verify a far-out MM3 result?

[Karl Irikura <karl.irikura at.at nist.gov>]

I recently obtained the following enthalpy of formation (298 K) for
 bicyclo[2.1.0]pent-2-ene (C5H6; cyclobutene with a cyclopropane ring
 fused to the edge opposite the double bond):
 MM3	545 kJ/mol	(expt	333 kJ/mol)
 It's so far off I wonder if there's some problem with my copy of MM3
 (came with Tripos' "Alchemy2000" software).  It reports missing
 bond-equivalent parameters for two bonds (the sp2-sp3 bonds)--could
 they be worth -106 kJ/mol each? The corresponding bonds in the
 bicyclopentane are worth +23 kJ/mol each.
 The results look great for these related molecules:
 cyclobutene		MM3 = 158	expt = 157 kJ/mol
 cyclopropane		MM3 = 54	expt = 53
 bicyclo[2.1.0]pentane	MM3 = 156	expt = 158
 I don't think the experimental value is bad:  from G2 I get 320
 kJ/mol and from B3LYP//cc-pVTZ//6-31G* I get 331 kJ/mol.
 I'd be grateful to learn the MM3 result obtained from other software
 (or even from an independent installation of Alchemy2000).
 Thanks for any help!
 Karl
 ----------------------------------------------
 Dr. Karl K. Irikura
 National Institute of Standards and Technology
 100 Bureau Drive, Stop 8380
 Gaithersburg, MD  20899-8380
 voice: 301-975-2510	fax: 301-869-4020
 e-mail: karl.irikura at.at nist.gov
 http://www.nist.gov/compchem/
 ----------------------------------------------