Hot Deals, Four Months Later: Does a Long Streak on the List Mean a Better Buy?
In July we published our first look at the Hot Deals list: 30 days of data, and an experiment in which someone bought every card on the list every day. Its first tip for spotting the good ones was "persistence beats novelty". The biggest gainers kept coming back to the list as they climbed, so a card that stays on the list for a week looked like a real supply squeeze, and a card that appears once looked like a blip.
We liked that idea enough to build it into the site. Every card on Short Term deals now shows a Top 50 Streak: how many of the last 30 and 60 days it spent in the daily Top 50. Four months later we have enough history to test it properly. The answer is not the one we expected, and it has changed what that page shows.
The short version
- The list itself works. Buying everything on it returned +17.4% after 30 days. The rest of the market returned +1.7% over the same dates. A card on the list was six times as likely to gain 50% in a month as a card that was not.
- The streak predicts nothing. Cards on their first day and cards on their twentieth did about equally well. If anything, long streaks fell slightly more often.
- Our July tip was hindsight. It came from looking back at cards after the fact, and a card only builds a long streak if its price keeps rising. Measured on the day you could actually have bought, the effect disappears.
- What does predict the outcome is the price. For cards over €5, the ones still at the price they had when they first appeared on the list returned +20% on average. The ones already up 50% or more returned about 2%, and half of them fell.
- So the site now shows it. Every card on Short Term deals shows its price on the day it joined the list and how far it has moved since. Cards over €5 also get a Better, Typical or Worse odds label. We tested a percentage instead, and it promised too much, so we show the label.
How we tested it
The site saves every card's Short-Term score every day. Since 27 April those scores have been written live by the daily job, never backfilled, so we can rebuild the exact Top 50 the site showed each day. That is the same ranking the streak counter uses.
Every card on every daily list counts as one purchase at that day's cheapest NM English listing on CardMarket. We then look at the cheapest listing 30 days later. Purchase days run from 27 May to 14 September: the first gives every card a full 30 days of streak history, and the last leaves 30 days to measure the result. That gives 4,702 purchases across 467 different cards.
The most important rule: the streak is counted only up to the day of purchase. A card on its fourth day on the list has a streak of four, even if it goes on to stay there for a month. That is what you see on the page when you decide, so that is what we test.
For comparison, we took every other card priced at €0.50 or more, bought on every Monday in the same window and held for the same 30 days. That gives 17,520 control purchases.
One more measure needs a word. The Hot Deals list is full of cards under €1, and when one of them goes from €0.30 to €3.89 it can move a basket total on its own. So next to the basket return we report a capped average, in which no single purchase can count for more than +100%. Where the capped average and the basket disagree, a few lottery tickets are doing the work.
The list works, again
| 30 days after buying | Purchases | Basket return | Capped average | Went up | Went down | Gained 50%+ |
|---|---|---|---|---|---|---|
| On the Top 50 | 4,702 | +17.4% | +16.6% | 43% | 23% | 18% |
| Every other card | 17,520 | +1.7% | +0.2% | 27% | 38% | 3% |
Before getting to the streak, this is worth stating plainly, because it is good news for the list. Over four months, including the falling market of August and September, the cards it flagged did much better than the cards it did not. They were more likely to go up, less likely to go down, and six times as likely to gain 50% within a month.
The median purchase was still flat, as it was in July. Most cards on the list do nothing in 30 days. The return comes from a minority that move a lot.
The clearest result is in the €5–50 range, where most collectors actually buy:
| €5–50 cards, bought in | Top 50: basket | Top 50: went up | Other cards: basket | Other cards: went up |
|---|---|---|---|---|
| June | +29.3% | 66% | +11.7% | 48% |
| July | +13.2% | 55% | −4.2% | 27% |
| August | +9.2% | 50% | −6.1% | 18% |
| 1–14 September | +8.7% | 38% | −4.7% | 21% |
| Whole window | +15.6% | 54% | −0.1% | 30% |
The list beat the rest of the market in every month. In July, August and September, when typical €5–50 cards lost 4–6%, the ones on the list still gained 9–13%. The median €5–50 purchase from the list gained +8.9%, against 0.0% for everything else. The margin has narrowed since June, but it never went away.
The streak test
Now the question we set out to answer. Here are the same 4,702 purchases, grouped by how many of the previous 30 days the card had already spent on the list:
| Streak on the day you bought | Purchases | Cards | Basket return | Capped average | Went up | Went down | Gained 50%+ |
|---|---|---|---|---|---|---|---|
| First day on the list | 446 | 403 | +12.9% | +17.6% | 48% | 20% | 18% |
| 2–4 days | 1,071 | 373 | +21.7% | +16.0% | 45% | 21% | 16% |
| 5–9 days | 1,222 | 273 | +19.0% | +16.3% | 44% | 23% | 18% |
| 10–19 days | 1,337 | 159 | +9.6% | +16.2% | 39% | 23% | 18% |
| 20–30 days | 626 | 50 | +18.6% | +18.1% | 40% | 28% | 23% |
The capped average is between +16% and +18% in every row. The basket return jumps around with no direction. The chance of a 50% gain barely moves. The only steady trend runs the wrong way: the longer a card had been on the list, the less often it went up (48% → 40%) and the more often it went down (20% → 28%).
We checked this with a 14-day holding period instead of 30, and without EB02. The capped average stayed flat both times. Counting unbroken runs of days in a row gives a mixed result: cards on day 8 or later of a run had a higher capped average (+24%) but a lower basket return (+11.5%), which is a few cheap cards jumping rather than a reliable advantage. On no measure was a card on day twenty clearly a better buy than a card on day one.
Why it looked true in July
So where did "persistence beats novelty" come from? Here is the same data, sorted the way the July article effectively sorted it: each card's first purchase, grouped by how many days it ended up spending on the list.
| Total days on the list (known only afterwards) | Cards | Basket return from first day | Capped average | Gained 50%+ |
|---|---|---|---|---|
| Once | 65 | −0.1% | +7.9% | 11% |
| 2–4 days | 125 | +20.3% | +16.3% | 12% |
| 5–9 days | 109 | +30.5% | +14.6% | 15% |
| 10–19 days | 101 | +56.6% | +27.9% | 33% |
| 20–30 days | 67 | +38.5% | +23.0% | 27% |
This is the pattern we saw in July. Cards that went on to stay on the list for weeks had great returns from their first appearance, and cards that appeared once went nowhere. It is real, and it is no use to a buyer, because you only learn how long a card will stay on the list after it has stayed there.
The reason is how the list works. The Short-Term score is high when a card's offer book is thin and its copies are selling. A card whose copies keep selling keeps a high score and stays on the list. A card whose squeeze fades drops off and ends up with a short streak, whatever happens to its price afterwards. So the long stay does not cause the rise. The rise causes the long stay. Looking back, every big winner has a long streak. Looking forward, a long streak tells you very little.
What actually predicts the result: the price since it arrived
If the streak is not the signal, what is? For cards over €5, the clearest answer is how much the price has already moved since the card first appeared on the list:
| Cards over €5: price since first day on the list | Purchases | Cards | Capped average | Median | Went up | Went down | Fell 30%+ |
|---|---|---|---|---|---|---|---|
| First day on the list | 218 | 198 | +21.4% | +10.5% | 61% | 23% | 4% |
| Still at that price, or lower | 861 | 160 | +19.8% | +11.1% | 59% | 16% | 3% |
| Up 2–20% | 156 | 46 | +15.4% | +7.1% | 51% | 37% | 3% |
| Up 20–50% | 233 | 55 | +7.3% | +0.1% | 49% | 37% | 12% |
| Up more than 50% | 249 | 50 | +1.8% | −0.1% | 37% | 49% | 17% |
This is a clean slope. A card that has been on the list for a while without its price moving was as good a buy as a card on its first day: 59% went up and only 16% went down. A card that had already risen more than 50% while on the list was roughly a coin flip with a bad tail. Half of those purchases lost money, and one in six lost 30% or more within a month.
It also holds in both halves of the window. From late May to mid-July, the "still at the same price" group averaged +34% and the "up 50%+" group +1%. From mid-July to mid-September the figures were +15% and +3%. It is not one rally, and it is not one set. Taking out the ST01 starter deck, which had a spectacular June, changes nothing.
Under €5 the pattern does not hold. Cheap cards rose on roughly a third of purchases whether they were on their first day, still at their starting price or already up 50%. The only group that stood out was cheap cards that had dropped more than 20% since joining the list. About half of those rose, but a third fell, so they swing both ways rather than being a better buy. At these prices one sale moves the cheapest listing by tens of percent, and that noise drowns out the signal.
So the Top 50 Streak is useful, but not in the way we described it. A long streak on a card whose price hasn't moved means the squeeze is real and still unpriced. A long streak on a card that has already doubled means you are looking at the result, not the opportunity. The streak only means something next to the price change, and that is what the site now shows.
Three cards that show the difference
- Shanks OP13-028 (alt art). Joined the list on 23 June at €8, and its cheapest copy then sat at €8 on all but one of the next 18 days while the card stayed on the list. Every one of those purchases was worth €14.50–18.95 a month later, roughly +100%. On 11 July the €8 copy finally sold, the price stepped to €12.10, and the card climbed to #2 on the list. Buying at #2 returned about +24%, a quarter of what the quiet weeks paid.
- Roronoa Zoro ST01-013. Joined the list on 5 June at €89.99, during the starter-deck rally from our June report. Buying in its first three days returned +33–67%. By 11 June the cheapest copy was €324.95, and it stayed on the list at €190 until 16 June. Every purchase from that week was worth €135–140 a month later, a loss of 26–57%.
- Sabo OP13-004 (Leader alt art). Joined on 17 August at €18.99, jumped to €35 on 22 August, and fell back to €18.83 three days later. It then stayed on the list at €18.83 for another three weeks, finishing with a 23-day streak, and was worth €17.69 a month later. The longest streak in this group came with no gain at all.
What changes on Short Term deals
Next to the Top 50 Streak, every card on Short Term deals now shows "Since €…": its cheapest NM English price on its first day in the Top 50 during the last 30 days, and how far it has moved since. A card showing "Since €7.99 +75%" has already had most of its run. A card showing "Since €10.00 +0%" has not moved yet.
The obvious next step was to turn the table above into odds: "62% of cards like this rose within a month". We built that, then checked it the only fair way, by replaying it. For every list day from late June to mid-September, we rebuilt the odds the site would have shown that day using only results already known by then, and compared them with what actually happened 30 days later. That is 3,317 replayed predictions.
The percentages promised too much. For cards over €5 they said, on average, that 71% would rise. 49% did. A card's result is only known a month after it is on the list, so the odds always describe the market of one to three months earlier. Those were the rising months of June and July, and they were applied to the falling months of August and September. A longer or shorter look-back did not fix it. This lag is built in, and more data will not remove it.
The ranking held up. Cards the odds rated higher really did rise more often than cards they rated lower, in both halves of the replay. So the site shows the comparison and not the percentage. Each card over €5 gets one of three labels, based on how similar past cards did against the average card on the list:
| Replay, cards over €5: label shown | Predictions | Cards | Rose within 30 days | Fell within 30 days |
|---|---|---|---|---|
| Better odds | 670 | 117 | 51% | 20% |
| Typical odds | 482 | 95 | 50% | 30% |
| Worse odds | 93 | 25 | 24% | 53% |
"Better" and "Typical" rose about equally often, but "Better" fell much less often. "Worse" is the label to take seriously: in the replay those purchases fell twice as often as they rose. It mostly goes to cards already up more than 50% since joining the list.
Cards under €5 get no label, for the reason in the previous section. When we replayed the odds on cheap cards, they ranked them worse than chance. The "Since €…" price is still shown, because it is a fact rather than a forecast.
The labels are rebuilt every day from every past list purchase whose result is known, and we will rerun the replay regularly. If the ranking stops holding, the labels come off.
Rank on the list barely matters either
A related question: is #1 a better buy than #45? Only slightly. Cards ranked #1–10 had a capped average of +19.1% and a 22% chance of a 50% gain, against +15.6% and 17% for #26–50. The basket return pointed the other way, because the lower half of the list held most of the expensive cards that did well. Treat the whole Top 50 as one shortlist, not a ranking to buy from the top down.
The cheap end is still a lottery
| Price when bought | Purchases | Basket return | Capped average | Median | Went up | Other cards: went up |
|---|---|---|---|---|---|---|
| Under €1 | 1,265 | +113% | +16.1% | 0.0% | 31% | 13% |
| €1–5 | 1,720 | +39.2% | +18.2% | 0.0% | 40% | 21% |
| €5–50 | 1,496 | +15.6% | +15.6% | +8.9% | 54% | 30% |
| €50+ | 221 | +16.0% | +13.3% | +5.3% | 52% | 34% |
The +113% basket for cards under €1 is a handful of cards like the Pirates Docking Six OP15-088, whose cheapest copy sat at €0.30 for three weeks and then stepped to €3.89 when it sold. The capped average is the same as every other price range, the median is zero, and only 31% went up. And as our volatility study put it, percent is not money. Doubling a €0.50 card earns you €0.50, which is less than the postage to sell it. Under €5, the list finds real squeezes, but you cannot make a living on them after fees.
€5–50 is still where the list is most reliable, as it was in July. More than half of purchases rose, the median gained 8.9%, and the advantage over the rest of the market held in every month.
What to actually do with this
- Use the list as a shortlist. Being on it at all is the signal: 18% of purchases gained 50% in a month, against 3% of everything else.
- Don't wait for a streak to "confirm" a card. Day one was as good as day twenty. Waiting only gives the price time to move.
- Check "Since €…" before you buy. Still at the price it joined the list at: the squeeze has not been priced in. Up 50% or more: you are probably late.
- Treat "Worse odds" as a warning, and the other labels as history. A label compares a card with the rest of the list, not with your bank balance. Even "Better odds" cards fell on about one purchase in five in the replay, and in a falling month the whole list does worse.
- Focus on €5–50. That is where the list beat the market most consistently and where the gains are large enough to survive fees and shipping.
- Don't buy from the top down. Cards ranked #1–10 did only slightly better than #26–50.
The method behind the Short-Term score is in How OPMT Scores Work, and today's list is on Short Term deals.
Methodology: the daily Top 50 is rebuilt from OPMT's stored Short-Term scores, ranked by score, which is the same ranking the Top 50 Streak counter uses. Scores from 27 April 2026 onward were written live by the daily job; nothing used here was backfilled. Purchases are every card on every daily list from 27 May to 14 September 2026 (4,702 purchases, 467 cards), at that day's cheapest listed NM English copy on CardMarket. The exit is the cheapest listing 30 days later (the latest price in the seven days up to that date). A version that exits on a 7-day average instead gives the same results to within a point. The streak is the number of days in the Top 50 during the 30 days up to and including the purchase day. "Price since first day" compares the purchase price with the price on the card's first list day within those 30 days. The control group is every other card priced at €0.50 or more on each Monday in the window that was not on that day's list (17,520 purchases). Prices at or below €0.10 (the bulk floor) or above €50,000 are dropped, and one known fake listing (a €49 poster sold as Sabo's Wanted Poster OP13-120) is excluded. The capped average limits each purchase to +100%. "Went up / down" means a move of more than 2%. The same card appears on many days, so purchases are not independent; the long-streak groups contain few distinct cards (50 in the 20–30 day group), and the real uncertainty is wider than the purchase counts suggest. Excluding EB02 changes no conclusion (it was 71 of the 4,702 purchases). The odds replay covers list days from 27 June to 15 September 2026. On each day the rates were rebuilt only from purchases whose 30-day result was already known that day, grouped by price (under €5 or €5 and over) and by the change since joining the list (first day, down more than 20%, flat or down, up 2–20%, up 20–50%, up more than 50%). Groups with fewer than 80 past purchases used their price group's overall rate. A label is "Better" or "Worse" when the group's rise rate is at least 5 points above or below its price group's. The live labels use the same filters as the rest of this article, including the fake-listing exclusion. Excludes fees and shipping. Historical analysis, not financial advice.
Read next
The One Piece TCG Market in September 2026 — Monthly Report
The index fell −6.4% in September, but the typical card fell −2.9%. OP16 stopped falling, OP17 singles slid 11% after launch week, and last month's biggest "gainers" turned out to be August's dips coming back.
Which Rarities Let You Sleep at Night? Volatility, Drawdowns and Growth by Tier
Five months of daily prices across 914 cards. SPs almost never drop 30% overnight and every drop recovered; SECs drop twice as often and stay down. What each rarity tier costs you in sleep.
What Is Actually in Each Price Bracket — and the Order the Summer Rally Moved Through Them
Price brackets are rarity ladders with a Luffy tax on every rung. Set EB02 aside and the 2026 rally has a clear order: Mangas first, then SSPs, then SPs, and alt arts last — just in time to fall.