The link that George Parker posted http://www.sqlskills.com/blogs/kimberly/transaction-log-vlfs-too-many-or-too-few/ is a good place to start.

Basically, observe the size of your log. During a maintenance window shrink that puppy down as small as possible then expand it to the observed max size. Adjust your auto growth

2k+ is not uncommon, but not ideal, in my environment 99+ would be too high. Using Pedro's query you can check the size each growth iteration to see if you are bugged. I've seen databases where 200k VLFs were reached by 248KB growth per iteration (fixed by applying the latest service pack)

You might want to start a new thread for this specific question so that some of the pros can chime in