TWS2: Logistics - maximizing efficiency

[Maltz]

D. The Big Picture

In a conventional strategy game where the human player's goal is to expand, the power of the human player can be roughly expressed by straight line, or a curve if a "retarding" mechanics such as the increasing Admin Cost is in place (such as in STW2). The AI opposition's power is usually higher than the human's at the beginning (increasingly higher on a higher difficulty). Let's draw a simple graph to make the concept more easily understood.



Veterans of strategy games should be familiar with the feeling that "only the beginning of the campaign is interesting, then it is just a test of patience". Indeed, once the human player's power surpass that of the AI opposition (Victory Certainty), there is not much challenge left. Furthermore, when the human player expands further, the AIs due to their stupidity usually do not expand as fast. So they will have a slower growth of powers, and eventually the human grows so powerful that the opposition crumbles rapidly.

What really sets STW2 apart from other strategy games is a very dramatic mechanics to turn up the challenge, the Realm Divide (RD). The RD naturally has received a lot of criticism from legions of frustrated players who are slaughtered by the AIs.



What happens in STW2 is that when the human player's province reaches a certain point, the Realm Divide is triggered, and very quickly all of the AI factions will stop fighting each other, but rather unite to oppose the human player. The power of the AI opposition increases dramatically after the Realm Divide, until the rest of Japan has all aligned against the human player (maximum challenge).

An extraordinary human player might be able to manage to expand further even under overwhelming odds. Eventually, the AIs' alliance against human is defeated and rapidly crumble. But the vast majority of human players cannot manage to overcome the huge power difference between the AI opposition and themselves. They cannot expand any further, and their campaign is stuck. Failed. Owned.

Note: There is actually a second point where the AI opposition's power is significantly higher than the human's - the very beginning. Fortunately, STW2 is friendly enough that it does not give each AI one full stack at the beginning of the game even at legendary mode.

Fortunately, there is a very easy way to work around Realm Divide, as has been pointed out by various guides and stories.



For a conventional strategy game, there is more or less a linear relationship between the human player's number of provinces and turn number. In many games, there is an exponential relationship - the border of the kingdom grows longer and the expansion accelerates. But Japan is mostly linear that we can pretend the expansion will be quite linear.

To combat realm divide, we can revise our expansion strategy. We will purposely hold off our expansion just before we hit the point of Realm Divide. The exact timing of the pause is very easily decided - just watch our yellow fame bar and stops expansion right before it fills up. We will use this waiting period to seriously strengthen our economy and then upgrade our army. After our power receives a significant boost, we comfortably trigger the RD and steamroll the AIs even if they are all against us.



This is what the power vs. # of province graph looks like with the revised strategy. Just before RD, the human player receives a large boost of power because we know the game better than they do. The AI opposition tries to match up after RD, but the human uses its superior power to suppress the AIs. Even if the rest of Japan all unite against the human player, the human is still a little more powerful. Quickly, the AI reaches the breaking point and the human wins another sweeping victory.

The above concept is probably nothing new to you. So where does logistics come into play?



The answer is that we need to find out the optimal waiting period in order to win the game as early as possible with your style. If we wait for too little, our power will not be enough to fully suppress the AI after Realm Divide, so we will expand rather slowly. If we wait for too long, sure we will overkill the AIs, but we have wasted too much time waiting.

The optimal waiting period is integrated with the consideration of all previous sections - the logistics of army, buildings, and research. Ideally, by the time that we are ready to trigger Realm Divide, our invincible armies are ready, our bank account is full and our income is ever growing, and we have researched every essential tech on the list. Most of this is gaming experience and learning from past mistakes.

Here are two graphs that I made from an actual campaign recently.

The first one is about incomes vs. number of provinces. You can see the "power boost" during the waiting period at 22 provinces.


The second picture is about number of province vs. turn number. The waiting period is between turn 54 and 78.

Concluding notes: Other than personal styles, the game's difficulty is affected by many factors: choice of clan (I have read people beating legendary Oda in 7.5 hours only by auto-resolve), game versions (subsequent patches often fix exploits and "overpowered" aspects), and random factors such as agents' mission outcome, retainers, AI's aggressiveness against the human player, and events. I think we should care less about playing the game as efficiently as possible (and never care at all whether we finish the game earlier than another person), but to find fun from exploring unfamiliar elements. When we gained more knowledge from our explorations, we can naturally make better decisions. If we make each of our thousands of decisions in a campaign based on deep knowledge, careful reasoning, and sound judgement, we have truly mastered the game, and we probably have played our style as efficiently as we can.

Thanks for reading the guide and I hope it helped!

[faker01]

The point of the tactic without preparing for realm divide is to defeat the neighbors before they can react. So it gives you some turns before you could get problems. If you conquered the center of japan before realm divide, you can get about 10 further provinces before the other attack you -> more stacks for you, less stacks may attack -> no problems.
btw, I tryed one time to win as fast as possible on legendary with my tactics. I took the 65th province in turn 61. (without loading old saves for optimize anything)

[Maltz]

You must be that extraordinary player that I could only imagine!
It will be great if you can write up a story/AAR. People are now talking about having a story/AAR competition after the video contest.

[faker01]

I'm not good on writing storys.
I think it's interessting how different we play. While you play quite inflexible and well organized, I play very flexible and reckless.
My typical logistic for reinforcments is: the nearest provinces recruit, thats all. I rarely upgrade something else then farms and mines and I rarely research C#8. I typically use only yari ashigaru. They are the efficientest unit and they make my armys flexible. My armys typically consists of 1-5 Stacks, my main-army of 1-3 Stacks. I use 2-3 armys. When units of my mainstacks got weakened to much(or got extinguished, which is normal for me), I simply exchange them. Thats why I prefer to research B#10.
edit: Normaly I split my armys for faster conquering and sometimes one stack joins another army(when they go to far for a job).

[Jacob_Sterlov]

B. Coordination of Buildings
2. Hire agents

Agents can make magnificant contributions that often surpass what military can accomplish - and their operation fee is just a fraction of the military's upkeep. For example, a ninja can forever stop an invasion as long as we have the money for it. A monk can single handedly turn many provinces in to rebels. Metsuke can bring fortunes equivalent to the richest trade node. Agents are the second priority just behind military only because they can be hired very quickly; the economy buildings can wait further.

Depending on people's preference, there are mainly two ways to use agents. The first style is more of an RPG or action game style - if an agent dies, restart from the latest save. Essentially the agents never die on the final record, although we could have compiled a long movie of all the gruesome death scenes in between. This style of course allows best efficiency of agents, and we only need 5 provinces to hire all 15 agents (who will last forever).

My personal preference is the order of Ninja -> Monks -> Metsuke. At the start of the game, I would build a Sake Den in every province. Even if my capital starts with something else, I demolish it and build a Sake Den. When I have 5 provinces, I already have 5 ninja. Then I demolish the Sake Dens to build 5 Temples. After hiring 5 monks, I demolish the Temples to build 5 Markets and hire 5 Metsuke. The markests stay to make money for the rest of the game.

Why Ninja first? Ninja often makes a big difference right from the beginning. They cripple an enemy so much that it almost guarantees the success of our early expansion effort. Training ninja by sabotage is cheap, as the lowest-level building only costs 100 koku to sabotage.

Monks start to shine when they can incite rebellion. It would be nice if we can find a Holy Site (Mori - Aki, Hattori - Yamato, Oda - Ise, Date/Uesugi - Uzen, etc.) to give the fresh recruits level 3, so they can recite rebellion right away. In this case, we can demolish the Sake Den one by one, since we only hire one monk from the designated Holy Site province each season. The province takes longer but the better quality monks should make up for it. Training monk is free (demoralize army).

Metsuke is the most difficult agent to level up as they have to seek enemy agents (only a few clans produce them consistently - Ashikaga in Kyoto and Honma on Sado island comes to mind) or bribe crap units from a warring clan. So it is better to wait for a province with the school capacity (Mori, Shimazu, and Chosokabe look for Tsukushi, Hattori and Oda look for Settsu, etc.) to hire them. Once they reach level 4 (four missions above level 3), they can reach 11 stars of Overseeing Towns, making them very good money-makers in our richest provinces.

I tried the save-game trick with the ninja, but it's not working out well.

I know enough of statistics to understand my ninja can get even 100 fails in a row with a 50% chance.

However, since I know PC games are not truly randomic, I am wondering if there is something about the clock... or something else: namely, odds are calculated at the start of the turn so you have to reload the turn before and not the actual turn in order to "re-roll" the dice.

Is that so? I am writing this after 20 fails out of 20 attempts at 50%. I can go on if it is just a matter of chance... but if there is something about the game mechanism, then I shall stop.