A few less commonly available attacks can add significantly to Mallory's chances of success:

Adaptive chosen plain text: This attack is just a more specialized version of a general chosen plain text attack. Each time Mallory inserts one chosen plain text and intercepts its encrypted version, he determines some statistical property of the encryption. Later chosen plain texts are selected in order to exercise different properties of the encryption.

Chosen key: An attacker might have a means of encrypting messages using a specified key. Or the specified key might only have certain desired properties. For example, if a key is indirectly derived from a different part of the protocol, an attacker might be able to hack that other part of the protocol, creating usable key properties.

Chosen cipher text: An attacker might be able to determine how selected cipher texts become decrypted. For example, Mallory might spoof an encrypted message from Bob to Alice. Upon attempting to decrypt the message, Alice will wind up with gibberish. But Alice might then mail this gibberish back to Bob or store it in an insecure way. By choosing cipher texts (or really pseudo-cipher texts) with desired properties, Mallory might gain insight into the actual decryption.