How it actually works

Stop Guessing.
Start Winning.

You've been burned before. Full price on a game that dropped 60% a month later. Pricetradamus exists so that never happens again. Here's how the Oracle actually makes its calls.

How confident is the Oracle?

Every prediction gets a confidence score from 0 to 100. It starts in the middle and moves based on what the data actually shows. A game that goes on sale like clockwork every few months scores higher than one with no history. A game sitting at its all-time low right now? That's a different story entirely.

Base Score
Every game starts somewhere in the middle. No assumptions.
neutral
Overdue for Sale
Been a long time since this game went on sale and the pattern says it should have by now
strong boost
Nearly Due
Sale is coming up but not quite yet based on historical timing
solid boost
Recently on Sale
Just came off a discount. Less likely to drop again right away. Does not apply when the game is currently on sale.
small hit
No Sale History
Never been discounted. Nothing to work from.
big hit
Sale Window Imminent
A major sale event like the Steam Summer Sale or Black Friday is right around the corner
major boost
Sale Window Approaching
A big sale event is coming up in the next couple of months
good boost
Sale Window Distant
There is a sale on the horizon but it is still a few months out
small boost
Mature Title
Game has been out for years. Its discount patterns are well established at this point.
solid boost
Seasoned Title
A year or two old. Starting to have enough history to read.
small boost
New Release
Just came out. No pattern to learn from yet. Check back in a few months.
solid hit
Rich Price History
Lots of sale events on record. The Oracle can spot real patterns here.
solid boost
Single Price Point
Only ever seen at one price. No way to assess discount potential.
major hit
Sparse Data
Not enough price points to build a reliable picture yet
moderate hit
At Historical Low
At or near the lowest price it has ever been and currently on sale. This flips the verdict to BUY NOW and locks in a high confidence floor regardless of other signals.
score locked high

Three possible verdicts

After scoring, the Oracle checks two things: is the price near its all-time low, and is it currently on sale? Those two answers decide what you see.

🔮 BUY NOW

Buy Now

The price is at or near its all-time low and it is currently on sale. It is not going lower. Stop waiting and buy it.

At or near historical low price
Currently discounted
High confidence floor applied regardless of other signals
⏳ WAIT

Wait

This game has been cheaper before and probably will be again. You get a predicted price and a rough timeline based on sale history and upcoming events.

Price has room to drop further
Predicted price is based on historical lows
Predicted date is tied to the next known sale window
UNKNOWN

Unknown

Not enough data to say anything useful. Usually a new release or a game that has never gone on sale. Come back once it has some history.

Never been on sale or very limited price history
No pattern to work from yet
Check back after the first sale hits

Where the data comes from

The predictions are only as good as the data behind them. Here is exactly where it comes from.

IsThereAnyDeal (ITAD)

The backbone of the price history. ITAD has been tracking game prices across Steam, PSN, Xbox, and Nintendo since 2012. Millions of data points covering sale events, discount depths, and exact timing.

IGDB (Internet Game Database)

Game metadata. Titles, cover art, release dates, publishers. Maintained by Twitch and the most complete game database available over an API.

Sale Window Calendar

A curated calendar of known seasonal sales: Steam Summer Sale, Black Friday, PSN Holiday, Nintendo eShop events. Each one has a typical timing range and average discount depth baked in.

Update Cadence

Prices refresh daily across all platforms. Predictions recalculate every night so confidence scores always reflect what is actually happening right now.