When to cut a loss

Every flipper can name the position that ruined a month. It is almost never the one that fell the furthest. It is the one that fell a little and was kept for six weeks because selling would have made the loss official.

The arithmetic of an instant exit is covered in Dumping vs waiting it out. This lesson is about the decision that comes before the arithmetic, which is whether the position is still the position you bought.

The gold is already gone

The price you paid is a fact about the past. It has no influence on what the item will do next, and the market has no idea what it cost you. Holding a position because selling would confirm a loss is holding it for a reason that exists only inside your head.

The useful reframe is that you own two things: an item, and a quantity of gold that is currently that item. The only question worth answering is whether that gold is better spent as this item or as something else. Your entry price is not part of that question.

The replacement test

One sentence covers most cases: if I held the gold instead of the item, would I buy it right now at this price?

If yes, you are holding it because you want to, not because you are stuck, and that is a legitimate position. If no, you are holding it only because you already own it, which is not a reason. The awkward middle answer, "maybe, but not this much of it", is a real and common one, and it tells you to cut part of the position rather than all of it.

The test works because it strips out the entry price entirely. Ask it out loud once a day on any position that is under water and it will resolve most of them without any further analysis.

Dip or broken thesis

A price falling is not information. Why it is falling is. Sort the cause into one of these before doing anything else.

Usually a dip

• A large seller cleared the buy side over twenty minutes and nothing else changed.
• The fall happened during the quiet hours of the day and the buy side was thin anyway.
• Everything in the same category moved by a similar small amount, with no update behind it.
• The price is inside its recent range and has been in this part of the range before.

Usually a broken thesis

• An update changed how the item is obtained, used, or made unnecessary.
• A new source of supply appeared, or an old source got much faster.
• The content that consumed the item stopped being played.
• The price left its recent range in one move and stayed outside it.
• You cannot explain the fall at all, on an item you did not really research.

That last one deserves its own line. Not knowing why is not the same as a dip. If you had no thesis to begin with, there is nothing to break and nothing to wait for, and a position you cannot explain should be the first one you close when you need a slot.

Four stops, in order of usefulness

1. The thesis stop

Written before you buy, in one line: what has to stay true for this to work. If that line stops being true, you sell, at whatever price the market is offering. This is the only stop that reliably prevents the six week disaster, because it triggers on the cause rather than on the number.

2. The time stop

Every position gets a deadline appropriate to its type. A fast supply flip gets a day. A gear position gets a week. When the deadline arrives, you either sell or explicitly rewrite the plan with a new deadline, and rewriting more than once means the position is being managed by hope.

3. The size stop

A cap on what any one position may cost you. If no single flip is allowed to lose more than about 5% of the bank, then the maximum position size follows from the worst realistic fall for that item type, and the sizing does most of the protecting before anything goes wrong.

4. The price stop

Sell if it falls X%. This is the weakest of the four, because a thin item can trade through your level for ten minutes on almost no volume and take you out of a position that was fine. Use it as a prompt to reread the thesis rather than as an automatic exit.

Recovery differs by item type

Waiting is a bet that the price comes back. How good that bet is depends almost entirely on what kind of item you are holding.

• High volume consumables. Runes, food, common potions. Demand is continuous and enormous, so ordinary dips fill back in quickly. The catch is that margins here are thin, so a long wait destroys the economics even when the price does return. Cut on time, not on price.
• Skilling supplies. Logs, ores, bars, planks. These are supply driven, and supply changes are sticky. A new efficient method or a wave of farmed output can hold a price down for weeks. Waiting is a real bet here, not a formality.
• Gear and mid tier uniques. Abyssal whip has traded between roughly 766K and 930K recently, a band of over 20%, on around 7,000 trades a day. Items like this drift within wide ranges for months and can settle at a new level after a combat change. Patience is often right, and it needs to be measured in weeks, which means the position must be small enough to forget about.
• Update driven items. Anything that moved because of a patch. Assume the new price is the price. The mechanism did not go away because you bought before reading it.
• Very high value items. Dragon claws at around 37M to 38M trades roughly a thousand times a day, so a single position is a meaningful share of daily volume. Exits take time regardless of your decision, so the cutting choice has to be made earlier than it feels necessary.

A worked decision

Same position, same loss on the screen, opposite answers. The only input that changed was the cause, which is why the cause is the first thing to establish and the number is the second.

Recording the loss

A cut loss you do not write down teaches you nothing and quietly disappears from your sense of how you are doing. My Flips records the closed position with the real after tax figure whether it was a gain or a loss, which is the point: a profit total that only contains good outcomes is not a profit total.

Add one line of your own next to it. Why you bought, why you sold, and which of the four stops fired. After a month of doing that you will find that the same stop fires for the same reason over and over, and that is the single most useful thing your own records will ever tell you.

Better still, arm the stop rather than remembering it. Every tracked flip can carry one price alert, set from the row or from the Alerts page, and it fires exactly once when the price reaches your figure. Put it at the level you decided on while you still had a view, and the decision reaches you instead of you watching a chart looking for permission to change your mind. That is the entire difference between a stop and an intention. Nothing is armed by default, so this is a choice you make per position, and the once-only firing means a price hovering around your number will not turn into a stream of notifications you learn to ignore.

Case B above is the one this changes most. A three day deadline and a target at break even is a good plan and an impossible one to hold, because holding it means not looking, and not looking means missing the moment it comes back. An alert at break even lets you genuinely stop looking. One automated warning exists on top of whatever you set, for demand collapsing while a flip is under water, and it can be switched off on the same page if you would rather make every call yourself.

Then move on properly. The gold that came back from a cut position is ordinary gold. It has no memory of the loss and no obligation to earn it back in the same item, and trying to win it back in the item that took it is how one bad position becomes three.

When not to cut

Three situations where the reflex is wrong.

• Into an empty buy side. If the only bids left are far below the market, dumping fills you at each of them in turn and the average is far worse than the price you read. Split the exit over hours instead.
• During the thinnest hours. The buy side is smallest when fewest people are online, so a panic exit then pays the widest possible cost for the privilege.
• Because the number is red. Every position that has not yet worked looks like a loss. The tax alone means a flip is under water the moment it is opened. Red is the normal state of a young position, not a signal.

The rest of the time, the honest version of this lesson is short. Decide in advance what would make you wrong, check it once a day, and when it happens, sell. The Margins board will tell you in seconds whether the spread you were waiting for still exists, and the Flip Finder will give the freed gold somewhere better to be.

Item pages used above

Live prices, volume and buy limits for the items in this lesson: Saradomin brew(4), Abyssal whip and Dragon claws.

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.