Today we’re improving how Gielinor Gains suggests trades, on the Market and in Trader. Here’s what’s new:
Amounts that sell. Suggested amounts now stay within what’s really selling at your price, with no more than 3M GP in one item.
Expected profit. Market suggestions now show the profit you can expect, and the Market sorts by it.
Chance. See how often trades like it made money within 4 hours.
Smarter Trader picks. Trader’s Auto-fill now ranks trades by expected profit and skips trades that haven’t been paying off.
Track records. We check every suggested trade against what happened next. Trader uses free slots for smaller trades with a good record.
Clearer labels. Profit that assumes everything sells is now labeled “if it all sells.”
Results
We tested the new suggestions against the old ones using a month of real Grand Exchange prices and trading volume.
| Result | Now | Before |
|---|---|---|
| Market: top 10 suggestionsCombined profit over 4 hours | 158K GP | 44K GP |
| Market: top 10 suggestions4-hour periods that made money | 88% | 80% |
| Trader: 10M budgetAverage profit over 4 hours | 134K GP | 71K GP |
| Trader: 10M budget4-hour periods that made money | 94% | 78% |
| Trader: 50M budgetAverage profit over 4 hours | 147K GP | 69K GP |
| Trader: 50M budget4-hour periods that made money | 96% | 77% |
| Trader: 200M budgetAverage profit over 4 hours | 147K GP | 89K GP |
| Trader: 50M and 200M budgetsWorst single 4 hours | 0.2M loss | 5.2M loss |
Results cover 342 starting times, 2 hours apart, from 17 August to 14 September 2026. Each suggestion was followed for the next 4 hours, after GE tax. “Before” is the site as it was until this update.
Expected profit
The old profit number assumed every item would buy and sell at our prices. Real offers sometimes take longer or don’t fill. With the new amounts, suggestions made about a third of that full profit overall.
Expected profit allows for this. The full number is still shown below it, labeled “if it all sells.” On Trader and Planner cards, profit is now labeled “Profit if it all sells,” so it’s clear what the number means.
On the Market, sorting by expected profit made the biggest difference. With the new amounts, the top 10 went from 73K to 158K GP.
Chance
Chance replaces Timing on Market suggestions. It tells you how often trades like this one made money within 4 hours. It isn’t a promise about your trade, but in our tests it closely matched what really happened:
| Average Chance we showed | How often those trades made money |
|---|---|
| 65% | 65% |
| 75% | 75% |
| 79% | 80% |
| 83% | 82% |
| 88% | 85% |
Chance is also better than Timing at telling which trades will make money, so it replaces Timing everywhere on the Market, including search results.
Trader
Trader’s Auto-fill now ranks trades by expected profit instead of the full “if it all sells” amount, and it follows the same amount limits as the Market.
Trader also keeps a track record for every suggested trade. Every 2 hours we note the trade, then check how it did against the trades that came next, with recent checks counting most. Trader skips trades that haven’t been paying off. It used to leave some slots empty, so now it fills them with smaller trades that have a good record. You’ll see the record on each trade in Trader, and on the Market when you hover over Chance.
Its average profit rose by 65% to 114%, depending on budget. It made money more often and lost much less in its worst 4 hours.
10M budget
Average profit over 4 hours
+88%10M budget, Average profit over 4 hours: before 71K GP, now 134K GP.
50M budget
Average profit over 4 hours
+114%50M budget, Average profit over 4 hours: before 69K GP, now 147K GP.
200M budget
Average profit over 4 hours
+65%200M budget, Average profit over 4 hours: before 89K GP, now 147K GP.
50M and 200M budgets
Worst single 4 hours, lower is better
96% smaller loss50M and 200M budgets, Worst single 4 hours, lower is better: before 5.2M loss, now 0.2M loss.
With 50M and 200M budgets, Trader used about 8M on average in our tests and left the rest unused. That’s on purpose: it won’t put GP into trades that probably can’t sell in time.
What to keep in mind
These numbers come from testing. They’re a guide, not a promise. A few things to know:
- Chance ran a little high for the most likely trades. When we showed 88%, 85% made money.
- A trade with no track record yet is treated as before until it has been checked a couple of times.
- Some very expensive or slow-selling items won’t be suggested anymore. You can still search for them and see their prices.
- Dump Alerts are sized separately and don’t follow the new amount limits.
- Our tests use past prices and trading volume reported through RuneLite. They can’t see your place in line or what other players will do next.
- Nothing is held for you, and you place every offer yourself.
How we tested
We rebuilt what the site would have suggested every 2 hours, using the site’s own suggestion code. Then we used real Grand Exchange prices and trading volume from the next 4 hours to estimate how each suggestion would have done, after GE tax. Buy offers that didn’t fill were cancelled, and anything not sold after 4 hours counted as sold at the price buyers were paying then.
We built every change using one stretch of market history. Before releasing, we tested the final version once, on a separate month we had set aside, with pass rules written down in advance. The Results table, the charts, and the Chance table all come from that month.
During development, we also tried moving offer prices up or down by half a percent. Every version did worse, so prices stay the same.
What’s next
We’ll keep testing our suggestions on new market data and compare them with real Trader results from players who choose to share them. When something stops working, we’ll fix it and tell you here.