Marriott currently has a live Bonvoy Moments auction in Prague that is unusually revealing, not just because of what’s on offer, but because of how the auction itself is behaving.

This is not a hypothetical case study. It is running right now. You can open it in another tab and watch the numbers move.

At the time of writing (early February 2026), with roughly ten days remaining before close, the auction is already doing something structurally interesting.

The package is titled:

Prague Unlocked + Private Heritage Immersion

15–17 May 2026

At its core is a private visit to the Musoleum of David Černý, led by the artist himself.

Černý is not a ceremonial name attached for marketing effect. He is one of the most provocative contemporary artists in Central Europe, known internationally for works like the rotating Kafka Head and Entropa. His Musoleum is not open to the public. This is not something a concierge can arrange.

Around that anchor, Marriott has built a tightly sequenced cultural program:

The structure is deliberate: Prague as a dialogue between disruption and tradition, rather than postcard beauty alone. It is exactly the kind of Moment that attracts serious interest.

Marriott listed two identical packages for this experience. The dates, hotel, access, and itinerary are the same. They are auctioned separately, but otherwise identical.

With about ten days remaining, Package 1 sits at 182,500 points with 39 bids, while Package 2 sits at 152,500 points with 32 bids.

Update as of February 6, 2026: Bids current at time of writing; refresh Moments for latest.

The difference isn’t about the experience at all; it reflects how information is revealed independently in each auction.

Marriott’s Moments platform logs both manual bids and automatic bids triggered by max bids. When two members have competing max bids, the system advances the auction in fixed 2,500-point increments until one bidder’s maximum is reached.

That behavior is quite clear here.

On January 25, around 8:10 a.m. CST, two bidders (masked as ******49 and ******57) ran into each other:

The bid log shows a rapid, perfectly spaced sequence:

150,000 → 152,500 → 155,000 → 157,500 → 160,000 → 162,500 → 165,000 → 167,500 → 180,000 → 182,500

All within seconds.

This was not organic, sequential bidding. It was the system in receipt of two high max bids, and working through the appropriate sequence to knock out the lower of the two.

Once that happened, Package 1’s visible price was effectively decided. The auction did not “discover” that the experience was worth 182,500 points. It simply revealed what the second-highest bidder was willing to pay.

Package 2 shows the same mechanism at work, but resolving at a different point.

Here, two competing max bids emerged in the 110,000 to 132,500 point range. The system went to work in the same way as for Package 1, fully clearing out one bidder. The identical timestamps on the January 24 prints suggest that a higher max bid was introduced, forcing the system to advance through the existing bids in order to make that displacement explicit.

The result is the same early disclosure of two bidders’ willingness to pay, but occurring at a lower level and unfolding more gradually than in Package 1.

Looking at masked bidder IDs across both auctions is revealing.

Several mid-tier bidders appear in both Package 1 and Package 2. These bidders entered early, escalated into roughly the 80k–110k range, and then stopped once prices ran away. That pattern is consistent with disciplined bidding.

What does not appear is also important:

This does not mean they are unaware. It almost certainly means they are waiting.

A rational bidder who has already revealed a ~180k ceiling has every incentive to delay placing a max bid in Package 2, precisely to avoid triggering the same public stair-step again.

Nothing about this outcome requires assumptions about emotional bidding or miscalculation. It follows from how the Moments auction is set up and how information becomes visible over time.

Max bids are not sealed in practice. When two bids collide, the system advances the price publicly until one bidder is cleared out, making that willingness to pay visible to everyone else. Because there is no bid extension, that information lands early and cannot be revised or tested by later entrants.

Each package is also auctioned on its own. There is no mechanism that links the two or allows prices to converge without bidders revealing themselves again. Once a high bid is disclosed in one auction, that information stays local to that auction, even if an identical package is running alongside it.

The result is not mispricing so much as path dependence. Once an early confrontation resolves, the auction has already shown more of its demand curve than some bidders are prepared to engage with, and participation thins accordingly.

I have bid on Moments before, and I have also walked away from many. What makes this one interesting is not whether it is “worth it,” but that you can watch the pricing mechanism fail in real time.

This Prague auction shows how loyalty auctions can drift away from rational pricing without anyone behaving badly. The system is simply very good at exposing the second-highest willingness to pay, and very bad at protecting price discovery once that happens.

Once two high max bids meet, the system resolves them publicly in tight, evenly spaced 2500-point jumps. At that point the platform has already disclosed something concrete: how far the second-highest bidder was willing to go for that package. One participant’s valuation has been forced into the open.

Before this, bidders are operating without a reference point. After it, anyone entering the auction is no longer bidding into a blank range. They are deciding whether they are comfortable bidding above a level that has already been tested and rejected for someone else.

That does not make further bidding irrational or mistaken. It simply changes the decision being made.

A useful way to read the bid ladder once this happens is to watch the timing and spacing. Near-instant bids that move in perfectly regular increments are not emotional escalation. They are the system stepping through competing max bids that were already in place.

By the time those jumps appear, the auction has already answered one important question. What remains is not price discovery in the dark, but individual judgment, made in the knowledge that at least one other participant was willing to go higher than what has already been revealed.

This Prague Moment will likely move further before it closes. Package 2 may step up if a higher max bid enters late (either the earlier second-high bidder from Package 1, or someone else entirely), or it may finish where it is.

Either outcome is instructive.

The interest here is not just the destination or the final price. It is the way scarcity, public max bids, and independently run auctions interact while the process is still unfolding.

That interaction is visible now, and it will be the same in future Moments long after this one has closed.

If you want to follow along, both packages are live on Marriott Bonvoy Moments now. Watching the bid ladders often tells you more than the headline price does.