Our July scorecard: every price we published, and how it resolved

On July 21 we published every price on the Ravioli board with a timestamp, so they could be checked later instead of quietly forgotten. Sixteen markets, sixteen numbers, one date. This is what happened to all of them.

Most forecasting write-ups appear after the fact, when the author already knows the answer. That is not a record, it is a highlight reel. The only version worth reading is the one written before the outcome and scored afterwards, misses included. So here is ours, two weeks on.

First, an honest note about what these numbers are

Ravioli is early. On July 21 most of the markets in that table had no trading in them at all, which means the price you saw was the opening line the market was seeded with, not a crowd of people disagreeing with each other in public. Two of the sixteen had real trades behind them. The rest were the board's opening estimate.

That distinction matters, so we are going to keep making it. When a price has traders behind it, we will say so. When it does not, we will call it the opening line. A number is only as good as the process that produced it, and pretending otherwise is how forecasting records get worthless.

The three that resolved

| Market | Board said | Traded? | Actual | Result | |---|---|---|---|---| | Boeing resumes 737 MAX airworthiness certificates by Jul 24 | 75% yes | Yes | Yes | Called it | | NYC Air Quality Index hits 'unhealthy' by Jul 30 | 70% yes | Yes | Yes | Called it | | BTS 'Normal' passes 50M YouTube views by Jul 31 | 72% yes | No | No | Missed it |

Two right, one wrong. The two that came in were the two with actual trading behind them. The one that missed was an opening line nobody had touched.

That is a sample of three and we are not going to pretend it proves anything. It is worth noticing anyway, because it is the whole argument for prediction markets in miniature: a number gets better when someone has a reason to correct it. We wrote about why that works when the site launched, and this is the first time our own board has said anything about it either way.

The miss, in full

BTS's 'Normal' video was at 72% yes to pass 50 million YouTube views by July 31. It did not get there.

Seventy-two percent is not a coin flip. It is a real opinion, stated in public, with a deadline attached, and it was wrong. It was also the second most confident call in the whole table, which is the uncomfortable part. Confidence and accuracy are different things and this is what the difference looks like.

The three we cannot score

Three markets from the July 21 table closed on schedule and are still sitting without a recorded outcome:

• Randy Orton appearing at WWE's Saturday Night's Main Event, July 25 — 70% yes
• Jazz Chisholm Jr. to homer in the Yankees' July 26 game — 42% yes
• Canadian wildfire smoke alerts persisting past August 1 — 55% yes

We are not going to score these from memory. Two of the three we would probably have got right, and saying so without a recorded resolution is exactly the kind of shortcut this post exists to refuse.

This is a gap on our end. A market that closes and never resolves is a broken promise, and the fix is ours to make, not yours to work around. Resolution is the part of a prediction market that has to be boring and reliable, and ours is not there yet.

Still open, for anyone who wants to check us later

| Market | Board said | Closes | |---|---|---| | Phillies finish 2026 with a better record than the Mets | 78% yes | Sep 28 | | Madonna 'Confessions II' holds the Billboard 200 top 10 | 70% yes | Aug 5 | | Tyler Kolek in an official Knicks promotion | 70% yes | Aug 15 | | CDC declares the Taco Bell cyclospora outbreak over | 70% yes | Aug 15 | | Netflix announces InterPositive AI integration | 70% yes | Aug 31 | | Generative AI publicly named as replacing a major role | 58% yes | Aug 31 | | Andrew and Tristan Tate face formal US charges | 58% yes | Aug 15 | | Abdul El-Sayed wins Michigan's Democratic Senate nomination | 42% yes | Aug 4 | | Péter Magyar's bill to replace Hungary's constitution | 40% yes | Sep 30 |

Prices as published on July 21, 2026. Every one of these is still live on the board, and every one of them will appear in a future scorecard with an outcome next to it.

Why we do this

Every platform in this category will tell you its numbers are sharp. Almost none of them publish a list of the times the numbers were wrong, at the same size, on the same page. That is not because the misses do not exist. It is because a highlight reel converts better.

We would rather be checkable. From this week on, the scorecard runs weekly: what we said, what happened, and what we could not score. If you think the board is wrong about something in the open table above, the useful response is not to argue with us. It is to trade against it. It costs you nothing, and if you are right, the price moves and everyone reading it gets a better number.

Start calling them yourself

Ravioli is the free-to-play prediction market. You trade with Sense, the free in-app currency, on real questions in sports, politics, culture and tech. No deposit, no ID verification, nothing at risk, and real prizes for the people who call it right.

Create a free account or browse the board first and see whether you disagree with any of the prices above.