From Rough to Polished: Planning, Yield, and the History of Cutting

Turning rough into a polished diamond is not a linear operation whose goal is to “preserve as many carats as possible.” It is a sequence of irreversible decisions in which geometry, quality, safety, market value, time, and risk are optimized simultaneously.

Essential

The process begins before cutting. Rough must be identified, documented, and planned. An external 3D model describes shape, but internal inclusions, fracture systems, strain, and color zoning can change every later step.

A common manufacturing flow can include:

documentation → 3D scan → internal mapping → planning → rough division → bruting → blocking → brillianteering/polishing → inspection.

Rough can be divided by cleaving, sawing, laser, or a combination of methods. Laser provides great geometric freedom, but it does not have “zero” loss: kerf and thermal risk must be included in the calculation.

Bruting forms the outline, blocking establishes the principal facet architecture, and later operations complete the geometry and surface. Diamond anisotropy is critical to efficient polishing; polishing direction is not arbitrary.

Yield Is Not a Goal by Itself

Polished yield indicates how much rough weight was retained in the polished product. It does not automatically say anything about overall quality, value, or profit.

The plan with the greatest retained weight can be inferior if it creates poor light performance, an excessively thick girdle, hidden weight, or a durability risk. Conversely, aggressively removing every inclusion can unnecessarily destroy weight and economic value.

At market carat thresholds, a small difference in finished weight can have a large commercial effect, but preserving the threshold does not justify unsafe or poorly executed geometry.

Fancy shapes can use nonideal rough more efficiently than a round. With fancy-color material, color and its face-up distribution can be more important than maximum weight yield.

Planning as a Decision Under Uncertainty

Professional planning compares multiple scenarios. It should include:

• expected final weight;
• quality and optical result;
• risk of fracture or manufacturing error;
• kerf and polishing allowance;
• cost and time;
• liquidity and probability of sale;
• the value of information that an additional scan or analysis could provide.

Software can search a large number of geometric options, but it cannot correct wrong input data. AI and planning systems should therefore be controlled by planned-versus-achieved feedback rather than treated as infallible authorities.

History Changes How a Cut Is Read

The development from natural crystals and early table cuts, through rose, old mine, and old European traditions, to the modern round brilliant demonstrates that “good cutting” has not historically meant one unchanging geometry.

Tolkowsky’s 1919 model is a key historical reference, but it is not a universal modern law defining one ideal point. Modern tools—laser, 3D scanning, and computer planning—have increased control, but they have not removed the need for expert choice among tradeoffs.

Practical Framework: Optimize the Result, Not Just Weight

A modern plan begins with documentation and 3D geometry, but the external model is not enough. Inclusions, fractures, strain, and color distribution must be mapped in volume because they can change orientation or even the entire strategy for dividing the rough. Software can compare scenarios, but it cannot compensate for data that were not detected or were labeled incorrectly.

After planning comes a sequence of irreversible steps: division by cleaving, sawing, or laser; bruting; blocking; brillianteering; final polishing and inspection. Laser provides great geometric freedom, but it is not a “lossless cut”: kerf, thermal effects, and the required allowance enter the real plan. After each critical operation, the plan should be checked again because new internal information may become exposed only then.

Polished yield measures retained weight, not quality or profit. High yield can preserve hidden weight in an overly thick stone and produce a poorer face-up result. Conversely, a lower weight yield can be rational if it produces substantially greater value, a safer construction, or a more desirable fancy shape. With fancy-color rough, retaining and intensifying perceived color can be more important than maximum weight.

The economic decision should therefore be decomposed into expected polished weight, quality, market weight threshold, face-up appearance, liquidity, time to sale, manufacturing risk, and the value of alternative plans. The highest bid for rough does not prove the best assessment; sometimes it merely marks the largest error in assumptions. AI and planning tools are useful when they display uncertainty and compare planned with achieved outcomes.

When to Escalate

Escalate when a fracture, inclusion, strain, color zoning, or unclear internal geometry can change the decision about sawing, cleaving, laser processing, recutting, or yield. In that case, the external 3D model should be supplemented with appropriate internal imaging and a plan from an experienced manufacturing specialist; an irreversible step must not rest only on automated optimization or assumed value.

Quick Check Before Reaching a Conclusion

Before accepting a technical, purchasing, or documentary conclusion, run this short check:

• Has the external 3D model been supplemented by a map of internal features?
• Are kerf, polishing allowance, and manufacturing uncertainty included in the plan?
• Am I comparing polished yield with value and risk rather than treating it as a stand-alone goal?
• Am I checking the plan after each irreversible operation?
• Is the software/AI recommendation tied to high-quality input data and an explicit representation of uncertainty?

Common Mistakes

“The highest yield means the best plan.”
No. Weight is only one variable.

“Laser has no loss and no risk.”
No. Laser processing also has kerf and process limitations.

“If the software says the plan is optimal, a person is no longer needed.”
No. A model is only as good as its input data, constraints, and assumptions.

“A historical cut is simply a worse version of a modern one.”
No. Historical styles have their own architecture, context, and sometimes collector value.

Remember

The best manufacturing plan does not maximize one metric. It seeks the best overall result after accounting for weight, quality, optics, safety, cost, market, and uncertainty.

Go Deeper in The Book

• Chapter 14 — From Rough to Polished Diamond
• Chapter 15 — Polished Yield and the Economics of Cutting Decisions
• Chapter 16 — The History of Diamond Cutting