What a Flipping Win Rate Really Means
A win rate is easy to quote and easy to fake. This is how ours is measured, why it only counts trades that actually happened, and why a win rate on its own never tells you whether a strategy makes money.
Every tool that recommends flips will happily tell you its win rate, and almost none of them tell you how it was measured. That gap is where most misleading numbers live. A win rate is only as honest as the rules behind it, so this guide walks through exactly how ours is defined, what it deliberately does not count, and why you should never judge a flipping strategy on win rate alone. The actual current figure is published live on the track record; this guide is about how to read it.
What a win rate is
A win rate is simply the share of flips that ended in a profit rather than a loss. If a hundred flips are graded and seventy of them made money after tax, the win rate is seventy percent. It is a useful headline because it answers one plain question: when the tool says a flip is worth taking, how often is it right? But that plainness is also its weakness, because the number hides everything about how big the wins and losses were, and everything about which trades got counted in the first place. A win rate is a starting point for judging a strategy, never the whole verdict.
How every flip is graded
The important word for any win rate is forward-tested. It is trivial to look back over history, pick the flips that happened to work, and quote a glowing number. That is marketing, not evidence. A forward-tested record instead posts a call first and grades it later, against prices that had not happened yet when the call was made. That is how GE Uncut's record works: a flip is posted, and then, as real prices come in over the following window, it is scored against what actually occurred. Nothing is graded on the day it is posted, and nothing is quietly removed later because it went the wrong way.
This matters because it removes hindsight from the equation. A backward-looking number tells you a strategy would have worked on data it was built from, which proves very little. A forward-tested number tells you it kept working on data it had never seen, which is the only kind of evidence worth much. The full, ungrouped record, including the flips that lost, is on the verified signals page, and every item page carries that item's own record too.
Only trades that happened
There is one honesty rule that quietly matters more than any other: a win rate should only count trades that could actually be made. A posted flip whose buy offer would never have filled at the suggested price is not a win and not a loss; it simply did not happen, and counting it either way would distort the number. GE Uncut measures its win rate only over flips that genuinely completed, so the percentage describes real, realisable trades rather than paper opportunities that no player could have captured.
Why this cuts both ways: excluding trades that never filled keeps the win rate honest, but it also means the headline number describes flips you could have made, not every flip ever suggested. Always read a win rate as "of the flips that actually completed," because that is the only version that means anything.
Why win rate is not profit
Here is the trap that catches almost everyone. A high win rate does not guarantee profit, and a lower one does not mean a losing strategy. What matters is the win rate combined with the size of the wins and losses. A strategy that wins ninety percent of the time but gives back everything on the occasional large loss can easily lose money overall. A strategy that wins only half its trades but wins bigger than it loses can be very profitable. Win rate answers "how often," and you always need "how much" alongside it before you can say anything about whether a strategy actually makes gold.
Strategy A: wins 90% of flips at +1% each, but the losing 10% average -12%.
Average result per flip = (0.90 × 1%) + (0.10 × -12%) = -0.3%. A great win rate that still loses.
The lesson: never read a win rate without asking what the wins and the losses are worth.
This is why our record publishes more than a single percentage. Average return per flip sits alongside the win rate precisely so the two can be read together, because either one on its own can flatter or mislead. When you look at any flipping tool's win rate, the first question to ask is always "and what is the average size of a win versus a loss?"
The survivorship trap
The other way win rates get inflated is survivorship: quietly leaving out the trades that went wrong. It happens in obvious ways, like deleting bad calls, and in subtle ones, like only counting flips that were closed and ignoring the ones left open and underwater. A losing position that is never closed never books its loss, so a record that only looks at closed trades can look far healthier than reality. An honest record has to account for what is still open, not just what was tidily completed, and it has to keep the losers in the sample. When you judge any published win rate, look for whether the losses are visible; if you cannot find them, assume the number is flattered.
How to actually use it
Treat a win rate as one input, not a verdict. Use it to sense-check whether a tool's suggestions are better than a coin flip and stable over time, then look immediately at the average win and loss sizes to see whether the strategy actually compounds. Prefer records that are forward-tested, that count only trades that could have happened, and that show you their losers. On an individual flip, the item's own record is more useful than the global number, because it tells you how that specific item has behaved rather than the average across everything.
In the app that is a two-second check rather than a research project. Every signal in the Flip Finder carries the item's own verified record on the card, so the confidence number is next to the flip it applies to and you never have to go looking for it. If you want the fuller picture before committing, open the item's page from the card, or use search to pull it up directly; the same record sits there alongside the live prices and the volume. Asking the assistant works too, and is usually quicker mid-session: name the item and ask how its calls have landed, and it reads the same record back to you in a sentence.
One habit is worth forming here. When an item's own record is thin, that is not a red flag, it is an absence of evidence, and the right response is to lean harder on your own read of the market rather than on the percentage. When the record is well populated and steady, you can let it carry more weight. Reading the record as a confidence dial rather than a promise is the whole skill.
The number that should matter most to you is your own, and the same rules apply to measuring it. Track a flip when you place it and it sits in My Flips as an open position; file the sale and it moves into your history with the after-tax profit worked out. Your history is then a record built under exactly the honest conditions described above: it counts only trades that actually happened, it keeps the losers, and nothing is graded before the gold has moved. Reading your own wins and losses side by side is the fastest way to find out whether a high hit rate is quietly being paid for by a handful of large losses, which is the trap this whole guide is about.
Be careful of the survivorship trap in your own books too. A losing position you never file is a loss you never counted, and a My Flips page with several old, underwater positions still sitting open is telling you something your completed history is not. Sort those out honestly and your own win rate starts describing your actual flipping rather than the flattering half of it.
All of that is why GE Uncut keeps the record public and forward-tested rather than asking you to trust a marketing figure. You can read the live win rate and average return on the track record, see any single item's history on its own page, and decide for yourself. The next lesson looks at a different measurable question that a win rate never captures: how long flips actually take to fill.
GE Uncut is an unofficial, fan-made tool, not affiliated with or endorsed by Jagex Limited. RuneScape and Old School RuneScape are trademarks of Jagex Limited; in-game content and item images are the property of Jagex Limited, used for reference only. Price data is from the OSRS Wiki real-time prices API. See data attribution.