8127/frame/fm

a.
Predictive Value of a Simulation
It is common, particularly when attempting a reconstruction, for the crim-
inalist to try to recreate or simulate the circumstances he thinks led to the
scenario found at the crime scene. This kind of experimentation can be
informative, but its limits must be well understood. All too often, having set
up conditions that appear to duplicate the physical observations, the inves-
tigator will then conclude that he has discerned what happened. Doing an
experiment is a good idea, but it is crucial to understand that an experiment
can only tell you that the evidence 
could
have been produced using your
simulated conditions. It does 
not
tell you either that this is the only way to
produce the evidence, or that the incident happened in the way you envision.
However, failure to recreate an expected result can be helpful in eliminating
that set of circumstances and actions from consideration, thus reducing the
number of possibilities. A simulation supports a reconstruction, it does not
prove one.
In one case, it was suggested that fine blood spatter found in a particular
pattern was produced by a contact gunshot wound with the victim and
suspects in particular positions. The analyst simulated the circumstances
that he thought had led to the scene as it was first found. Sure enough, he
reproduced what he was looking for, including particular sizes of blood
spatter drops at a specific distance from the victim and in a similar pattern.
He then testified that his simulation proved his reconstruction, and that the
events happened in exactly the way he envisioned.
This criminalist could not conceive that an alternate explanation for the
evidence might exist; he was caught in his own expectations and prejudices
about how the crime could have occurred.