Sengoku Clans · Ethan Ding

Each ribbon is land (in millions of koku) passing from the defeated to the victor. Red ribbons = clan destroyed or usurped. Gold ribbons = forced submission or transfer — the clan survives, reduced, under the victor's suzerainty. Hover any ribbon for the battle or event behind it. Every stream ends at the Tokugawa.

Why they were so powerful. Three structural advantages, compounding for a century. (1) The best land in Japan: the eight Kantō provinces assessed at ≈2.4M koku — for scale, Takeda Shingen's entire home province of Kai was 0.23M; the Hōjō were farming a prize ten times the size of their rivals' bases. (2) Five generations without a succession war (Sōun → Ujitsuna → Ujiyasu → Ujimasa → Ujinao, 1493–1590) while every rival tore itself apart — Shingen had to depose his own father, Kenshin's death triggered the Otate civil war, the Imagawa fought two succession crises. (3) The era's most modern administration: cadastral surveys, fixed taxes at "four to the lord, six to the people" (≈40%, low for the age), rational magistrates and post-stations — peasants fled into Hōjō territory. Behind it all sat Odawara, a fortress-city that shrugged off sieges by both Uesugi Kenshin (1561) and Takeda Shingen (1569).

Why they stayed on top while Takeda, Uesugi and Oda warred. Strategy: they never marched on Kyoto. Kenshin and Shingen bled each other white in five battles at Kawanakajima and endless Kantō campaigns; Nobunaga fought a decade of wars in the center. The Hōjō fought defensive, regional wars and spent the difference digesting the Kantō. By 1589 they were at their all-time peak — ≈2M koku — having never lost their home.

Why they fell. The same strategy that made them rich made them alone. The map inverted: instead of the biggest of several regional powers, by 1589 they were the only independent power left, facing a unified ≈16M-koku Japan. They misread Hideyoshi as one more regional rival who could be stalled — delaying homage, haggling over hostages — until a retainer's seizure of Nagurumi Castle (1589), in violation of Hideyoshi's peace edict, handed him his casus belli. Their war council debated fight-or-submit so interminably that Odawara hyōjō ("an Odawara conference") is still the Japanese idiom for a meeting that decides nothing. In 1590, ~210,000 troops and a naval blockade surrounded Odawara; it surrendered in three months without a decisive battle. Ujimasa was ordered to seppuku, and the 2M-koku Kantō was handed to Tokugawa Ieyasu — the transfer that made Ieyasu the largest landholder in Japan.

Handing Ieyasu the richest region in Japan looks insane until you see what Hideyoshi took in exchange. It was a forced swap, not a gift: Ieyasu had to surrender his five ancestral provinces — Mikawa, Tōtōmi, Suruga, Kai, Shinano (≈1.4M koku) — the lands where his family had ruled for generations and where every village headman and retainer network was personally loyal to him. In return he got ≈2.5M koku of freshly conquered, unpacified territory, a ruined castle town in a marsh (Edo), and an administration to rebuild from zero.

From Hideyoshi's chair, the logic was containment: (1) it uprooted Ieyasu from a 40-year power base; (2) it moved him 300 km farther from Kyoto and Osaka, behind the Hakone mountain passes; (3) Hideyoshi packed Ieyasu's old provinces with his own loyalists (Yamauchi, Horio, Nakamura and others in the Tōkaidō castles), building a wall between Ieyasu and the capital; and (4) destroying Ieyasu wasn't on the menu — Komaki–Nagakute (1584) had already proven Ieyasu could beat Toyotomi armies in the field, so Hideyoshi bought him instead. Land was the only currency of loyalty in this system: Hideyoshi kept just ≈2.2M koku directly and ruled by granting the rest.

The gamble failed on every axis. The Kantō turned out to be under-assessed and Edo boomed; the clean-sheet move let Ieyasu build the most rationalized domain administration in Japan, unencumbered by old obligations; the Hakone "wall" worked both ways, making his base unattackable; and the moment Hideyoshi died, the Tōkaidō buffer lords defected east. The exile became the throne — Edo, the marsh town, became Tokyo.

Hideyoshi died leaving a five-year-old heir and a Council of Five Regents (go-tairō) sworn to protect him, checked by five commissioners (go-bugyō) led by Ishida Mitsunari. The bars show why the arrangement was doomed: Ieyasu alone held more land than the next two regents combined, and the only elder with the stature to restrain him — Maeda Toshiie — died within a year of Hideyoshi. Ishida, a 190k-koku administrator, tried to rally the rest against Ieyasu: three of the five regents ended up on the losing side. Hover a bar for each man's fate.

All of the era's major houses, in order of when they folded. "Folded" takes three forms: destroyed (line extinguished or stripped of all land), submitted (surrendered independence, kept a reduced domain), or survived to the Meiji era (1868). Note the ending: the two great clans the Tokugawa merely reduced rather than destroyed — the Mōri (Chōshū) and Shimazu (Satsuma) — returned 268 years later to lead the revolution that destroyed the Tokugawa shogunate.

How the koku estimates were made (Fermi method): Province-level kokudaka from the 1598 Taikō survey are the anchors — e.g. Kai ≈ 228k, Shinano ≈ 408k, Echigo ≈ 391k, Owari ≈ 572k, Musashi ≈ 667k, Satsuma ≈ 283k. Each clan's holding in a given year ≈ (provinces fully held) + (fractions of contested provinces) + a premium for non-land income where it was decisive (Ōuchi's China tally trade, Takeda's gold mines, Oda's control of the capital region's commerce ≈ +10–20%). Documented checkpoints keep the estimates honest: Mōri Terumoto assessed at 1.12M koku (1591); Tokugawa Ieyasu's Kantō transfer at ≈ 2.5M koku (1590) — making Ieyasu larger than Hideyoshi's own direct domain of ≈ 2.2M koku plus ≈ 2M held by Toyotomi house kin; Uesugi Kagekatsu moved to Aizu at 1.2M koku (1598); Shimazu confirmed at ≈ 0.6M koku after their 1587 submission; Mōri cut to 0.37M and Uesugi to 0.3M in the 1600–01 post-Sekigahara reckoning. Pre-survey decades use the same land at its later assessed yield, so early figures are koku-equivalents. Error bars are roughly ±30%.

Two things that look wrong but aren't. (1) The total never shrinks: the stack always sums to 18.5M koku — land doesn't disappear when a clan falls, it's regranted. The bottom of the chart shows this explicitly: Independent minor clans (gray — hundreds of unaligned houses in 1467, squeezed to exactly zero at Odawara in 1590) drain into three vassal bands tinted by their master's color — Oda vassal lords (1568–82), reborn as Toyotomi vassal lords after Honnō-ji, converted almost overnight into Tokugawa vassal lords at Sekigahara, when ≈6.5M confiscated koku were regranted to the Eastern army. Watch the tint change hands at 1582 and 1600: those recolorings are the transfers of national power — often the very same lords (Fukushima, Kuroda, Hosokawa) under a new master. (2) The Ashikaga at ~11% of Japan in 1467 is, if anything, generous. The Muromachi shogunate was land-poor by design: its direct demesne (goryōsho) was small and scattered, and it financed itself by taxing Kyoto's moneylenders and sake brewers, Ming trade licenses, and its monopoly on appointing shugo. Its true asset was legitimacy — which is why, once the Ōnin War burned that legitimacy down, there was almost no land underneath to fall back on.