While I like theory as much as the next guy, this is an area where I have to side with statistical analysis.
Assuming a balanced map implies a 50% win chance, then a map which deviates from this by a statistically significant amount can be called: imbalanced. We can control for Racial imbalance by using the current win rates; maps that deviate from that rate are considered a factor in imbalance beyond racial imbalance. To further help control for skill level I'm only going to consider top ranked matches, so we don't have to worry about mistaken matches between Bronze/Diamond players.
For my data source I will use TLPD.
Here are the calculations.
From this table, the most likely imbalanced maps are:
- PvZ Metal
- ZvT Blistering Sands
- TvZ Jungle Basin
- PvT Xel'Naga
- ZvT Shakuras
All of these maps have P values of less than .01 (many scientific articles require P < .05)
Thus a Terran should feel comfortable against a Zerg on Jungle Basin, but uncomfortable vs Zerg on Shakuras and Blistering Sands, and uncomfortable against Protoss on Xel'Naga. It's worth noting that Terran are favored in all maps, but that these are the only ones that deviate significantly from the standard favoring.
Note: Many of the PvZ maps did not have enough games to provide for a low enough P value.