Furthermore, players often fear asymmetric layouts, preferring to have an advantage that may be smaller but is more stable. Spectators, conversely, are more interested in tense asymmetric layouts with "play for 3 results" — like "queen vs 2 rooks" or "2 bishops vs 2 knights," etc. Applied to Heroes, these are, for example, layouts of "control vs wealth" or "mage vs warrior" or "shooters vs melee units," etc.
So—in the current state of esports in general and Heroes in particular, the interests of the spectator are so marginalized that it is quite acceptable to consciously marginalize the interests of the players if it benefits the spectacle.
For example, most players find it more interesting to play with a long timer, thinking about their move during the opponent's turn, etc. Spectators, however, prefer non-stop action, even if accompanied by many blunders or even gross errors.
Therefore, in the future, it will be necessary to reduce the standard timer values on all templates. Not necessarily down to blitz mode (8+4+1), although experience shows that even Jeb is quite playable here, but at least down to rapid (12+6+1).
It is also desirable to avoid templates with long stretches between forks; the typical distance between forks should be no more than one day's march for a low-level unit, which makes the game more energetic.
It is also desirable to avoid templates with off-road treasure zones, because you can easily gain +1 week of wealth there. Also, the size of the zones should not be too large so as not to encourage long slogs in the off-road areas. An exemplary clearing of a treasure zone is: take the village, spread out with low-level units, collect everything interesting in a day or two, and leave.
Thus, the optimization of templates for HotA will continue in accordance with the criteria specified in this discussion.