the most interesting problem in fed cts so far is the problem of bad representation. in what ways can your atty's mistakes be chalked up to you? when should they be? if you're a judge (or a prosecutor) and you know damn well that the atty on the other side is hurting their client, or not doing as much good for their client as they could, what do you think? how can you consider that justice?
cases can be "valued" based on the percentage chance of victory and of the relief that will likely result. a good lawyer, presumably, will move your chances of winning up some not-insignificant percentage. (if you don't believe this, then you believe a lot of corporations are wasting a ton of money on representation.)
let's say it's a 15% bump, the good-lawyer bump. i think that's probably very conservative, particularly in light of how bad attys can be (self-policing doesn't work for at least this profession). so if you were a defendant and you had a 60% chance of getting off the hook with a great atty, and only a 45% chance with a bad atty, it's effectively the statistical difference between guilt and innocence.
conclusion being that justice can be bought. not in every single case, because there's always only a chance of winning or losing. a single verdict can't be guaranteed any more than a long life can be guaranteed. but in the same way that across enough people, chances of very good health can be bought, similarly, chances of very good verdicts can be bought.