The current theory is a combination of marine sediaments and the bituminization stage of coal via permeable surface pathways from surrounding areas.
just one theory that is unsubstantiated. I believe you are refering to an article that Mike Stanton put in Discovery digest awhile back where he theorized that the tar sands pre-cursor oil came from the extensive Cretaceous coal seams. Unfortunately there is no evidence to support this theory. The work that is best substantiated is that done by Martin Fowler, a geochemist at the GSC and his wife Cyndi Riedeger who was a geochemistry professor and now works at one of the larger oil companies. What these researchers found was that the biomarkers present in the Athabasca and Wabasca Lake heavy oils were similar to that noted for all Missippian and Cretaceous conventional oils. The source rock that matches those biomarkers closest is the Devonian/Mississippian organically rich shales of the Exshaw Fm. Based on burial/generation modeling it can be shown that the migration pathways for Athabasca oil were long and likely convolute, involving a number of fill/spill and remigration events.
Interestingly enough Martin Fowler is at odds with what was conventional wisdom regarding biodegration. His view (at least at one time) is that biodegradation was caused by anerobic rather than aerobic bacteria.