But looking at the statistics, in normal games with no tides of battle, Stark and Greyjoy are by far the most likely houses to win. The Tyrell player will be discouraged, knowing that this house has only won 9.3% of games. Let's look at the stats for no ToB:
Stark: 22.8%
Greyjoy: 22.7%
Baratheon: 17.1%
Martell: 16.6%
Lannister: 11.4%
Tyrell: 9.3%
This type of game would be impossible to balance win rates between all houses, and I don't think anyone would really enjoy that anyway. The geography of the map itself is unbalanced. However, let's look at what ToB does for the win stats:
Stark: 24.1%
Greyjoy: 21.6%
Martell: 15.5%
Baratheon: 14.5%
Lannister: 13.2%
Tyrell: 11%
So Stark, the leading house, enjoys a 2% increase. But the other three leading houses, Baratheon, Greyjoy, and Martell, all suffer a decrease, and the bottom two houses Lannister and Tyrell enjoy a 2% increase in win rates.
It appears then that ToB does offer a small benefit in making the game slightly more balanced. This benefit won't sit well with players who enjoy being able to predict every single variable in the battle, but it does reflect a bit more realism is uncontrollable variables in battles (such as the weather, morale, officership, and other intangibles).
For me, this change is worth it, and I will be using ToB in more of my hosted games in the future. It's also a refreshing difference from years of playing without them.
What does the community think? Is ToB the best way to even out the win rates of the houses, or are there other less random ways to achieve this? The statistics tab on this website is a wonderful tool for analyzing this.