Betting the best can be best value
I did not bet Frankel; I did bet Winter Memories. My strategy with the latter was to single in the pick four, but honestly, 11-to-20 to win probably offered more value than any other pool considering there is not a doubt in my mind that Winter Memories wins that race way more than 64.5% of the time; 80% (1-to-4) is probably a better estimate.
In the long run, betting 11-to-20 shots who should be 1-to-4 has about the same value expectation as betting 6-to-1 shots who should be 9-to-2 with the added benefit of less variance. I realize it's far less glamorous to cash $31 on a $20 bet than $140, and touting to your friends that you were all over that $3.10 winner hardly builds cache, but sometimes the price just makes sense.
Back to Frankel, who was 4-to-5 (55.5% chance) here in the states and probably should have been no worse than 1-to-2 (a 66.67% chance and 20% overlay from his off odds). Sinking your teeth into that kind of overlay is like betting a fair-value 2-to-1 horse at 5-to-2. Again, it's not glamorous, but neither is betting 6-to-5 on a coin flip, and you'd do that all day long, right?