Mammoth for T-Rex. T-Rex for Lightning. Lightning for Spirit. Spirit for Gravity. Gravity for Blizzard. Put each of those one-for-one swaps into our Blox Fruits trade calculator and every one comes back "Fair trade". Make them in that order and a 2,350 VP fruit has become a 2,000 VP fruit: 350 VP, 14.9 percent of where the run started, gone without a single verdict saying so.
Nothing is wrong with any one of those verdicts. The figures here come from our Blox Fruits value list as stamped September 17, 2026. Values on it are community estimates, not prices Roblox publishes, and the swaps below are illustrations built from the list, not a record of real trades.
Fair is a window with two edges
The calculator weights each item by its demand tier (low 0.85, medium 1.0, high 1.15, insane 1.35), totals each side, and calls the trade fair when the gap between the two weighted totals is under 5 percent of their average. All four of our calculators use the same rule. What a tier is worth is worked through in what a demand tier costs in a one-for-one trade; this piece is about the window itself.
A window has two edges. At the bottom one, what you receive is worth 39/41 of what you give, about 95.1 percent, and the verdict still reads fair. At the top one, you receive 41/39, about 105.1 percent. The edges are not mirror images because the gap is measured against the average of the two sides, not against your side alone.
For a single trade, a window is the right design. From 900 VP up, every live value on the list is a multiple of 50, so near 2,000 VP the smallest difference the list can express is 50, or 2.5 percent. A verdict sharper than that would be reading precision into numbers that do not have it.
The edges compound
The window resets with every trade. Your inventory does not. Here is what happens to a starting value when every trade in a run lands on the same edge:
| Trades in a row | Value left at the losing edge | Value at the winning edge |
|---|---|---|
| 1 | 95.1% | 105.1% |
| 3 | 86.1% | 116.2% |
| 5 | 77.9% | 128.4% |
| 10 | 60.6% | 164.9% |
| 14 | 49.7% | 201.4% |
Fourteen fair verdicts at the losing edge leave less than half of what you started with. The same fourteen at the other edge double it. A trade exactly on the line reads as a loss, so these are limits, and a run of trades just inside the edge keeps a hair more.
Real runs do not land on an edge every time. The five swaps at the top of this page give up about 3.2 percent a step on average, well inside the window, and still cost 14.9 percent. Where a run ends is set by the average lean of its trades, and no verdict shows the lean.
The run, step by step
| You give | You get | Gap the calculator shows | Verdict |
|---|---|---|---|
| Mammoth, 2,350 | T-Rex, 2,300 | 50 (2.2% of average) | Fair trade |
| T-Rex, 2,300 | Lightning, 2,250 | 50 (2.2% of average) | Fair trade |
| Lightning, 2,250 | Spirit, 2,150 | 100 (4.5% of average) | Fair trade |
| Spirit, 2,150 | Gravity, 2,100 | 50 (2.4% of average) | Fair trade |
| Gravity, 2,100 | Blizzard, 2,000 | 100 (4.9% of average) | Fair trade |
All six fruits carry medium demand, so the weights cancel and the weighted totals equal the VP totals. Twelve pairs of live rows on the list get a fair verdict one-for-one, and these five link end to end into the longest downhill run the list allows.
Read the table from the bottom and it is a climb: Blizzard to Mammoth, 2,000 to 2,350, up 17.5 percent, with every step called fair. That is the trade-up ladder climbed without a single trade the calculator would call a win. The fruits and the verdicts are identical in both directions. Only the order changed.
One caution about the middle of the run. Lightning at 2,250 is one of the list's held rungs: the two trackers the list is checked against put Lightning and Gas in opposite orders, so the row stays where it was rather than moving on one source. When your two sources split explains the hold.
The band is a shortlist, not a verdict
Every row on the value list also prints a one-for-one swap band, 0.75× to 1.35× of its value, and opening a row lists the closest fruits inside it under "Fair trades". The band is far wider than the calculator's window, and its floor compounds faster: three swaps at 0.75× leave 42.2 percent.
On the list it looks like this. Take the cheapest live row inside each band in turn and Dragon (West) goes to Gas, Gas to Gravity, and Gravity to Pain: 3,600 VP to 1,650 in three swaps, every one inside the band. Weighted for demand it is worse, 4,860 to 1,402.5, because Dragon (West) carries insane demand and Pain low. The calculator scores all three swaps as losses. Use the band to find candidates and the calculator to decide.
Which side of fair you are on
The verdict reads the same from both sides of a trade. Put Shadow on your side and Buddha on theirs and the calculator says "Fair trade" with "Adjusted gap: 75 (4.3% of average)" underneath. Swap the sides and it says exactly the same thing. The two totals are plain VP sums, 1,800 against 1,500, and the gap is weighted and printed without a sign, so neither line tells you who is behind.
To find out, weight each side yourself: multiply every row by its tier weight and add. Shadow is medium demand (1,800 × 1.0 = 1,800) and Buddha is high (1,500 × 1.15 = 1,725), so the side giving Shadow is the one 75 behind.
When both sides sit in the same tier, the plain totals already point the right way. When they do not, the plain totals can point the wrong way. Give Magma (700, medium) for Love (820, low) and the totals read 700 against 820, which looks like 120 in your favour. Weighted, it is 700 against 697, three points the other way. Small, but in the other direction.
Counters that move you across the window
When a fair verdict leaves you on the losing side, the fix can be as small as one rung:
- Ask for the rung above. Holding Gravity and offered Blizzard, you are 100 behind (4.9 percent). Counter for Spirit instead and you are 50 ahead (2.4 percent). Both verdicts read fair, and the counter is a 150 VP swing. Of the twelve fair pairs on this list, five leave the side that is behind a counter like that.
- Ask for a small add. Holding Spirit and offered Gravity, you are 50 behind. A Dark on their side (80 VP, low demand, 68 weighted) turns that into 18 ahead, and the verdict still reads fair.
- Build the offer yourself. An offer arrives already built, and whoever built it chose the pairing. The verdict cannot tell you which edge they chose. When you make the offer, you choose it.
How to negotiate a trade covers the conversation around a counter.
Keep a running number
The calculator checks one trade at a time; only you can see a run. A note of the weighted total you hold before and after each trade is enough. If it drifts down across trades that were all called fair, the drift is the lean, and that note is the only place it shows.
Taking the losing edge on purpose is sometimes the right trade, when the fruit you get is easier to trade on or is the one you were after. What an overpay is, and when to pay it is about exactly that choice. The difference is whether you made it.
The limits of this
The demand weights are the calculator's rule of thumb, not a measured market premium, and the values are community estimates from one list on one date. A re-check can move a rung and break the run above or extend it. The five swaps and the three band swaps were picked to show the arithmetic at its plainest. Most real trades land somewhere inside the window rather than at an edge, and a run usually mixes small wins with small losses.
None of that changes the point. A fair verdict is a statement about one trade, and the value you hold is the sum of all of them.