Pottery Sherd Estimator

What Is the Pottery Sherd Estimator?

The Pottery Sherd Estimator works out three different quantification measures for a ceramic assemblage, sherd count, sherd weight, and Estimated Vessel Equivalent (EVE), so you can compare how each one represents the true proportions of ware types in a collection. Archaeologists use these figures to characterize a site's ceramic economy, compare assemblages between sites, and track how ware proportions change across a sequence. According to Orton and Tyers' foundational statistical work on ceramic assemblage quantification, the vessel-equivalent measure is the only one of the common methods that gives an unbiased estimate of the true proportions of different wares in a population.

Figure out which measure your project needs before treating any single figure as definitive, since count, weight, and EVE can genuinely disagree about which ware dominates the same assemblage.

Sherd Count and Weight: Two Simple but Biased Methods

Sherd count, simply tallying every identifiable fragment by ware type, is the fastest method to record but is directly biased by how finely a vessel happens to have broken. A thin, brittle fineware vessel can shatter into far more countable fragments than a thick-walled coarseware vessel of similar or even smaller original size, inflating the fineware's apparent share purely from breakage pattern rather than how many actual vessels were present. Recording guidance from the Chartered Institute for Archaeologists' pottery recording toolkit instructs specialists to count sherds precisely, to the nearest sherd rather than an estimated quantity, since even the raw count itself needs to be reliable before any bias correction is worth applying.

Sherd weight corrects some of this bias but introduces its own: thick, heavy fabrics such as amphorae or storage jars weigh disproportionately more per vessel than thin, delicate finewares, so weight-based proportions tend to overstate the heavier wares regardless of how many vessels they actually represent. Neither measure on its own reliably answers the question "how many vessels of each ware were really here," which is exactly why the same CIfA guidance recommends pairing count and weight with a genuine vessel-estimation method wherever the assemblage allows it.

Estimated Vessel Equivalent (EVE): The Unbiased Standard

EVE works around both problems by measuring the percentage of a vessel's rim circumference each rim sherd represents, using a rim chart, then summing those percentages across all sherds of a given ware. A complete rim scores 100%, half a rim scores 50%, and so on; set out and sum enough fragments to reach 180%, for example, and that ware's EVE comes to 1.8. Because the measure is based on the surviving proportion of a standardized reference point, the rim, rather than fragment count or mass, it is not distorted by how completely or how finely individual vessels happen to have broken.

Rim charts and recording guidance for this method are freely available through resources such as the Potsherd pottery research reference site's tools page, which provides printable rim charts in standard formats for fieldwork use.

Why the Three Methods Can Disagree Dramatically

Come back to this comparison whenever a report leans on a single quantification method to characterize an assemblage, since the gap between methods is not always small. A coarseware that dominates by sherd count can genuinely represent fewer actual vessels than a fineware once EVE is calculated, if the coarseware happened to break into many more countable pieces per vessel. This calculator flags a meaningful spread automatically, since a large gap between a ware's count-based and EVE-based share is exactly the situation where relying on count alone would produce a misleading picture of the assemblage, a divergence Orton and Tyers' statistical treatment of ceramic assemblages addresses directly when explaining why vessel-equivalent measures were developed in the first place.

Work out all three figures side by side rather than reporting only whichever was fastest to collect, particularly before using ware proportions to support an interpretive claim about site economy, status, or trade access, since that is exactly the kind of conclusion a count-only bias can distort. Look into your assemblage's rim sherd coverage before assuming EVE alone tells the whole story, since prehistoric or heavily worn pottery without identifiable rim profiles often cannot be quantified by EVE at all. Once your ware proportions are settled, the Artifact Density Calculator can relate those figures back to the specific unit or context they came from, and the MNI Calculator applies a related counting logic for faunal remains from the same site.

Accuracy and Limitations

The arithmetic in this calculator, summing counts, weights, and rim percentages and converting each to a proportional share, is exact given accurate input data. That said, EVE still depends on having enough rim sherds recorded to be statistically meaningful; a ware represented by only one or two small rim fragments can produce an EVE figure that is technically unbiased in method but still unreliable in practice due to small sample size. This calculator cannot detect that kind of small-sample unreliability on its own, so treat an EVE figure built from very few rim sherds with the same caution you would apply to any small sample in statistics. It is also worth knowing that EVE and simpler maximum-vessel-count methods do not always agree even when both are calculable; CIfA's pottery recording guidance notes that rim-based EVE often produces lower vessel estimates than a maximum vessel count approach, so state clearly which method underlies any vessel-count figure you report.

The Most Common Pottery Quantification Mistake

The mistake I see most often is reporting sherd count alone as though it directly reflects how many vessels of each ware were present, when count is specifically the measure most distorted by differential breakage between fabrics. With that in mind, always calculate EVE alongside count and weight whenever rim sherds are present in sufficient numbers, and flag any case where the methods disagree by a wide margin rather than quietly picking whichever number supports the interpretation already in mind. On top of that, be transparent in a report about which method underlies any stated proportion, since a reader comparing your figures to another site's needs to know whether both used the same quantification method before treating the comparison as valid. This consistency point matters enough that professional guidance treats it as a requirement rather than a preference; CIfA's own toolkit states plainly that consistent methodology within a period specialism is necessary, and that time allocated to quantification should be driven by what the assemblage needs rather than by budget alone.