inf, 1 art, and 3 fighters. On turn 2 (the first turn that we’re going to score) you deliver (1 inf * 3 siege points) + (1 art * 4 siege points) + (3 ftr * 6 siege points) = 25 siege points. On turn 3, you deliver (1 inf * 3 siege) + (1 art * 4 siege) + (3 bomber * 7 siege points) = 28 siege points, but we have to discount that by 33%, for an adjusted total of 19 siege points. On turn 4, you’ve run out of artillery, so you deliver (2 inf * 3 siege) + (3 bomber * 7 siege) = 27 siege, and you have to discount it by 33% twice, yielding 27 * (2/3) * (2/3) = 12 adjusted siege points. On turn 5, you have a fresh supply of artillery from your first wave of ground unit construction, so you deliver (1 inf * 3 siege) + (1 art * 4 siege) + (3 fighter * 6 siege) = 25 siege. Multiply by 2/3 three times to adjust for the discount rate, and you get 7 siege points. The same delivery on turn 6 is worth 5 siege points because you have to multiply by 2/3 four times. Let’s stop there – we’re going to be dealing with tinier and tinier numbers as we move forward in time, and the troops you deliver to Yunnan after turn 6 might not reach Moscow before the game is decided, and at this point America is probably breathing down your neck anyway.

    Your air-heavy strategy, relying on only your starting transport, yields 25 + 19 + 12 + 7 + 5 = 68 discounted siege points. Can we do better?

    Build Three Transports, for a total of Four Transports

    Here’s the opposite strategy: suppose we build three transports on turn 1, hoping to rely exclusively on ground troops for our attack. There are 8 ground units available from setup, 2 of which will be unloaded on turn 1, so we have to replace those units with the turn 1 build in order to fill all 4 transports on turn 2. That means the turn 1 build is something like 3 transports, 1 inf, 1 tank. Starting on turn 2, we’ll need to build 8 ground units to make sure all 4 transports are full. Since we have plenty of money, that means a build of 4 inf, 4 tnk each turn.

    On turn 2, the first turn that gets scored, we deliver 4 inf, 2 art, 2 tanks for a total of (43) + (24) + (25) = 30 siege points. On Turn 3, we deliver 4 inf, 4 tnk for a total of (43) + (4*5) = 32 siege points, adjusted down by 33% for an adjusted total of 21 siege points. On Turn 4, we deliver the same 4 inf, 4 tnk for 21 * 2/3 = 14 siege points. Turn 5 is worth 9 siege points, and Turn 6 is worth 6 siege points. This strategy yielded 30 + 21 + 14 + 9 + 6 = 80 discounted siege points … much better than the 68 discounted siege points from the all-air strategy! It looks like Japanese transports built on turn 1, assuming they survive through Japan’s turn 6 (reasonable in a KGF game), will more than pay for the cost of their investment.

    Build Four Transports, for a total of Five Transports

    There’s one more option we need to discuss: a fifth transport! What can Japan do with a fifth transport? For starters, Japan can deliver its initial setup pieces more quickly – even though Japan can’t sustain a delivery of 10 ground units per turn from Tokyo (because it can only build 8 units per turn in Tokyo), it will take a long time for the starting units to run out. Let’s see how this plays out:

    Turn 1, build 4 transports.
    Turns 2 through 6, build 4 inf, 4 tnk.

    Turn 2, deliver 4 inf, 2 art, 2 tnk (30 pts). We have a fifth transport available, but it does not have any ground troops available to transport. Send that transport to Iwo Jima and Okinawa to collect the 2 infantry there.

    Turn 3, deliver 6 inf, 4 tnk (38 * 2/3 = 25 pts) by dropping off the Iwo Jima and Okinawa infantry.

    Turn 4, deliver 4 inf, 4 tnk (32 * 4/9 = 14 pts), and send the extra transport to East Indies for 2 more infantry.

    Turn 5, deliver 6 inf, 4 tnk (38 * 8/27 = 11 pts) by delivering the infantry from the East Indies

    Turn 6, deliver 4 inf, 4 tnk (32 * 16/81 = 6 pts) because the fifth transport can go fetch infantry from New Guinea, but they won’t get back to China until turn 7.

    Total score: 30 + 25 + 14 + 11 + 6 = 86 points…a noticeable improvement over the 80-point four-transport strategy!

    Build Five Transports, for a total of Six Transports

    Now we’re getting deep into crazytown – can Japan make efficient use of a sixth transport?

    Turn 1, build 4 transports, save 2 IPC
    Turn 2, build 1 transport, 4 inf, 2 art, 2 tnk
    Turns 3 through 6, build 4 inf, 4 tnk

    Turn 2, deliver 4 inf, 2 art, 2 tnk (30 pts). We have a fifth transport available, but it does not have any ground troops available to transport. Send that transport to Iwo Jima and Okinawa to collect the 2 infantry there.

    Turn 3, deliver 6 inf, 2 art, 2 tnk (36 * 2/3 = 24 pts) by dropping off the Iwo Jima and Okinawa infantry. Send the sixth transport to the Caroline Islands to pick up one infantry there.

    Turn 4, deliver 5 inf, 4 tnk (35 * 4/9 = 16 pts) by dropping off the Caroline infantry. Send the fifth transport to the East Indies to pick up two infantry there.

    Turn 5, deliver 6 inf, 4 tnk (38 * 8/27 = 11 pts) by delivering the infantry from the East Indies. Send the sixth transport to Borneo to pick up one infantry there.

    Turn 6, deliver 5 inf, 4 tnk (35 * 16/81 = 7 pts) by delivering the infantry from Borneo. The fifth transport can be sent to New Guinea to fetch the one infantry there, but it won’t make it back to China until at least turn 7, so it doesn’t count.

    Total score: 30 + 24 + 16 + 11 + 7 = 88 points. Amazing! The sixth transport manages to pay for itself and deliver a small profit. The score on turn 3 (24 pts) is slightly weaker because building the sixth transport cuts into the funds we need to build ground units, so we have to swap out a pair of tanks for a pair of artillery. But that’s OK, because we deliver an extra infantry on turn 4 and on turn 6, which together are more important than the downgrade on turn 3.

    Notice that a seventh transport would not pay for itself, because instead of just downgrading units, we’d probably lose a unit altogether. Worse, the seventh transport wouldn’t be able to deliver two new units until turn 6 (when the units are worth considerably less) because it has to ferry them all the way from New Guinea and the Solomons.


    So – if you buy the initial assumptions that we’re ferrying cargo strictly to Yunnan beach, and that it has to arrive by turn 6 at the latest, and that Japanese income is constant on turns 2 through 6, and that we’re not building any factories, and that Japan never has to devote any resources to defending its transports (other than its starting navy and air force), and that the discount rate is 33%, then the optimal number of transports for Japan to build is 5 transports, so it can wind up with a fleet of 6 transports altogether.

    Obviously, in a real game, these assumptions will have to be varied somewhat – but I think in a Kill Germany First (“KGF”) game, these assumptions are realistic enough that the underlying lesson still carries some value. Many players work with a fleet of 4 transports or even just 3 transports for Japan, and in a KGF game, that just doesn’t seem like enough to me. It’s better to build a couple ‘extra’ transports to help get your troops off the islands than to insist on a ‘perfect’ 4 inf + 4 tnk build or to try to rely on air power to make up the transport gap.

    In my next post, I’ll discuss how things change when we add factories as an option. If the series attracts enough interest, future posts will also discuss leaving a budget to fend off American attacks in the Pacific, higher and lower discount rates, and a more detailed model of how Japanese income changes over time.

  • This is a pretty neat write-up. I’d love to see more. I always get bogged down with not being able to get my units over to Asia to take India, even though my economy is big enough to shred India, China, and Russia.

    I don’t think that I will build an extra 5 transports because if America invests even a little into navy, then that extra transport can’t snarf up the island infantry without an escort.

  • '19 '18 '17 '16 '15

    Without finishing all the details i can immediately tell this is a fantastic post!  Great job!

  • Hello, I was reading this strategy and when I got to the section about building 4 starting transports to end up with 5, the math isn’t working out for me on the delivery of turn 2.  Maybe there is something I’m missing?

    Turn 1, build 4 transports.
    Turns 2 through 6, build 4 inf, 4 tnk.

    Turn 2, deliver 4 inf, 2 art, 2 tnk (30 pts)

    Where do the 1 art and 1 tank come from?  Japan starts with 8 ground units to unload, you unload 2 on the first turn, leaving you with 6 and since you don’t buy a tank/inf on turn 1, were do those extra two units come from?

  • 2021 '20 '19 '18 '17 '16

    Thanks, guys! I’m glad people are enjoying the topic. Jagula correctly notes that I have been getting the math wrong on the Japanese starting units. Thank you, Jagula! To fix this problem, I have moved all of my math onto a spreadsheet (attached) and started ‘scoring’ on Turn 1 instead of on Turn 2. Also, for added realism, I am now modeling the Japanese economy as collecting 30 IPCs on Turn 1, 32 IPCs on Turn 2, then 34, 36, 38, and then 40 IPCs on Turn 6, and I am scoring the 6th turn of production at a discount rate of 9% instead of ignoring it entirely. Other than that, all the assumptions from the previous post (KGF game, British kill one Japanese starting transport and then retreat, etc.) all still apply.

    Anyway, onto Part 2: Factories!

    How do Factory Locations Affect the Speed of Your Attack?

    The most important benefit to building a Japanese industrial complex (“factory”) is that units built in a factory sitting on the Asian mainland are at least one turn closer to attacking their final targets. If you build an infantry in Tokyo on turn 2, then the earliest you can move it to Yunnan is turn 3, and it won’t reach Burma until turn 4 and India until turn 5. On the other hand, if you build an infantry in a Kiangsu factory on turn 2, then you can move it to Burma on turn 3 and India on turn 4.

    We can represent this extra speed mathematically by applying one less “discount rate” to the siege point value of units build in mainland factories. For example, an infantry built on turn 2 in Tokyo will be delivered to the mainland on turn 3, so it’s worth (3) * (1 - 33%) = 2 siege points. An infantry built on turn 2 on the mainland is “delivered” to the mainland on turn 2 itself, so it’s worth a full (3) * (1 - 0%) = 3 siege points. Of course, the one-turn speed advantage is just an average. Your exact speed advantage will depend on where you build your factory.

    Where should Japan build Factories?

    There are three plausible locations for Japan to build factories in the opening: Manchuria, Kiangsu, and French Indochina Thailand (or “Vietnam” for short). Building a factory on the 4-IPC islands of East Indies or Borneo seems attractive because of their high IPC values, but it’s almost always a trap: if you build a factory on an island, you’re just going to have to build more transports to ship the units you built there off the island. The point is to get units into mainland Asia, and building factories on islands does not accomplish that goal.

    Manchuria has the obvious advantage that it lets you build three units per turn. Unfortunately, it’s positioned way too far north to be useful for an assault on India: five territories away. It’s also four territories away from the Caucasus and Moscow. The only thing Manchuria is really convenient to is Siberia, which isn’t worth much: all the 1-IPC territories make lean eating for a hungry Japanese Empire. You might want one tank a turn in Siberia, but after that the units really need to travel south toward the juicier territories. To reflect this, we’ll score two out of the three units built in Manchuria as if they were built in Tokyo, i.e., with no special speed advantage.

    Kiangsu, by contrast, is ideally located for all kinds of offensive campaigns: it’s three moves from Moscow, three moves from the Caucasus, and three moves from India. Very nicely balanced, and as an added bonus, once America starts attacking you, you can use a Kiangsu factory to build destroyers in the same sea zone that you use to drop off Tokyo’s infantry in Yunnan – Yunnan and Kiangsu share the same sea zone.

    Vietnam is another good choice for a factory site if you plan to attack Britain – it’s two moves from India, four moves from Caucasus, and four moves from Moscow. (In the very early game, two moves from India might be too close for comfort. Don’t make the embarassing mistake of handing Britain a gift-wrapped factory! Even if you recover the factory, you’ll never recover the lost momentum.)

    Other Benefits for Japanese Factories?

    Other than speed, there are a few other potential benefits for mainland Japanese factories. One is that they increase the total number of units you can build each turn – if you never build or capture a factory, you’re limited to building eight units per turn. Once you run out of ‘free infantry’ to pull from island garrisons, that cap will mean you can’t deliver more than 8 units per turn to the mainland.

    Another benefit is that you can (theoretically) build a unit mix that has more than 50% artillery and tanks. This is rarely a good idea, because infantry are so efficient and are needed as cannon fodder to absorb your inevitable casualties. No matter how dominant your army is, you’re going to take a few hits along the way, and you may as well take those hits on cheap infantry units. Still, as you’re gearing up for the final push on India or Moscow, sometimes it’s nice to be able to crank out tanks close to the scene of the action.

    Finally, some people think factories offer an implicit defensive benefit, because even if America permanently sinks your fleet (preventing you from using transports as ferries), you can still build troops on the mainland with your factories. In my opinion, this is a pretty thin benefit: yes, an Asian factory can be useful if America attacks you with a fleet of pure submarines or some other extreme tactic, but if America is serious about attacking Japan, then they will usually send along a few loaded transports of their own. The Americans’ initial goal might be to seize the Philippines for a victory city, or Borneo for cash – but if you have a juicy factory sitting in Manchuria, there’s no reason why the Americans can’t take that territory first, and then move on to their original target(s). In any scenario where you fleet is being sunk, you’ll get at most one or two extra rounds of Asian production out of a mainland factory, and then the factory will be a gift for the Americans.

    How Many Factories Should Japan Build in a KGF Game? When should they be built?

    Sticking with the same assumptions as last time (no significant American threat in the Pacific, 33% discount rate, primary Japanese goal is to conquer India, Caucasus, and Moscow as soon as possible via a land invasion), how many factories should Japan build? None, one, two, more? And should you built them on turn 1, turn 2, or turn 3? There are lots of different combinations here. I don’t have time to test them all, but I’ve tried out a dozen of the most obvious strategies on my spreadsheet, which is attached.

    If you want to check my work, or if you want to try out strategies of your own, private message me and I’ll be happy to share my spreadsheet with you (it’s a Google Doc).

    Otherwise, my bottom line is that you want to build either three transports with zero factories (61 siege points), four transports with zero factories (62 siege points), two factories followed by one transport (61 siege points), or two transports and a factory followed by a second factory (63 siege points). Any of these strategies are much better than zero transports with zero factories (53 siege points).

    Other than making sure to build a reasonable number of transports and factories, the most important factor in getting a high siege score is to make sure any transports and factories you do build get fully utilized each turn. If you build a factory, make sure you can max out its production each turn. Transports and factories will both pay for themselves and yield a small but important profit in just about any quantity or combination you care to build them, but only if you fully utilize them – if you leave them empty, even for one turn, then you probably would have been better off sticking with a smaller supply chain and supplementing your buildswith air power.

    What about defense?

    In the third and last post of this series, I’ll take a look at how the Japanese supply chain changes when you have to worry about defending against a mid-game American naval invasion. How many transports should you build when you have to save a slot for subs and destroyers? How many factories can you build if you’re going to be focusing on holding east Asia defensively rather than on seizing Moscow? Stay tuned.

    Japanese Logistics Tables.pdf

  • 2021 '20 '19 '18 '17 '16


    Because this series is about Japanese logistics, rather than just Japanese strategy in general, I’m not going to cover how to go toe-to-toe with the US Pacific fleet and win. Instead, I’m going to cover how to get as many troops as possible to the Asian mainland as fast as possible, despite significant interference from the US navy.

    As I hope to show you in this post, even if your only goal is to maximize the troops you deliver to the Asian mainland, you still want to invest in a significant naval defense budget, because defending your sea zone will buy you additional turns in which to keep purchasing and delivering additional cargo.

    Of course, a reasonable player will have other goals besides just delivering troops to Asia, like keeping the Japanese capital out of American hands. Remember, every turn Tokyo stays alive isn’t just another turn of Japanese production – it’s also another turn during which Japanese troops that have already reached the mainland can continue their march on Moscow.

    Note that in a KFJ opening, you usually don’t even want Japan to take Moscow if you can help it, because the income from a German-owned Moscow will be safer as the US closes in on Tokyo. That means your focus should be less about a death march of tanks, and more about eating away at Russian income, sending fighters to help support forward-deployed German troop stacks, and light trading in the territories bordering Moscow to force Russia to deplete its infantry reserves.


    Here are the five biggest threats that Japan wants to defend against, in descending order of priority:

    1. Capital kill: If they take Tokyo, that’s very bad. Make sure they don’t take Tokyo. Unless Germany has taken all of Africa and most of Siberia, trading Tokyo for Moscow means that the Allies win. One way to help protect Tokyo is to make a significant infantry purchase each turn in Tokyo. If you have two transports shipping 3 inf, 1 art to the mainland each turn, that’s 4 extra land units defending Tokyo. In fact, consider overbuilding your land units on purpose for a couple of turns once you see the first loaded American transports venture west of Hawaii. In other words, if you’ve got 2 transports, build 6 land units. The extra land units won’t make it to Asia next turn, but they will make sure Tokyo survives, and they’ll ‘stick around’ for future turns so that Tokyo is always that much more protected from amphibious invasion. If America does invade and knocks out a couple of your troops, the surplus means you’ll still be able to make your regularly scheduled Asian delivery.

    2. Factory steal: If you build a factory on the Asian mainland, and the Allies come in with a big fleet of transports, they can capture the factory and start using it to crank out new units next turn. This is a big problem because your mainland factories (especially the 2-IPC ones) usually can’t crank out units quickly enough to defend against a large transport attack – by the time you see the problem coming, it’s too late. Coping mechanisms include keeping a stack of fighters around to reinforce factories on short notice, including some tanks in your mix of Asian production so that you can retreat the tanks if necessary without totally shutting down your westward progress, the Kwangtung shuck-shuck (discussed below), and not building mainland factories.

    3. Transport sink: If you’re relying on a big stack of transports to get troops across to the mainland, and the Allies take the sea zone where your transports live (or even just occupy the sea zone where your transports need to unload), then your pipeline of new troops is finished. On the other hand, if this doesn’t happen until the late middlegame, it might not be a problem – you’ll need those troops for defending Tokyo, anyway, and the troops that are already on their way on the mainland might be enough to weaken Russia for Germany’s killing blow. Particularly if the Allies sent a fleet that’s heavy on warships and light on transports, losing your transport stack might be a blessing in disguise. This is one reason why transports are more powerful than than factories in a KJF opening.

    4. Victory City grab: While you’re busy on the mainland, the Allies can often keep or retake some combination of Honolulu, Manila, Calcutta, and maybe even Shanghai. That’s usually not a huge problem – if you wind up with both Moscow and Tokyo, then the Axis should have time to get super-rich and push the Allies back off of those victory cities. Still, keep an eye out for Allied surprise attacks on Paris, Rome, and Leningrad. Because victory city wins are scored at the end of the US’s turn, there’s a chance that the US could push the Allies past the critical VC threshold for a win all at once, without giving you a chance to respond.

    5. Money Island grab: Many KJF strategies advise the US to recapture the Philippines, Borneo, and the East Indies (the so-called “money islands”, worth a total of 11 IPCs/turn). While that is indeed big money, the islands aren’t great sites for Allied factories because they require Allied transports to get any troops off those islands. An Allied stack in Borneo is way out of position to defend Moscow, and it’s not even in a great position to attack Tokyo – the Allies may wind up splitting their fleet between Hawaii and Borneo, allowing you to attack one of the fleets and win or at least trade. You should be able to make up economic losses from the money islands by conquering valuable mainland territories like India, Caucasus, Egypt, Kazakh, and Novosibirsk (worth a total of 13 IPCs/turn). If the US moves in force to take the money islands, let 'em go. This is not an existential threat.


    On average, it takes about four turns for the American navy to show up in Tokyo Bay if you don’t spend any money on defense. On A1, America builds a fleet of defensive warships, like a pair of loaded carriers and a destroyer, starts shifting troops to the Western US, and brings the Atlantic navy (if it survived) through the Panama Canal. On A2, America finishes stacking in the Western US and builds transports and infantry so that it has about 5 loaded transports off the coast of San Francisco. On A3, the transports and carriers move to a central island like Wake or Midway, and on A4, the transports and carriers can reach Tokyo, Manchuria, or Shanghai.

    A Turn 3 Blitz?

    A committed blitz that’s supported by the British can reach Tokyo in only 3 turns: if the British, e.g., use their bid to put a sub off the coast of India, then the combined Indian and Australian fleets can probably trade at least evenly against the Caroline Islands carrier and battleship in SZ37. Meanwhile, if America builds nothing in the Atlantic and diverts all fighters to the Pacific, it can probably afford to build both a small defensive fleet and a couple extra transports on A1, picking up troops from Hawaii and Midway as the fleet cruises west on A2, and landing in Tokyo or Manchuria on A3. The downside to this blitz is that America is only bringing 3 loaded transports, which may not be effective against hardened targets (e.g., Tokyo), and America and Britain are totally neglecting the European theater. Germany could be so dominant in Africa that it actually breaks out through the Suez and supports the Japanese efforts against India. Germany may also be able to stack and hold Karelia and the Caucasus very early. While a turn-3 blitz might work against a Japan that was totally refusing to spend money on defense, in practice Japan could drop, e.g., 5 infantry and 3 fighters into Tokyo on turn 3, blocking the capital attack and setting up a turn 4 kill of the US Pacific fleet if the US reinforces the fleet with additional transports instead of additional warships.

    The Turn-4 Baseline

    So, while it’s not impossible for the US to reach Tokyo in 3 turns, it probably makes more sense to use 4 turns as the baseline for when the US can arrive in Tokyo without Japanese defensive spending. Any money that the Japanese spends on defense has to get justified in terms of its ability to hold off the American navy for at least one more full turn so that Japan has an extra turn in which to deliver troops to the Asian mainland.

    In other words, the siege points ‘lost’ by Japan when it builds subs or carriers for defense on turns 1-4 need to be made up by siege points ‘gained’ by Japan on turns 5+ when Japan continues to deliver troops to Asia that could not have been delivered if the American navy were present on the scene.


    Again, this post isn’t going to analyze which kinds of fleets are exactly optimal for defense, but as a rule of thumb, the two main options for a defensive fleet are (a) carriers, fighters, and destroyers, or (b) subs and bombers. Whichever option you go with, the key is to exploit the extra turns that the US needs to move its fleet across the Pacific (as well as the US’s need to build loaded transports) to deliver a big defensive punch on a small budget, so that most of your budget can go directly into assaulting mainland Asia.

    Recall that if you spend zero IPCs on defense, the US can safely destroy your fleet on turn 4. So, the goal when spending IPCs on defense is to acquire lots of “bonus” turns (relative to the turn-4 baseline) during which the US doesn’t dare to attack you for a relatively small investment in defense, so that the extra siege you deliver during the bonus turns is worth more than the siege you sacrifice by diverting funds to defense.

    Carriers, Fighters, and Destroyers

    If you go with option A, you want a mix of fighters for their high defensive value supported by destroyers for relatively cheap fodder. This can make it very hard for your opponent to attack you cost effectively. Let’s say you spend 10 IPCs/turn on Pacific defense vs. America’s 30 IPCs/turn on Pacific offense. Because your destroyers and carriers are sitting in the sea zone you’re trying to defend, and because Japan goes before America in the turn order, and because Tokyo is two turns away from San Francisco, you’ll get two extra builds compared to the builds in the fleet America attacks you with. In other words, if America launches its offensive fleet on turn 4, then that fleet only has three turns worth of builds in it, but the fleet won’t arrive until after Japan builds ships on turn 5. America’s 30 IPCs / turn add up to 90 IPCs of offense, while Japan’s 10 IPCs / turn add up to 50 IPCs of defense.

    You can spend your 50 IPCs purely on defending the naval zone – let’s say 2 destroyers, 1 carrier, and 2 fighters.

    America has to split its 90 IPCs of offense between warships and transports and marines.

    Four loaded transports (the minimum for a serious attack on Japanese core territories) cost 28 + 12 + 16 = 56 IPCs, leaving America with only 34 IPCs for offense. Let’s say America spends that on 1 sub, 1 DD, and 1 BB. That battle wouldn’t even be close against your defending carrier group –  the battle calculator gives 86% odds to Japan. Even though America is outspending Japan by a factor of 3:1 in the water, America would lose any attack on Tokyo Bay on turn 4 OR on turn 5 because of Japan’s defensive fleet. That means that America needs at least two more full turns of buildup (relative to a zero-Japanese-sea-defense baseline) to be able to invade core Japanese territories. (Even on turn 6, the battle will be close to 50-50 for Japan, or even better if you up your turn 6 defensive spending to 20-25 IPCs).

    This is exactly the kind of trade-off that Japan is looking for – because with the 50 IPCs you spend on defense, you gain two turns, and with that additional time, you can drop another 70 IPCs or so into mainland Asia, as well as just having additional time to march units in Asia toward Moscow. Your total assault on Asia is stronger over the course of 6 turns if you spend 10 IPCs/turn on defense than if you spend 0 IPCs on defense and then drop nothing in Asia on turns 5-6 because your fleet is dead.

    Note that to maximize the effectiveness of your siege, you can drop most of your warships in the water after America moves to Hawaii. Let’s stick with the 2DD + 1 CV + 2 Ftr example. Build the fighters first, and send them to Asia – they can fight in Asia and still make it back to Tokyo Bay by turn 5. Build the DDs on turns 3 and 4 – they’re cheap, and won’t distract much from your siege. Finally, build the carrier on turn 5. That way you get a consistent 7 build slots left over for land units, and your heaviest investment in defense comes at the end of the opening, when additional land troops are less useful for Japan (because of the discount rate) and when the American invasion is already (erroneously) underway.

    Subs, Bombers, and Deadzones

    The other defensive option is subs and bombers. Again, you want to build the bombers first, because they can be used to fight in Asia and still make it back to Tokyo Bay in time to help punish an American invasion. The subs can be built later, but unlike destroyers, they need to be built one turn before the US moves to Hawaii, because you need time to attack with the subs and bombers, rather than just letting the US come to you. With subs and bombers, you are not literally defending Tokyo Bay. Instead, you’re turning Tokyo Bay into a dead zone – you want to make it so that if America moves its fleet to Tokyo Bay, you can kill the fleet at a profit. (In context, killing “at a profit” might mean that you kill it while losing only 10-20 IPCs more than the Americans lose, because in the time it takes for the Americans to build a second fleet, you might have already captured Moscow).

    Let’s say America again is spending 30 IPCs/turn on Pacific offense, so they come with 1 sub, 1 DD, 1 BB, and 4 loaded transports, moving to Hawaii (or Midway or Wake; it doesn’t matter) on turn 4. You spent 10 IPC/turn for 4 turns and built 2 bombers and 3 subs. On turn 5, you attack the American fleet in Hawaii, and you’ve got 80% odds to win that battle, sinking his boats and transports. You might lose the subs to a counter-attack from San Francisco, but your bombers can fly safely back to Japan.

    So again, your 10 IPC/turn in defense buys you at least two extra turns to wreak havoc in Asia, and since a lot of your ‘defense’ consists of a pair of bombers which are perfectly useful in Asia, it’s a much better option than zero defense.

    The Optimal Japanese Naval Defense Budget

    As you can see from the attached spreadsheet, I calculate that the optimal Japanese defense budget is about 15 IPCs per round. I’m making a number of assumptions here – the US is spending $30 per round every round in the Pacific, leaving them with enough cash to help support Africa or Norway, Japan is doing reasonably well in mainland Asia and so its income is going up 2 IPCs per turn each turn (e.g. 30 on turn 1, 32 on turn 2, 34 on turn 3, 36 on turn 4…), the US will build 4 fully loaded transports, the US won’t attack in Tokyo Bay unless the US can get at least a 5-to-4 ratio of total warship unit value in the Pacific, and once the US sinks Japan’s transports, all further deliveries to Asia are permanently cut off.

    These assumptions are necessarily simplified – in real life you could, e.g., send some planes to Asia even after you lost your transports, and America might attack with a 4.5-to-4 ratio of total warship unit value if they have plenty of subs and a good mental odds calculator – but they let me build a model that helps illustrate the central idea of the post: as Japan increases its defense budget up from $0/turn, Japan gains total value because it gains more turns in which to build and deliver cargo, but eventually further increases in defense spending become counter-productive and Japan loses total value, because spending that much on defense takes too much away from delivering troops to Asia, and so even if you can maintain the tiny deliveries for many more turns, the trade-off still isn’t worth it.

    The main reason why that trade-off happens is that a big part of Japan’s defensive advantage comes from the 2-turn tempo advantage based on America’s need to (a) cross the Pacific, and (b) attack at a point in the round after Japan has already built new boats. If Japan spends nothing on defense, then the tempo advantage is lost, because (x+2) times nothing is still nothing. If Japan spends buckets and buckets on defense, the tempo advantage is eroded, because (10,000 + 2) isn’t that much more than (10,000). Japan wants to spend just enough cash on warships to exploit the tempo advantage for a few turns while still devoting the bulk of Japanese purchases to transports, tanks, and infantry.

    (This post is continued in the next post due to character count limitations.)

    Japanese Logistics on Defense.pdf

  • 2021 '20 '19 '18 '17 '16


    The Kwangtung Factory on Defense

    At the start of the game, the only sea zone you absolutely have to defend is Tokyo Bay. Once you build a transport fleet and start ferrying troops to Asia, you also have to defend whatever sea zone you’re using to unload your troops, because that’s where your transports end their turn. This is typically the Yunnan / Kwangtung sea zone.

    Most Japanese factories create an additional defensive liability – if you build a factory in Vietnam, now you’ve got to defend three sea zones: Tokyo, Kwangtung, and Vietnam. A Manchurian factory is slightly less of a problem because a strong Japanese fleet in Tokyo mostly screens off the Manchurian sea zone, but you’re still vulnerable to an American approach along the northern route, via Alaska and the Soviet Far East. The best defensive site for a factory is in Kwangtung itself – that way, you’re still only defending two sea zones: Kwangtung (where your transports unload and your factory sits) and Tokyo (where your transports load and your capital sits).

    An added bonus for a Kwangtung factory is that, as Black Elk has pointed out, you can use it to efficiently drop new boats in the water. When you’re playing defense, it’s very important that any new boats you build immediately link up with the rest of your fleet. But if your fleet is assigned to guard your transports, and your transports end each turn off the coast of Yunnan, then you also need to build new boats off the coast of Yunnan – which you can do if you have a factory in Kwangtung.

    All of this assumes that you’re using what TripleA players affectionately refer to as a “shuck-shuck” strategy, named after the sound that you make with your mouth while you pick up a fleet of transports to move it back to your capital (“shuck”) and the sound that you make with your mouth again (“shuck”) while you pick up that same fleet of transports later in the same turn and move it, loaded, to its destination. Try saying “shuck-shuck” while you move your transports in an endless loop. It’s fun! And it reminds you not to disperse your transports all over the Pacific where they’ll be vulnerable and ineffective. If you find yourself saying “shuck” instead of “shuck-shuck,” you’re doing it wrong. Transports are not a single-use item. Recycle!

    Optimizing the Battle of Midway for a Tokyo vs. Moscow race

    There are three general options for what to do about the Battle of Midway (attacking the American ships off the coast of Hawaii on turn 1 – it’s frequently called Pearl Harbor, but that’s not accurate, since there’s carriers there instead of battleships, and the start year is 1942).

    Option 1 is to just skip it – consolidate your fleet in Japan or off the coast of China, and let the Americans be. Option 2 is to attack with everything that can reach – a carrier, the cruiser, the sub, your bomber, and three fighters. Option 3 is to attack “lightly”, using only your cruiser, sub, bomber, and two fighters.

    The advantages of Option 1 are that you get to use your boats to support bombardments in China, and you might buy some extra time or push America into a premature attack by making it more difficult for America to calculate when American can safely move into Tokyo Bay. The more units there are on the map, the more difficult it is for players to make a correct decision, and it’s America that has to make the go/no-go decision about invading Japanese territory. The disadvantages of Option 1 are that America could send its fleet east into the Atlantic, helping to crush Germany ahead of schedule, and you give up whatever opportunities you had for a profitable battle based on a “surprise” turn 1 attack.

    The advantage of Option 2 is that you force America to commit essentially all available boats and planes to take out your fleet, seriously delaying America’s attack on Germany. The disadvantage is that you wind up losing more IPCs’ worth of units than America does, and you also commit your entire air force to the battle, which will seriously slow you down in China.

    The advantage of Option 3 is that you achieve a tactical victory by sinking more IPCs’ worth of American boats then you’re likely to lose on the counterattack, and it still leaves most of your fleet roughly in position for your next turn. The disadvantage of Option 3 is that it can be disrupted by the British Australian navy (either by attacking your Caroline Island carrier or by killing your bomber after the bomber lands), and it does slow down your attack on Asia relative to Option 1 (although not as badly as Option 2).

    My advice for the context we’re discussing here – a game where Japan is being attacked but Japan is nevertheless choosing to invest significant resources in conquering mainland Asia – is to go with Option 1 and just skip the Battle of Midway. Yes, Option 3 can turn you a pretty reliable profit – but after subtracting your losses in the inevitable American counterattack, that profit is small, on the order of 10-20 IPCs. Gaining a 20 IPC edge against the American navy for a battle that’s going to happen on turn 6 or 7 isn’t worth moving 40+ IPC of cruisers and planes out of position from your assault on China.

    Maximizing Offensive Punch With Fewer Build Slots

    One final thought: if you avoid factories because they’ll inevitably be captured by the Americans, and you set aside a couple of build slots each turn for naval defense, you’ll have to work with only about 5-6 build slots for your transports and your entire amphibious assault force.

    This has a couple of implications. First, you probably won’t need four transports for Japan. You might build a fourth transport to go pick up infantry from around the islands, but that idea is less exciting when you know America is coming for you – you may as well make it harder for them to land their planes. So, three transports is probably plenty. You won’t be able to empty Japan on turns 1-2 with only three transports – but that’s just fine, because the second implication is that sometimes you’ll want to overbuild infantry so that you never waste a build slot.

    If you build two transports and six infantry on turn 1, you’ll have way more infantry than you can ship to Asia on turn 2 – but that’s fine, because eventually you’ll have enough income to build 3 planes/ships and 5 infantry, and then you’ll start depleting your stock of infantry. For example, 1 fighter, 2 destroyers, 3 infantry, 1 artillery, and 1 tank cost $45 – well within the range of a successful mid-game Japan. In the worst case scenario where you can’t ever unload your infantry, the extra infantry serves as a home guard that makes it that much harder and more expensive for America to muster a credible threat on your capital.


    That’s all for this series. Thanks for sticking with me through a highly technical, mathematical journey! I hope you enjoyed the posts, and I’ll be around to read and reply to comments and questions. Next up on my to-do list, real life permitting, is a series on general Japanese opening strategy.

  • '19 '18 '17 '16 '15

    Excellent read with bullet points made! Thumbs up

Suggested Topics

  • 3
  • 23
  • 5
  • 5
  • 4
  • 20
  • 18
  • 14
I Will Never Grow Up Games
Axis & Allies Boardgaming Custom Painted Miniatures
Dean's Army Guys