The day we decided a flower should rot
Same Seed has a rule that I am a little embarrassed by, and I want to tell you about it anyway. A bloom that sits too long after it opens loses value. By the third day it is worth less than it should be. By the sixth day it is worth half. This is a cozy garden game with a rot mechanic, and for a week I kept trying to delete it. The numbers told me to keep it. I may still be wrong, which is why I am writing this down.
The hole we were trying to fill
Same Seed is a twelve day garden. Nine plots, six species, three actions a day, thirty six in total. A daisy opens in three days, a sunflower in eight. The problem was what happens after a flower opens. Without any cost to holding, the best way to play was a logistics problem, and a logistics problem is a game where the interesting decision is made on day one and the next eleven days are bookkeeping, which is not a game at all, and I had spent a lot of time trying to make this feel like the opposite of a spreadsheet and the opposite of a spreadsheet is a garden where something is always quietly changing and you have to notice it.
The real problem ran deeper than boredom. Nothing was forcing a harvest day. With thirty six actions in the whole season, every action spent standing in a finished field is an action you no longer spend planting, and yet the game was quietly telling players that hoarding was fine. I wanted the calendar to be a third character in the game, next to the soil and the weather, and it could not be a character while it cost nothing to ignore.
The rule, and the floor I argued with myself about
Two constants do the work. A bloom is at full price for two days after it opens. After that it fades fifteen percent a day. The floor is fifty percent, so a held flower never becomes worthless. In the code the names are blunt: PRIME_DAYS, FADE_PER_DAY, FADE_FLOOR.
A sunflower is the most expensive flower in the game at 96 per bloom. With the fade applied it sells for 96, then 82, then 67, then 53, then 48, and 48 is where it stops. Leave it four days and you have given away half of it, and the remaining half is safe. A lily at 42 fades 42, 36, 29, 23, 21. Same shape, different weight, which is most of the point. The expensive flowers hurt more when you forget them, so the same rule plays differently on every plot.
The floor is the part I argued with myself about the longest. The first version faded all the way to zero. A sunflower left for six days was worth nothing. That felt wrong in a game whose entire pitch is that it is gentle. Losing a plant to a dry spell is the garden being fair. A bloom rotting to zero feels like the garden is punishing you for being a human who has other things to do. Fifty percent makes the worst case half price instead of nothing, and that difference is the whole argument for the rule existing at all.
The table I did not want to read
The rule is in the code. That was the easy part. The part I hate is the question that follows: did anyone actually change their behaviour? Feel cannot answer that, so I wrote a probe. It plays the same four hundred daily seeds the balance probe uses, with four bots, and counts how many harvests came in at prime value and how many came in faded.
The fully random player made 434 harvests, 60% at prime, 40% faded, and lost 638 plants to wilt along the way. The casual bot, the one that cuts whatever looks ready and waters only what is visibly parched, made 1,118 harvests and every single one of them was at prime value. The greedy bot made 3,626, with 92% at prime and 8% faded. The variety line made 3,771, and 99% of those were prime.
Read that table twice before you smile, because the awkward part is real. A bot that harvests the moment something is ready never sees a faded flower, by definition, and the casual bot is a bot that harvests the moment something is ready, so of course its faded count is zero, and the probe is telling you nothing about whether a human being who has a job and a family and a life outside this garden will still make the same choice on day nine when the sunflower is one day past prime and the kitchen is calling and the game is waiting for you to come back and the question is whether you care enough to notice. What the probe measures is the difference between planning and not planning. It cannot measure a player who plans and still gets lazy. If my only evidence is that table, then I have proved the rule is consistent, not that it does anything, and I want to be that honest about it.
The number that did move
The part that moved was somewhere else. Our goal probe checks small seasonal targets against the same bots, and one of them is cutting three flowers in a single day. The casual bot clears that goal 10% of the time. The greedy bot clears it 95%. That gap is the fade doing its job, because a three cut harvest day only exists if blooms are being lined up on purpose, and the fade is what makes lining them up a skill instead of a preference. Without it, three cuts in a day is just three cuts eventually. With it, you have to decide two days in advance which day is the harvest, and which flowers you are planting to get there.
The number I almost made too big
Fifteen percent a day is a compromise, and I want to show the arithmetic that kept it there. At twenty percent the sunflower curve runs 96, 77, 61, 50, and it hits the floor on day five. The rule stops doing its work exactly when I most wanted it working, on the last days of a season when a player is weighing whether to hold one more day. At fifteen percent the curve is still moving on day six, which leaves a real decision at the edge of the season, the kind where one action matters more than usual. I am not claiming fifteen is the right answer. I am claiming twenty was a wall, and a wall does not leave you a choice to make.
What I still do not know
Bots are bots. I do not know what a human feels when they come back to a bloom that has been waiting four days. Whether they feel the loss or do not look back. The probe cannot measure that, and I would rather say so than pretend the four hundred seeds settled it. The fifty percent floor is my best guess at the line where a price is high enough to make planning matter, and low enough that a player who returns after a long break is not punished by the garden for having a life. If the logs of real seasons show me a different shape, the fade changes, and because that would move a score it becomes RULESET 3, with every old submission still verifying under the rules it was played with.
A cozy game can have a cost for doing nothing. The question is what you charge. I chose half a flower. The table says that was probably right. The players will tell me for sure.