The optimal moment to sell most Diamonds in MLB The Show has already passed. By the time you noticed the card was getting good, the people who notice these things for a living had already sold to you.

I used to think that was a vibe. Then I pulled the numbers: 219,122 daily price records, every Diamond on the market, mid-March through early July. It's not a vibe. It's a decay curve, and it starts the day the card hits packs.

This is not advice for the floor. The floor is fine. This is advice for the cards in your inventory right now that you bought for two hundred thousand stubs, that are currently worth one hundred and twenty, and that you keep thinking will recover.

They will not recover. Here's the proof, and here's what to do about it.

One rule before the data: pull-only cards

Everything below is about cards you can only pull from a pack. I threw out the earned cards on purpose — the ones you grind out of an XP path, a program, or a collection. Those lose value for a different reason: the more people finish the grind, the more copies hit the market, and the price bleeds from supply you can't see on any pack odds. That's the grind talking, not the market. If you want to understand market timing, you have to look at cards whose supply comes from one place: people opening packs. So the timing rules below are all pull-only — but the earned pile decays too, differently and worse, and the contrast turns out to be the most useful thing here. I'll get to it.

Exhibit A: Garrett Crochet

The 94 Vintage Crochet hit packs at around 50,000 stubs. That was about the ceiling. Watch what happened next:

• Day 7: 44,000
• Day 14: 33,600
• Day 21: 25,750
• Today (~week 5): ~21,000

A clean 58% slide, no earning involved — just pack supply piling up while demand thinned out. And he had company. The whole Vintage series dropped together in May, and every diamond in it — Bell, Pendleton, Bogaerts, Sánchez, Utley — traced the same curve on the same days. That's the tell that it's the market, not the player: a dozen unrelated stars don't decay in lockstep by coincidence. They decay in lockstep because they all hit packs at once.

Crochet's 90-day price history. From a ~50k launch to the low 20s — the shape of a pack card with nowhere to go but down.

Crochet isn't weird. He's the shape of the whole pack market.

The half-life is real, and I can draw it

Every pull-only Diamond has a decay curve. I lined up the diamonds that debuted cleanly in our window and tracked what the median one was worth as the days ticked by, as a share of its launch price:

Median pull-only diamond, as a share of its debut-day price. Down ~7% in a week, ~24% in three, ~40% by six weeks. Steady, one-directional erosion.

Down ~7% by the end of week one. Down ~15% by two weeks, ~24% by three, ~40% by six weeks. That's the median — half the cards do worse, and the chase cards do a lot worse. "Half-life" isn't a metaphor I reached for. It's the shape of the data.

Three forces drive it, and they only point one way:

1. Supply accumulates. Every day, more copies get pulled and listed. Pack supply never shrinks.
2. Demand shrinks. Everyone who wanted the card and could afford it bought it in the first 48 hours. After that you're selling to a smaller and smaller pool.
3. Competition arrives. SDS drops new content every week, and every drop introduces cards fighting for the same lineup slots.

The only thing that reverses decay is an upgrade — a Roster Update boost, a re-release, a meta shift. Upgrades are rare. Decay is constant. We'll come back to the exception, because it's a real one.

Pulled or earned? The crater is a different shape

Here's the payoff for splitting the piles. Cards you grind out — programs, XP paths, collections — decay on a different clock. Their supply doesn't trickle in from packs a few at a time; it floods the market the week a wave of players finishes the grind. So the crater is front-loaded.

Same six weeks, two very different paths. Earned cards do most of their falling in the first week; pull-only cards bleed out slowly. They meet at the bottom.

A pulled diamond is down about 7% after a week. An earned one is down 22% — three times as fast. By day 45 the two lines meet around −40%, so they end up in the same hole. But the earned card digs it in a few brutal days; the pulled card takes six weeks.

That gap is the whole reason I kept them apart. Remember Corbin Burnes? The 94 Breakout was an XP-path card — he debuted around 450,000 and was worth 5,500 three weeks later. A 99% wipeout, and it had nothing to do with the pack market: it was a million people finishing the 3rd Inning program and listing him the same week. Bryce Harper's 94 2nd Half Heroes did the same thing, −96%. Every one of the ugliest craters in my data is an earned card.

So the rule bends depending on what you're holding. With a pull-only card you've got about a week to read the room. With an earned card, if you're not out by day two, you're already late.

The peak comes early — or it doesn't come at all

If the decay curve is the "what," peak timing is the "when." I checked, for every pull-only diamond, the day it hit its lifetime-high price:

When pull-only diamonds hit their lifetime high. Most peak on day zero; the small day-14+ group is almost all upgrade cards.

58% peak on debut day. 85% peak inside the first week. If you're holding a diamond for its price and it's more than a week old, the odds its best day is still ahead of it are about one in seven — and that one-in-seven is almost entirely cards that get an upgrade catalyst. For everything else, the peak is behind you the moment the hype cools.

The four moments to sell

If you only think about selling at four moments, you'll beat every grinder who sells reactively.

Moment 1 — Drop day, first 24 hours

If you pulled the card, you have about a day before supply catches demand. Most diamonds never beat their debut-week price. Selling here captures the ceiling. Selling later almost always means selling for less.

The exception is a card the market hasn't priced yet — a sleeper quirk combo, a position-scarcity card that takes a few days to be recognized. Those rise for 48–72 hours before they decay. Reading which is which means reading the card, not the price.

Moment 2 — First-week peak

Miss the drop-day window and the next-best exit is usually inside the first week. Hype builds, creators post, the price climbs for a few days, then rolls over. Watch the 24-hour sparkline on the market table: when it flips red and stays red, the peak's behind you. (More on reading that signal below — because the popular version of it is wrong.)

Moment 3 — Before a Roster Update

This is the one people get backwards, so here's the data. I split every card at each of the season's 12 Roster Updates into "got upgraded" and "didn't," and measured the price move around the update:

Median 3-day price move before a Roster Update. The cards that get boosted run up ~99%. Everything else bleeds.

Cards that get upgraded run up a median +99% in the three days before the update lands — the market front-runs the boost. Cards that don't get upgraded do the opposite: they bleed ~5% as players cycle stubs into upgrade candidates. So: if you own a diamond that isn't on anyone's expected-upgrade list, sell two or three days before the update. You're racing everyone else doing the same math, but the alternative is holding through the update and watching the floor drop on a card that got nothing.

That +99% is also the upgrade exception in action — the one thing that reverses the decay curve. It's real, and it's why the peak-timing chart has a day-14 tail. It's also unpredictable, which is why "hold everything just in case" is not a strategy. You're betting against a curve that wins most of the time.

Moment 4 — End-of-cycle clearance

Every content cycle ends. Series sunset, themed packs leave, Live Series cards get power-crept. If you've held a diamond more than a month and it isn't a Live Series card you actually play with, sell it. This isn't about hitting a peak — it's about not dragging zombie inventory into the next cycle.

The signals to watch (and the one everybody gets wrong)

Four signals, all on the market table:

The market table, signed in. The Signal column reads Buy / Wait / Overpriced — Stanton is flagged Overpriced at the bottom — right beside the 24h sparkline. The sparkline and Buy Now floor are free; Signal and Value Gap are Pro.

• The 24-hour sparkline (free). Your fastest read on direction. Red and staying red is a warning.
• The Buy Now floor (free). If it's at or near an all-time low for the card, the market has decided. Don't fight it.
• The Value Gap (Pro). Flags cards the market is overpaying for upgrade potential. Own one that's gone negative and you're holding a known-overpriced position.
• The Signal (Pro): Buy / Wait / Overpriced — a card is "Overpriced" when it's trading 15%+ above its own 7-day average, "Buy" when it's 10%+ below.

Now, the thing everybody gets wrong. The folk wisdom is "two red days in a row means sell, five red days means you already should have." It sounds right. It's backwards. I tested it on every liquid pull-only diamond — over a thousand cases:

Chance a card is lower three days later. After a streak of down days it's less likely to keep falling, not more. Consecutive red days mean-revert.

On any given day, a card is lower three days later about 51% of the time — basically a coin flip. But after two consecutive down days, that drops to 41%. After three, still 41%. A streak of red days makes a card less likely to keep falling, not more. It's overshoot, and it tends to bounce.

That's why the streak isn't the signal — the Signal is the signal. It measures the card against its own 7-day average, which is what actually predicts reversion. A card that's "Overpriced" (well above its average) is the one to lighten up on. A string of red candles that's already dumped the card below its average? That's frequently a bounce, not an exit. Trust the deviation, not the streak.

When NOT to sell

The shorter list, on purpose:

• Cards you actively play. Performance beats profit. If the card wins you games, the stubs you'd recover don't.
• Cards on a credible upgrade path. Pre-update buys are sometimes correct even at a premium, if the boost is likely. (That's different from the pre-update sell in Moment 3, which is for cards getting nothing.)
• Cards below the Quick Sell floor. If quick-selling beats the market, that's not a sell decision — that's the market having given up before you did.

What flash sales do to all of this

There's one more force big enough to move the whole diamond market in a single day: the flash sale. When SDS drops discounted packs into the Show Shop for a few hours, players open them and dump the pulls — which lands directly on these pull-only cards. Supply floods, and prices fall off a cliff.

I can show you exactly when it happened, because SDS announces flash sales on their socials and I can match the dates to the market. April 21 was a confirmed flash sale:

The diamond market around the confirmed April 21 flash sale: −25% on the day, +18% the next, fully recovered in three. The same pattern hit May 13 (−30%) and May 26 (−27%).

The diamond market dropped 25% in a single day, bounced 18% the next, and had recovered within three. It wasn't a one-off: the May 13 flash sale knocked it −30% (and it snapped back +43% the day after), and May 26 took it −27%. Same fingerprint every time — a sharp dip and a fast recovery.

The play cuts both ways. Don't panic-sell into the dip; the recovery is usually days away. And if you've been eyeing a card, a flash-sale afternoon is the cheapest you'll see it. (Stub sales — when stubs themselves get discounted — are rarer and push the other direction: more buying power, chase cards tick up. I haven't caught one cleanly enough this season to put a number on it.)

What this all means

Most people sell a diamond for one of three reasons: they need stubs and grab whatever they can stand to lose; they notice it decayed and panic-sell at the floor; or they never sell, and quick-sell it later for a fraction of what they paid. All three are the wrong move.

The right move is to plan the sell on the buy day. Decide your window before you own the card. Treat every Diamond like a fixed-period position, because that's what the data says it is.

You won't nail every sell. Nobody does. The point isn't precision — it's not being the person still holding the card at end-of-cycle because they never decided when they were getting out.

Where to go next

The buying side is the mirror image: the best time to buy is the floor of the decay curve, before any upgrade catalyst — or an hour into a flash sale. If you'd rather make stubs without timing the market at all, parallel grinding is the other lever, and OpScore tells you which cards are worth holding for reasons that have nothing to do with price.

The intraday question — which hour to click sell — needs our hourly price data to mature before I'll put real numbers on it. That post comes when the data does, not before.

Every number here comes from DiamondOps market data: 219,122 daily price records, March–July 2026. The per-card decay math is restricted to pull-only diamonds (earned cards excluded on purpose). Drafted and crunched with editorial AI on my data; the takes — and the being-wrong-about-red-days — are mine.

I'm Shaun. If something's off, the comments are open.