The liberal argument that punishment does not deter criminals is a curious one. There is probably some merit to the argument that those who break the law don't expect to get caught, but as the prison population testifies many do; and many offenders have string of previous convictions so surely experience would teach them that prosecution is a likely outcome. In this small instance from Stockton, the offender had 54 previous convictions. That's no typing error - by the age of 18 he had already been convicted fifty-four times of criminal activity when he decided to punch someone and steal their bicycle. It seems unlikely that he would discount completely the possibility of being caught a 55th time.
A rational criminal would look at the likelihood and the consequences of getting caught and decide if the reward justifies it. This simple rationality lies at the heart of most decisions to break the rules, from chancing 10 minutes without putting money in a parking meter to a multi-million-pound bank heist. By way of violent bicycle theft. Using this equation then dealing with crime is a matter of increasing the likelihood of detection and increasing the penalties for those convicted sufficiently to deter criminal activity. These are the two levers we have. There are crimes of passion where rage obscures reason, but it seems unlikely that the decision to steal a bicycle falls into this category.
Increased detection rates must surely be a goal of any police force, but as noted, at 54 previous convictions no rational person could discount the possibility completely. Dwayne is either a raving lunatic or he has no fear of the law.
The argument against deterrence then rests on the assumption that criminals have an irrational belief that either they won't be apprehended, or will suffer no ill consequences if they are. They simply act on random impulses to harm people and their property, and no harshness of sanction will alter this behaviour. The problem with this view is that you can not reason with irrational people. You can not change or predict their behaviour. It is random.
Dwayne in the article linked above doesn't quite seem to be acting on random, irrational impulses. He didn't, so far as we know, jump through a plate glass window, he didn't punch a baby or pretend to be a tree for hours on end. He violently stole a bicycle because he thought that he would either not be prosecuted at all or the consequences of prosecution were not severe enough to deter him.
If he is simply irrational and won't respond to disincentives then the only sensible course of action is to lock him up indefinitely to prevent him causing further harm, or to somehow re-educate him to look at their offending rationally, which still brings us back to the need for negative consequences. You can have rehabilitation efforts along-side tougher sentencing but without the starting point of a rational person who will respond to incentives and disincentives you have only random violence and theft. If you truly believe criminals are simply irrational then by definition you believe that they can never be safely released without a high probability of recidivism.
Dwayne was sentenced to 21 months in a young offenders' institution. With good behaviour he will probably be out in 12 months and would anyone be surprised if he re-offends? Whatever efforts have been made to reform Dwayne have failed on 54 previous occasions, but what is clear from his astonishing tally of convictions at his young age is that one thing that hasn't been tried is a meaningful period of incarceration.