I Built an Auction Game That Exposed My Terrible Price Instincts
I looked at a polished ammonite, admired its perfect internal spiral, and confidently bid $25,000.
The hammer price was $6,800.
A red SOLD stamp landed across the image, my score came back as 43, and the app helpfully informed me that I was off by 3.7×. That tiny moment — commit to a number, then discover how detached from reality it is — is the entire pleasure of Going Once, a mobile auction guessing game I built in Eazo.
Each round presents an unusual object with one compact catalogue note. I enter what I think it sold for, place my bid, and get the hammer price immediately. At the end, the app turns the whole session into a sale book showing every lot, every guess, every actual price, and the points I somehow managed to keep.
My final score was 43 out of 700. So, yes, the visual design went better than the auction strategy.
Step 1: Start with one object and one uncomfortable guess
The earliest useful version didn’t need a whole auction house. It needed one object, one number field, and a reveal.
I started with a prompt like this:
Build a mobile-first auction price guessing game called “Going Once.” Show one collectible object at a time with its name, category, image, and a one-sentence catalogue description. Let the player enter a dollar bid and submit it. After submission, reveal the object’s hammer price, compare it with the bid, award up to 100 points based on closeness, and offer a button to continue to the next lot.
That produced the core loop quickly enough to test in the Eazo preview: look, estimate, bid, reveal. It worked, but the first version behaved too much like a calculator. I entered a number and received another number. Technically correct. Emotionally flat.
An auction needs a point of no return. The player should feel the bid leave their hands before the answer arrives.
Step 2: Give the hammer price an entrance
The finished reveal does several jobs in one screen. A red SOLD stamp appears diagonally over the lot image. The bid interface disappears. In its place, three columns compare YOUR BID, HAMMER PRICE, and SCORE, followed by a plain-language ratio such as “Off by 3.7×.” Only then does the wide NEXT LOT button appear.
That hierarchy mattered more than adding a complicated animation. The stamp supplies the drama; the side-by-side numbers explain the result; the multiplier delivers the joke.
For the second pass, I asked for the reveal itself:
Make submitting a bid feel like an auction result rather than a form response. Lock the bid as soon as the player submits it, pause briefly, and place a large angled red “SOLD” stamp over the object image. Replace the input with a three-column result row for the player’s bid, the hammer price, and the score. Under that, state how many times higher or lower the guess was. Keep the next-lot action hidden until this reveal is complete.
This was the moment the game found its rhythm. During my run, I overvalued a folding fan, a glass-key typewriter, and a carved jade dragon by increasingly heroic amounts. The app didn’t need a sarcastic narrator. Seeing $30M beside $310K, plus a clean “Off by 97×,” was enough.
There was still a presentation problem, though. A generic card layout made the objects feel like items in a shopping app. They needed to look as if they had already been catalogued, numbered, and laid out for inspection.
Step 3: Make the phone feel like an auction catalogue
The visual direction came from printed ephemera rather than glossy commerce UI. The live app sits on a near-black background, framing a narrow cream sheet with a double ochre border. Fine rules divide the masthead, lot number, category, image, description, and controls. Serif type carries the object names; condensed uppercase labels handle the mechanics. A faint grid runs behind everything like old ledger paper.
Get I’m Not a Cat’s stories in your inbox
Join Medium for free to get updates from this writer.
It feels considered, but not precious. The black BID button is blunt. Categories such as NATURAL HISTORY, TIMEPIECE, and SCIENTIFIC INSTRUMENT appear in red, as if marked by the auction staff. The object illustration gets most of the screen because that is where the guess begins.
My art-direction prompt was more constrained than “make it vintage”:
Restyle the game as a compact printed auction catalogue viewed on a phone. Use warm ivory paper, a subtle ledger grid, black ink, aged-gold double borders, and restrained red accents. Pair an elegant editorial serif with small uppercase utility labels. Give the object image most of the page and separate each section with fine rules. Avoid gradients, glass effects, modern marketplace cards, and decorative clutter. It should feel archival and tactile while remaining easy to tap at 390px wide.
Because I could keep the app open beside the Eazo conversation, the follow-up was concrete: make the image taller, keep the price field visible without scrolling, and reserve the red accent for categories, hammer prices, scores, and the sold stamp. Those adjustments made the hierarchy much clearer than another broad style request would have.
The app also supports English and Chinese from a compact switch in the masthead. I tested the result screen in both languages; the score, bids, prices, controls, and full lot recap all changed together rather than translating only the headline.
Step 4: Turn a stack of guesses into a sale book
A single price reveal is amusing. A sequence becomes a game only when each mistake follows you.
The score in the top-right corner accumulates from lot to lot, while the lot counter makes the remaining distance visible. My session moved through seven objects, ending with a cut meteorite fragment. When the last hammer price appeared, NEXT LOT changed to SEE RESULTS.
I used one final prompt to give that ending some weight:
Add a complete auction-session structure. Randomize a small set of lots for each replay, show “Lot X / total” and a running points total in the masthead, and preserve every bid and result. After the final lot, replace the next button with “See results.” Build the results screen as a “Going, going… gone” sale book with the total score, Replay and Share actions, and a numbered table of every lot showing bid, actual price, and points. Replay should reset the score and begin a newly selected sale.
The sale book changed how the earlier rounds felt. I was no longer making isolated guesses; I was building a visible record of bad judgment. My one respectable ammonite score sat above six zeros, each row preserving the exact size of the miss. Brutal, but neatly typeset.
I also replayed the app after finishing. The score returned to zero and a different opening lot appeared, which gives the experience a reason to survive beyond the first result. You can learn the rough price range of one set of curiosities, press Replay, and immediately discover that a wax-flower cloche creates an entirely new crisis of confidence.
The best feedback was just four words and a multiplier
The build became more interesting each time the app got more specific.
The first pass established a functional price guess. The second made the reveal irreversible. The catalogue treatment gave the objects authority. The sale book turned individual surprises into a complete performance.
None of those steps required a giant all-at-once prompt. I described the interaction in Eazo, tried the working version, noticed where the tension disappeared, and made the next request about that exact moment. By the time the session flow was finished, the app was already live and shareable; there was no separate deployment chapter between “this feels right” and “try my terrible bids.”
You can play Going Once and find out whether your auction instincts are any better than mine. I built this game in Eazo, and if you have a collection or game idea of your own, you can use Eazo to turn it into something other people can play. Start with sneakers, film props, vintage cameras, or design chairs and see if your friends can survive the hammer price reveal.
