# Percentile calculations for input parameters, forecasts & EURs

After the probabilistic analysis is run, and all the runs are completed, Harmony Enterprise generates percentiles (P10, P50, P90) for input parameters, percentile forecasts, and percentile expected ultimate recovery (EUR) volumes. Details about these calculations are given below.

## Percentile (P10, P50, and P90) input parameters and OFIPs

Input parameters and original fluid-in-place (OFIP) values are calculated using the following steps:

Consider all the successful runs generated by the probabilistic analysis.

Calculate P10 for a certain input parameter as a P10 percentile of values for this parameter for all successful runs. (Do the same thing for the P50 and P90 values.)

Calculate P10-OFIP as a P10 percentile of values for OFIP for all successful runs. (Do the same for the P50 and P90 OFIPs.)

## Percentile (P10, P50, and P90) forecasts

These forecasts are calculated using the following steps:

Consider all the successful runs generated by the probabilistic analysis.

At each timestep, calculate a P10-rate as a P10 percentile of rates at this timestep for all successful runs. (Do the same for P50 and P90 rates.)

Calculate a P10 forecast by combining all the rates calculated in step 2. (Do the same for the P50 and P90 forecasts.)

Percentile forecasts do not correspond to any of the runs generated by the probabilistic analysis. Therefore, there is no combination of input parameters that results in a forecast that is equal to the percentile forecast.

## Percentile (P10, P50, and P90) EURs

EURs are calculated using the following steps:

Calculate P10, P50, and P90 remaining recoverable (RR) volumes by summing up production based on P10, P50, and P90 forecasts.

Calculate P10, P50, and P90 EURs as:

• P10 EUR = Historical cumulative production + P10 RR

• P50 EUR = Historical cumulative production + P50 RR

• P90 EUR = Historical cumulative production + P90 RR

The P10 EUR calculation shown above is not the same as considering EUR values for all the successful runs generated by a probabilistic analysis and calculating a P10 percentile among them. (This is also true for P50 and P90 EURs.)