We have source-code verificaton and results-testing as the two options for software verification. With results-testing, the credibility of the results is the issue: are they genuine results, or faked by the casino / tester? If that is established / establisheable, then you can move forward. With this alternative, that problem is removed, but there are others (please humour me if these questions are in any way irrelevant): how regularly is the code tested? how is one guarenteed that the "fair" code you examine is the same code run by the operator? Is code-testing foolproof? - meaning, a code which "passes inspection" cannot then entertain the possibility that non-random numbers may still be generated? Can a "fair" code not be somehow bypassed? Do you consider source-code verification a better alternative to result-testing? If so, why? And the old chestnut: What guarentees is it possible to offer the consumer (me) that you are on the level and that your pronouncements may be trusted?