The Gem Sea Journal

PELICAN TOWN – Every farm accumulates them: the seed packets that tumble out of cleared weeds, tilled dirt, and fishing chests, sell for 0g at every counter, and carry no season on the label. Mixed seeds are the Valley’s loose change, and the community treats them accordingly, scattering them wherever there happens to be dirt. The Journal’s finding is that the scattering is the mistake. Because the seed is free and never expires, the question is not whether to plant it but when, and the answer moves the same packet’s value by a factor of eleven. The mechanism that makes timing everything is a single line of the game’s code: the crop is not stored in the seed. It is rolled from the current season’s pool at the moment of planting.

§1The mechanics of a seasonless seed

The pools are small and the rules have teeth. Spring rolls parsnip, potato, or cauliflower; summer rolls corn, hot pepper, radish, or wheat; fall rolls artichoke, corn, eggplant, or pumpkin. Within a season the draws are uniform, with one documented exception: on the farm in spring, the game’s Crop::ResolveSeedId routine weights parsnips double against potato and cauliflower. Spring’s pool is rigged toward its cheapest member, and the desk found no forum myth here; it is in the code. Winter closes the mainland outright: mixed seeds cannot be planted outdoors at all, and an indoor winter planting rolls a crop from any season. One circulating belief fails its check: Powdermelon is not in the mixed-seed pool. Its seeds drop separately, from artifact spots between Fall 21 and Winter 20.

Supply is everywhere once it is looked for: a 5% drop from every weed, raised to 9% by reading the Raccoon Journal; tilled dirt; fishing chests; golems on the Wilderness farm; the Seed Maker’s 1.99% whiff; Mystery Boxes; the Forest Farm’s special weeds, which always drop them; and Krobus, who sells them for 30g every Thursday, a price this report audits in Section 6. Two legal oddities round out the mechanics, and both earn their keep later: fall mixed seeds can roll an artichoke in Year 1, a crop whose seeds are otherwise locked until Year 2; and one mixed seed plus 5 sap and 1 clay crafts Fiber Seeds, which need no watering, so every packet doubles as a voucher for free fiber.

§2What a day-1 seed becomes

The desk’s model holds Tiller, normal quality, free seeds, daily watering, and no fertilizer, since a free random seed does not merit Speed-Gro. Two facts organize the whole table. First, the best mainland outcome is not fall’s 352g pumpkin but summer corn: planted on Summer 1 it regrows every 4 days and survives the season change, delivering eleven harvests across two seasons for 605g, from a seed that cost nothing. Wheat shares that immortality at a humbler 27g, and between them they are why summer cannot bust. Second, spring has no jackpot at all. Its ceiling is a 192g cauliflower, and half of all spring rolls are a 38g parsnip, which is how the season’s expectation lands at 95g against summer’s 271g and fall’s 270g. The full roll table, odds and harvest counts included, is in the appendices.

§3Three ways to decay

Day-1 averages understate how differently the seasons age, and the aging is where the strategy lives. Spring is a step function: nothing in its pool regrows, so value holds flat at 95g until cauliflower busts on day 17 (down to 47g), potato on day 23 (19g), and everything by day 25. Summer is a gentle slope that never reaches zero: each passing day shaves regrowth harvests off corn and pepper, but nothing can bust, and a day-28 seed still returns 62g. Fall is a cliff: full value through day 2, corn unplantable after day 14, pumpkin after day 15, and the pool is 100% bust from day 24 on. A fall seed sheds about 70% of its value between day 1 and day 16, which the desk condenses to a rule a farmer can carry: fall mixed seeds are a first-fortnight product, and a late-fall planting is strictly worse than not planting, because it also wastes the seed.

§4Filler doctrine

Granting mixed seeds a dedicated tile for a whole season, replanting the moment a single-harvest crop comes off, the desk’s dynamic program returns season totals of 323g per tile in spring, 466g in summer, and 373g in fall: 12 to 17 gold per tile-day against roughly 40 and up for strawberries, 45 and up for blueberries, and 55 and up for cranberries on the same dirt. The seeds are free; the land and the watering are not. The correct deployment therefore follows what this desk calls filler doctrine: mixed seeds go on tiles that would otherwise sit idle, the ragged edges outside sprinkler range, the row that was tilled but could not be afforded, Year 1 dirt before capital exists. They should never displace a crop a farmer was willing to pay for. Their honest pitch is not gold per tile; it is gold per gold invested, and against a 0g basis that figure is infinite.

§5The island exception, and the honest answer for winter

Ginger Island rewrites the question. Island plantings ignore the season entirely and roll a fixed, uniform pool of blueberry, melon, pineapple, and rhubarb, and island crops never die to a season change. Two of the four are regrowers with no horizon: a rolled pineapple pays 330g every 7 days forever, a rolled blueberry about 166g every 4 days forever. Measured over the same 56-day window as a summer corn, one island seed expects about 1,081g, four times the best mainland day, and the figure is conservative, since the regrowers keep paying after the window closes. In steady state a dedicated island tile converges to roughly 35 to 40 gold per tile-day, free, forever, with a quarter of all rolls becoming permanent plants. So the question in this report’s title has a sincere answer and a better one: among mainland seasons, summer; but the globally optimal season to plant a mixed seed is winter, on Ginger Island, where the mainland offers 0g and the island offers its usual 1,081g.

The desk flags one tempting substitute as a trap. Winter greenhouse plantings do roll from all seasons, and a lucky corn regrows indefinitely indoors. But greenhouse dirt is the most valuable in the game: permanent corn earns about 14 gold per tile-day on a tile where raw ancient fruit earns roughly 85 before kegs. Planting mixed seeds in the greenhouse spends an 85g/day tile to earn 14. The island’s free dirt is the right home for gambling; the greenhouse is not.

§6The arbitrage desk

Because the seeds never expire, “which season is best” is really “how long should I hold?” A seed found in spring is worth 95g planted on the spot, 271g held eight weeks for Summer 1, a gain of +186%, and about 1,081g held for island access. The one honest argument against holding is Year 1 liquidity: early gold compounds through seed purchases, and 95g of free value on day 3 of a new farm beats 271g in five weeks. The desk’s banking rule follows: in Year 1 spring, plant on sight; every spring after, bank for summer; once the boat is repaired, bank for the island.

Krobus’s Thursday counter prices the same logic in reverse. At 30g a packet, every summer Thursday is a buy, from an 8.2× return on Summer 4 down to a still-respectable 2.5× on Summer 25, the one month where even the last Thursday pays. Spring 25 and Fall 25 are the two days Krobus is charging 30g for merchandise that is dead on arrival, unless the buyer is banking, in which case every Thursday of the year is a buy: 30g against a future 271g summer planting is a 9× trade. One clause serves the completionists: artichoke seeds do not reach Pierre’s shelf until Year 2, but fall mixed seeds roll them in Year 1 at one in four, so sixteen early-fall seeds give a 99% chance of the Year 1 artichoke a tight Full Shipment run needs. And three things the gamble cannot do, for the record: giants need a 3×3 of one crop and random rolls cannot be clustered, so mixed seeds never produce giant crops in practice; the Fall Crops bundle’s yam is not in any pool; and a strategy built on free random seeds has no business buying Deluxe Fertilizer. Ship what drops.

§7The verdict

The ranking, complete: the island, then summer, then the first two weeks of fall, then spring, then the greenhouse, then nothing at all, and then late fall, which is strictly worse than nothing because it spends the seed to earn it. The Valley hands every farmer a trickle of free lottery tickets, and the desk’s finding is that the tickets have a settlement date printed in invisible ink. Read it, and the loose change compounds; scatter them on sight, and the same change buys parsnips.