We developed Polygraph. It's based on the algorithm from http://pubs.acs.org/doi/abs/10.1021/jp905049d, which documents some of the preliminary results using a hybrid database/library search. A manuscript with updated results was just submitted for publication, and hopefully will be out the beginning of next year.
As well as what has been mentioned above, there are numerous other tools out there - all with slightly different approaches/algorithms (including spectral library approaches). You might want to take a look at http://www.ms-utils.org/wiki/pmwiki.php/Main/SoftwareList which contains a list of proteomics related software including protein identification software.
I presume you are aware that you can access Mascot in a cut down version from http://www.matrixscience.com/ website - I know it is limited and you do put your data "out there", but it is one of the ones that is accepted by journals, etc (albeit with some verification required now).
Personally of the free stuff that I can download and put on my computer, I use X!Tandem and OMSSA, the later of which I have found to be a little more easy to use although not by much. Running combinations of the free search software is usually beneficial too.