These are not easy questions but they are questions we need to be asking in the courtroom. If fingerprint evidence is introduced we need to tell our juries the error rate could be as low as .8% or as high as 34%. Neither should be acceptable. We should request a jury instruction that tells the jury to consider that error rate (and other factors) in determining what credibility should be given to the forensic science introduced in the case. Prosecutor has an expert who says there is no error rate? That claims the method is 100%? Silly. Point out there was literally a whole conference on this topic. They must not have attended it. Point to learned treatises that say otherwise. Ask them to produce a study that supports a 0% error rate.
This is an effective attack in a case where the forensic science is the only evidence connecting a client to a crime. Cold hit fingerprint or DNA hit? Error rate. No other evidence linking the client to the charge? Error rate. Point out that even if the error rate is low, that this is the exact type of case where an error can occur. Other cases usually have other evidence linking a defendant to a crime. A case that wholly relies on forensic science is a ripe case for error and a ripe case for a wrongful conviction.
Don't be scared that the whole case is based on forensic science. That is a sword for your client. Use it.