ORD - Open Results Data

On this page:


Introduction

ORD was initially developed during the Lloyd’s Lab innovation project (Cohort 3) in 2019, by a working group led by Oasis, that focussed on constructing model agnostic results formats and appropriate data formats. These model outputs cover an extensive suite of results that can be isolated by aspects of the exposure data, financial and statistical perspectives.

Like OED, the purpose of ORD is to ensure model outputs are consistent, transparent, and model agnostic. The ORD package is intended to contain the outputs of all current catastrophe models and to be extensible to cope with future developments. The ORD outputs are results tables that capture all of the possible data that a model could produce.

Listed below are the different result tables.


Tables in the ORD format

Table

Acronym

Description

Currently in Oasis

Sample Event Loss Table

SELT

Sample losses for each event at the appropriate summary level

Yes

Quantile Event Loss Table

QELT

Distribution of losses at user specified quantiles for each event at the appropriate summary level

No

Moment Event Loss Table

MELT

Summary stats (mean, SD etc) for each event at the appropriate summary level

Yes

Sample Period Loss Table

SPLT

Sample losses for each event within each period at the appropriate summary level

No

Quantile Period Loss Table

QPLT

Distribution of losses at user specified quantiles for each event within each period at the appropriate summary level

Yes

Moment Period Loss Table

MPLT

Summary stats (mean, SD etc) for each event within each period at the appropriate summary level

Yes

Average Loss Table

ALT

AAL (Average Annual Loss), SD

Yes

Exceedance Probability Table

EPT

Exceedance probabilities for VARs and TVaRs of maximum and aggregate event losses in a period

Yes

Per Sample Exceedance Probability

PSEPT

Exceedance probabilities for VARs and TVaRs of maximum and aggregate event losses in a period for each sample

Yes

Error Log

ELOG

Contains error codes for locations that were failed to be processed by the model

No

Aggregate summaries of exposure within model domain

DOMEXP

Model specific so needs to be in the Results area

No

Aggregate summaries of rejected exposure by reason rejected

REJEXP

Model specific so needs to be in the Results area

Partly

Note

A model may not produce every result table; some may not be applicable from that scenario. However, this package covers an extensive suite of model outputs for multiple perspectives and calculations. All the outputs are available in order to cater to all model types and ensure continued interoperability.

The results can then be used to enable users to conduct comprehensive catastrophe risk analysis, quantify potential losses, and make informed risk management decisions. Results can be generated at a variety of summary levels, including single-way summaries and multi-way summaries. The way such complexity is represented in ORD is shown in the following example:


ORD summary levels example

Consider an analysis run where the following output summaries are selected:

  • Summary 1: Single way summary of results by OccupancyCode

  • Summary 2: Single way summary of result by ConstructionCode

  • Summary 3: Multi-way summary of results by CountryCode x AreaCode x LOB

This is represented as follows (using mean and SD as a proxy for any kind of results). On the left are the link files and on the right are the summary files.

Summary 1

SummaryId

OccupancyCode

SummaryId

MeanLoss

SDLoss

1

1050

—

1

23

34.5

2

1150

—

2

353

529.5

Summary 2

SummaryId

ConstructionCode

SummaryId

MeanLoss

SDLoss

1

5050

—

1

234

351

2

5100

—

2

467

700.5

3

5150

—

2

346

519

Summary 3

SummaryId

CountryCode

AreaCode

LOB

SummaryId

MeanLoss

SDLoss

1

US

FL

A

—

1

942

1413

2

US

TX

A

—

2

256

384

3

US

FL

B

—

2

390

585