The focus of the thesis is elliptic curve public-key key cryptography. Bos et al. in "Elliptic curve cryptography in practice" analyzed EC keys from (Bitcoin and TLS, SSH) in order to reveal unique mistakes and vulnerabilities that arise in implementations of ECC. They generated their own keys using highly non-random scalars and tried to find a match with a database of real-world keys. The goal of this thesis is to generalize their approach in two ways: 1. Larger sets and different types of own scalars. 2. Fingerprinting of keys based on the difference of public keys, assuming patterns in differences of the corresponding private keys. The thesis will cover the following parts: 1. Student will collect ECC data from widely used sources - at least public keys from Bitcoin will be used. 2. Student will survey fails of the random number generators in the past. Based on the results of this survey he will propose types of patterns that can be expected for improperly generated RNGs. 3. Student will propose and implement the tool that will generate given types of patterns and extract private keys or fingerprint the public keys. 4. Student will analyze use of faulty real-world keys and summarize the findings.