Let's say you have Xeen, which 100 people played and 70 people liked; and Amberstar, which 10 people played and 8 people liked. In your version, Xeen will have a score of 70, while Amberstar will have a score of 8. In my version, Xeen will have a score of 70 (70% of people who played it, liked it), while Amberstar will have a score or 80 (80% of people who played it, liked it). Even if you correct it for Bayesian average which brings Amberstar's score to something like 65, it still results in much less difference between the two than in a popularity poll.
The advantage of a Google form vote is that we could create a more accurate ranking of blobbers. It has several disadvantages though:
1) Google form votes can be gamed, both by Codexers and non-Codexers.
2) It is more work to set it up and do the calculations afterwards.
3) Since there needs to be a cut-off date and it is an "inofficial" vote, I suspect that not as many would vote as in a poll here on the site.
A poll here on the site can run indefinitely and the results are visible right away. Also, if an obscure title has many votes, you can see that it is good - it's basically the same as doing calculations to find hidden gems.
