I think he was saying it's not a deterrent from committing more crime, not that it's not a deterrent from going back to prison. Which is to say, it's not deterring what we want it to be deterring.

Even if that part is useful though there's no real reason to impugn the integrity of the people the GP references, and the reply does not tell us why we should find those people more likely to be prone to self-rationalization, which it needs to.

In behavioral science it is well documented that punishment is an ineffective driver for behavioral modification. In fact, for punishment to work at all it has to be consistent and immediate. Prison is neither.

So our best theory of behavioral modification states that prisons is not a deterrent for crime, nor is there any evidence for it in the records. All the while prisons are an inhumane and terrible way to treat people. Can we just please abolish prisons.

Car thieves don't steal cars while in prison. Bank robbers don't rob banks while in prison. Some crime may continue within the walls of the prison, but the world outside is protected. If a person would commit crimes for 40 years, keeping that person in prison for 20 years is a 50% crime reduction.

There are other ways to confine people then prisons. E.g. you can place people under house arrest, curfew, etc.

Also your logic if flawed. Confining people after the deed seem logically a really poor and inefficient way to reduce crime. While one criminal is confined, there will always be another. So your 50% crime reduction is closer to 0%.

A contradiction is when two things both can't be true. Not just when things are unlikely to be true.

It would be a contradiction if people only ended up in prison because they wanted to go.

In reality many people commit crimes for other reasons, like lack of food and shelter.

Its much harder to get a job with a criminal record, so many former criminals end up forced back into criminal activities for food and shelter.