Takeda Shingen's legacy: banner, house law, flood banks, and a loud afterlife

This page is the afterlife, not the life. For the man himself, use Takeda Shingen. He was born in 1521, ruled Kai (the basin around modern Kofu, in Yamanashi), and died in May 1573 while campaigning toward Tokugawa territory. A sniper story at Noda Castle shows up in later retellings. The stronger reading is illness, and he was dead by May 1573. Two dates belong to his son Katsuyori and get stapled onto the father anyway: Nagashino in 1575, and the end of the house at Tenmokuzan in 1582. Tenmokuzan is also spelled Temmokuzan. Sekigahara, in 1600, is not on this list at all.

What remained is smaller than the parade, and more concrete. A four-character banner. A set of house laws. A flood bank on a bad river, plus a traditional map of wedges and openings around it. A rivalry with Uesugi Kenshin that never produced a clean winner. A crowd of officers whose names got packaged, much later, as twenty-four generals. Then a twentieth-century festival, game characters, and business books. The clan was destroyed. The name got louder afterward. That gap is the subject.

What remained after the house was gone

A legacy page earns its keep by naming jobs. Each row below is something people still point at when they say "Shingen," and the middle column is the job it actually had. The right column is the mess: later books, missing paperwork, or a tourist slogan wearing a historical costume.

If you want the wars themselves, use Shingen's battles and the clan timeline. This table is only the leftovers.

Four characters on a cloth

風林火山, read furinkazan, is wind, forest, fire, mountain. The line sits in Sun Tzu's chapter on maneuvering an army: be as fast as the wind, as unhurried as the forest, invade like fire, stand like a mountain. The Chinese sentence does not stop there. It goes on to say the army should be as hard to read as a shadow, and that when it does move it should move like thunder. The banner drops those last two images. People meet a slogan and think they have read the chapter.

A banner is a marked cloth on a pole. In a noisy field it tells men who cannot hear the command which behavior is wanted: speed, order, attack, or refusal. Takeda forces carried standards of this kind, and Kofu's Shingen Treasure Hall displays a "Sun Tzu flag" as a war banner traditionally said to have stood at the head of the army. "Traditionally said" is the honest museum tense. Do not upgrade that into a proven 1561 original with a battle scar and a signature.

The same four characters show up in games as a technique, sometimes with a flash of light. That is a control scheme. It does not document a rite, a curse, or a formation that made horses faster. For how the army actually fought, use Shingen's weapons and style. The banner branded a preference. It did not win the next century.

House law for Kai, not a national code

Koshu Hatto, fully Koshu hatto no shidai (甲州法度之次第), is the domain law of Kai under Takeda Harunobu, the name Shingen used before the priestly one. A University of Tokyo legal-history copy is one of the best surviving witnesses. Its catalog note lays out the usual scholarly story: a 26-article text made in the 6th month of 1547, later rearranged into 55 articles, with two more added in the 5th month of 1554. The copy itself is dated the 17th day of the 2nd month of Tensho 8 (1580), under Katsuyori, and it carries a seal the heir kept using. That volume is the base text in the Iwanami medieval house-law collection published in 1965.

There is a rival reading. Some historians argue the 26-article version is a later digest and that 55 articles were the working code from early on. Alternate titles float around the same material: Koshu shikimoku, Shingen kaho. If a summary gives you one article count and no caveat, it picked a side and hid the other. Both readings stay visible here because the paperwork really does split.

What the clauses did, on the points reference works agree on, was local discipline for retainers. The code sits close to the Imagawa house's kana mokuroku, the law book of the neighboring domain, which is a polite way of saying Kai did not invent the genre. Three jobs show up again and again in the dictionary entries:

• Kenka ryoseibai: if retainers fight a private quarrel, both sides take the penalty. The lord does not have to crown a winner, and the fight stops being a legal bet.
• A ban on private alliances, so captains cannot build a side coalition that the household has not approved.
• A ban on shuron, arguments about religious doctrine, so a sect fight does not become a split inside the camp.

That is administration. It tells you how a Sengoku domain tried to keep its own men from becoming a second government. It does not tell you tax rates. The text traveled because the Koyo Gunkan, the early-Edo military chronicle of the Takeda, copied a version into a book people actually read. Survival in a popular war story is not the same thing as enforcement in every village in 1547. Still, a law bad enough to ignore would not have been worth copying under Katsuyori in 1580, two years before the roof fell in.

Flood banks on the Kamanashi

West of Kofu, two rivers share a basin and a bad habit. The Kamanashi runs along the basin. The Midai drops off the Southern Alps onto a fan and, in a storm, used to shove gravel and water toward the fields. Together they feed the Fuji, one of Japan's steep river systems. Shingen-zutsumi is the name attached to the banks and to the wider plan for calming that junction. Kai City's page says the work is said to have taken 17 years and to have been finished in 1560. The national river office tells the same story in more steps. Both are public accounts of a tradition, not a set of surviving blueprints. You will not find a sixteenth-century height, length, or volume here, because repeating a guessed meter would be fiction.

A Japan for Sustainability write-up, leaning on the Kai kokushi of 1814, is blunter than the tourism boards. It treats the embankment itself as the piece those later writers were willing to pin on Shingen, and it says the other structures are still short on war-period records. The city, for its part, notes that the techniques were not brand-new gadgets. What changed, on that telling, was scale: one domain strong enough to run a basin-wide plan instead of each village patching its own leak. Read the sequence below as the plan those offices teach. Then keep the 1814 caution in your other hand.

1. Stone-stack spurs on the Midai fan were meant to push that steep flow north instead of letting it spread across fields. The city page counts six of them in the traditional layout.
2. Two wedges called shogi-gashira, one at Arino in Minami-Alps and one at Ryugaoka (also labeled Tatsuoka) in Nirasaki, were meant to split the current so it hit the Kamanashi in pieces. The shape is compared to a shogi piece's head, which is a description of the plan, not a claim that anyone played a board game on the dike.
3. A cut through the terrace, Horikiri, plus a line of sixteen large stones in the usual telling, was meant to settle the merge and aim the joined water at Takaiwa, a natural cliff that kicks the flow south, away from the middle of the basin.
4. Below the cliff, on the left bank near Ryuo, the embankment proper is the wall in front of the fields. Projecting banks and wooden frames called seigyu ("ox," from the horned triangle of the base) sit in the current, weighted with stone-filled baskets, so the water spends force on timber before it spends it on the dirt bank. Those frames get rebuilt. You are not looking at a log from 1560.
5. Openings in a kasumi-tei, a discontinuous levee, were the plan for failure: if water came over the top, it had a path back to the channel instead of sitting on the village. The river office also describes a maintenance settlement at Ryuo, excused from tax in return for keeping the works, and a flood-prevention procession, Omiyuki-san, still put on 15 April.

An Edo handbook, the Jikata hanrei roku, is the source for the claim that Shingen was the first person in Japan to use seigyu, and that the idea later spread to other rivers. That is a claim in a later manual. It is not a patent. After the Takeda fell, the basin still needed the bank. Edo-period Kai, for long stretches a Tokugawa holding, kept flood work on the local agenda. A modern interview in the magazine Mizu no bunka records the local story that village officers asked the new rulers to keep the arrangement. No reliable wording from Ieyasu belongs in quotes here. The solid point is duller: a useful ditch does not care which surname collects the tax.

The city is frank that the bank has been repaired over and over. Visitors at Ryuo can still see the line, and the tourism board points out ox frames along the water. "Still there" means the place is still defended and still explained. It does not mean the 1560 fill is untouched under your shoes.

A poster of generals, and a plain that would not end

"Twenty-four generals of Takeda" sounds like a council with a fixed chair count. It is a fame list. Japanese reference writing defines it as the two dozen retainers later war tales and storytelling rated highest, not as an office Shingen created. Edo ukiyo-e and puppet plays loved the set. The names move. Some pictures count Shingen himself as one of the twenty-four, which leaves twenty-three commanders. Some drop men, some add Katsuyori or a Sanada whose main career came after Shingen was dead. A print can show Itagaki Nobukata, killed at Uedahara in 1548, beside an officer who mattered in the 1570s. That is a poster, not a staff meeting.

The men inside the better-documented names were real, and several of the famous ones died with Katsuyori's army at Nagashino: Yamagata Masakage, Naito Masatoyo, Baba Nobuharu. Their deaths are part of the legacy of the list, because the list freezes them as Shingen's companions even when the battle that killed them was his son's. By 1582 almost nobody later painters liked to include was still carrying a Takeda banner. The number 24 did not save a payroll.

Kawanakajima is the other memory engine. Between 1553 and 1564, Shingen and Kenshin met five times on the plain north of today's Nagano. Neither house erased the other. The fourth meeting, in 1561, is the one later books load with a duel: Kenshin breaking into the headquarters and striking at Shingen, who takes the blow on a war fan. Shingen's brother Nobushige did die in that campaign. The duel itself is a scene from military tales, above all the Koyo Gunkan, which was put together after the clan was gone and which supplies a lot of the color we think is "the" battle. Treat the rivalry as real and long. Treat the single combat as a story people wanted. The Sengoku period is full of borders that grind for a decade and then get remembered as one afternoon.

Former Takeda men in other houses

Katsuyori inherited a working domain and a dead father's reputation. He pushed it until Nagashino, where prepared gun lines and field obstacles broke the attack and killed the veterans people now put on posters. Kai was still Takeda land after that. In 1582 Oda Nobunaga and Tokugawa Ieyasu invaded, Katsuyori ran, and he and his heir died at Tenmokuzan. The Takeda clan page covers the house as a house. Here the point is smaller: the surname ended as an independent power, and a lot of the expertise did not die on the mountain.

Survivors scattered. Anayama Nobukimi had already switched to the Oda-Tokugawa side, so "loyal until the last banner" was never true of every famous man. Others took new pay. The cluster historians can actually describe sits with Ii Naomasa. In the scramble of 1582 he brought in former Takeda vassals, among them men of the Tsuchiya, who had served Tsuchiya Masatsugu. Published counts of that oath disagree (one account says 70, another 120), so the headcount stays unset. The equipment story is firmer. Takeda units under Obu Toramasa and then Yamagata Masakage had fielded red-lacquered armor. Ii's troops became known for the same red, and the Ii house kept red as a military dress into the Edo period. Color and some of the men moved. The domain did not.

Texts moved too. Obata Kagenori, son of the Takeda retainer Obata Masamori, served the Tokugawa and is tied to Koshu-ryu, the "Kai style" of military study that daimyo schools later treated as a curriculum. He is also credited with a large share of the Koyo Gunkan. That book is why so many sharp little Shingen anecdotes feel firsthand. They were written up when the witnesses were old or already gone, inside a culture that wanted a model general. Use it as a window on memory. Do not use a speech from it as something a clerk wrote down in the camp in 1561.

Festivals, games, and the business-book Shingen

The headquarters outlasted the family as a place. Tsutsujigasaki, the residence Nobutora built in 1519, is a national historic site. Takeda Shrine was founded there in 1919, on the east bailey, after a 1915 grant of junior third rank tied to the Taisho emperor's accession gave local backers a reason to build. The shrine is a modern religious address on a real medieval footprint. It is not a hall Shingen dedicated.

The Shingen-ko Festival is younger than the shrine. Yamanashi's own festival history traces a blossom fair at Maizuru Castle Park from 1947, a remodel into the "Kofu Shingen Festival" in 1966, and the present name on 12 April 1970. The first years used prefectural staff as the stand-in Shingen. From 1995, after the 1988 NHK series Takeda Shingen had made him a national television face, the role went to actors. The main event, Koshu gundan shutsujin, restages a departure for the Kawanakajima fight. The official site puts the armored crowd above 1,000 people. The Kofu tourism association adds about 30 horses. Organizers call it the world's largest warrior procession. That is their billing.

It is not an annual rite from 1570, and it has not run as an unbroken "52 years in a row" machine. The 2026 edition was billed as the 52nd even though the first under this name was 1970, which only works if some years dropped out. The festival committee's own notice for 3-5 April 2026 reported 123,000 visitors across three days, and it mentioned rain on the Saturday. A rainy year is a count, not a law of nature. The April date still confuses people who have read "May 1573" and stop there. The festival follows the old calendar's 4th month, 12th day. Same memorial, different number printed on the page.

Screens did the rest. Long-running games such as Koei's Nobunaga's Ambition and Samurai Warriors put him on a select screen with a furinkazan flourish and a cavalry bonus. That teaches a personality: the cautious tiger, the rival of Kenshin, the man who would have reached Kyoto if the illness had waited. The last clause is a wish. He died on a Mikawa campaign short of Kyoto, and his son could not hold what he had.

Business shelves do a stranger rewrite. One real title in that pile is Matsumoto Yukio's Takeda Shingen no bijinesu seiko-gaku, a guide that pulls office lessons out of the warlord. It exists; Kinokuniya sells it. It is advice literature. It is not a source for the 1547 law code, the stone count on the Midai, or anything a retainer was paid. Once a destroyed clan becomes a life coach, the reader has left Kai and entered a bookshop. The levee, the house law, and the 1582 dead end are still there when the coach talk stops.

FAQs