Making the map playable to more variable number of people would hamper gameplay? Your logic really confuses me.
Your assumption that all maps must be playable by all numbers of players equally is not something that I subscribe to, in this map or in general. By that metric, campaigns lack versatility, AoS maps lack versatility, tower wars lack versatility... Some things are simply designed for more than 1 or 2 players, and I don't see that as a flaw and I submit that most people wouldn't. Could you play baseball with 3 people per team? Sure, but the gameplay sucks. Is poker as fun with 2 people as it is with 4? Not every game needs to have the sort of flexibility you suggest that it must have.
In games where the components can be compartmentalized (example, every TD I have made) such that the gameplay is essentially the same with 1 as it is with 8, it makes perfect sense - and it should a common practice to make compartmentalized gameplay so that it is playable for a range of players. But in a game where everything depends upon your team gameplay, where elements are indivisibly grouped, the more you reduce the 'ideal' number of players, the worse the gameplay becomes.
Yes, some maps of these maps can and should still cater to fewer players where it is reasonable, but despite your persistent badgering of the idea of reducing the number of spawns to fit the players is the solution for this problem, I assure you that it isn't. Here's why, in a nutshell:
Pace is the most critical element to the gameplay in this map. That may not have been so transparent to you, having played it with 2 players and not made it very far into the map, but I have played the map multiple times every day since it became playable and I promise you, if you destroy that pace, the game becomes boring, easy, and uninteresting.
An isometric reduction of the number of spawns would leave the map barren, with long pauses between combat sequences and despite what you suggest, doesn't make the map more 'playable'. A non-linear reduction in the number of spawns might be reasonable, and in theory might not hamper the gameplay as much, if it weren't for the following problems that result from inherited pace:
1)
Terrain: The terrain has been designed to provide cycles of intense gameplay with intentional lulls so that players can basically have a couple of moments to recover from the action and assess where to go and how to do go about it. I cannot change the terrain to compensate for fewer players - so this element of the pace is pre-determined. Reducing the respawn rate cuts across all bottlenecks equally, creating an inherited change in the pace.
2)
Abilties: Ward of Ice provides the players with a means of instantly killing a target. The ability itself can't be removed without destroying that element of the gameplay, where some situations require an emergency use of this ability. Cut the number of spawns down considerable and you've suddenly made the game far too easy - because players can just pass through large sections of the map freezing and slashing targets. While this is true of 4 players, as well, the mana pools have been designed to act as a control for this, forcing players to use this ability very sparingly or be unable to make that emergency save when they need it most and know that they will be able to recover the mana. As with the terrain, this ability + the mana pool + and the available mana in the game produce it's own inherited pace. If I reduce the spawns, I also need to recalibrate the mana pools and the availability of mana for each possible number of players. A feat that could be accomplished, but would require an intense amount of playtesting.
There are more reasonable approaches to solving this problem - if I really wanted to set it up for fewer players - I could handicap the health/regen of the demons and the Spirit, change the availability of health/mana/XP orbs (as I have done), and reduce the movement speed of the demons / increase the movement speed of spirt - all adjusted to the number of players. These solutions circumnavigate (or directly assault) the problems with inherited pace without creating problems with the gameplay, but they also would require intense playtesting to calibrate.
As an aside from this discussion, I did see the rationale of making the game so that it could be inspected by a single player, so that players who want to try and experiment with approaches and learn the best use of their array of abilities without a threat of being defeated nearly instantly - and as a result, added a feature that automatically makes it nearly impossible to lose when playing the game in single-player mode. I feel that this 'playtesting mode' will address the problem you suggested (for players who might want to inspect what the game is like, without a full house to play) that might turn new players away from the game.