Mathematical theory and empirical data show that discovery sizes from mature plays have lognormal distributions. This allows a simple transformation to be used that converts lognormal trends to straight lines from which valuable statistical data on exploration potential is derived. After correcting for the unreliable nature of data from non-commercial accumulations and then discounting the extreme high ends and low ends (P1), discovery size distributions from a variety of mature plays outside the Middle East show certain common properties that help constrain reserves ranges in individual prospects and new play fairways. These are: P99 is 0.3MM brl or less in non-DHI oil plays and 4BCF or less in gas plays at normal depths; P50 reserves fall within the range 7-35 MM brl or 15- 150 BCF (half of all discoveries are smaller than this); risked average discovery size is generally 10-100 MM brl or 20-200 BCF, there is somewhere between a 1 in 5 and 1 in 4 chance of finding reserves equal to or larger than the average size of discovery, the average minimum economic field size falls between P70 and P30, meaning that around half of all discoveries are too small to be commercial, except in onshore areas close to infrastructure; P90 values show the smallest variation over time as a play matures but these rarely become commercial, most established conventional plays contain 1-10 billion brl o.e. of recoverable resource. In contrast upside reserves, the number of discoveries and the chance of making a discovery are unique to each play. Also, rather than indicating the exploration maturity of a play per se, P10/ P90 ratios reflect play specific limits to discovery size, although it is true that the P10/P90 ratio decreases as the number of discoveries increases. As most large discoveries are made in emergent plays, companies looking for high impact (unusually large) prospects are unlikely to find these in established plays meaning that they need to be prepared to explore at higher risk, often in frontier areas.