Sift Threshold Worked Example
This deliberately small example is not a real product forecast. Its purpose is to make the calculation order auditable: all outcomes stay in the modeled draw, then low-priced drawn cards are removed from sellable value.
By The Expected Value editorial team · Published
Illustrative box inputs
| Outcome group | Expected copies | Price each | Gross contribution | Included at $1 sift? |
|---|---|---|---|---|
| Chase foil | 0.08 | $50.00 | $4.00 | Yes |
| High-value rare | 0.60 | $6.00 | $3.60 | Yes |
| Regular rare | 0.40 | $2.50 | $1.00 | Yes |
| Sellable uncommon | 2.00 | $1.10 | $2.20 | Yes |
| Low-price commons | 10.00 | $0.18 | $1.80 | No |
| Token | 1.00 | $0.40 | $0.40 | No |
The expected-copy values remain unchanged when the sift is enabled. The box still produces the common and token outcomes; the seller simply chooses not to count them as sellable inventory.
Step-by-step math
- Gross opened EV: $4.00 + $3.60 + $1.00 + $2.20 + $1.80 + $0.40 = $13.00.
- Value removed by a $1 sift: $1.80 + $0.40 = $2.20.
- Post-sift sellable EV: $13.00 - $2.20 = $10.80.
- Modeled percentage fees at 12%: $10.80 × 0.12 = $1.30 after rounding.
- Net proceeds after modeled percentage fees: $10.80 - $1.30 = $9.50.
If the sealed product cost $8.00, the simplified profit after acquisition cost would be $1.50 before unmodeled tax, fixed order fees, shipping, supplies, returns, and labor. Changing that product cost does not change opened EV; it changes the purchase economics.
What the wrong order would do
If the $0.18 commons and $0.40 token were removed from the card pool before drawing, their probability could be redistributed to the four higher-price groups. That would manufacture extra valuable pulls and overstate post-sift EV. The sift describes what the seller keeps or lists after opening; it does not rewrite the sealed product.
The same principle applies when a low-price outcome occupies a special slot. The physical branch remains part of the pack even if it contributes no sellable value under the chosen threshold.
What this example does—and does not—show
It shows the arithmetic of an expected-value summary. A real simulation draws integer card outcomes, produces a distribution of boxes, and can have a mean that differs sharply from its median. It also has price-coverage questions: if the chase foil price were unavailable, the $4 expected contribution would be missing rather than silently valued at zero.
Next, read the full fees and sift guide or compare two $100-EV products with different loss rates.