Island works because of the enormous amount of oversight you get as a defender. Attackers are running up a hill and it’s hard for attackers to slip past the net. For defenders this means there’s less risk in venturing outwards than in maps with dense geometry.
This is only true for the status quo. Any newcomers won’t suffice with the standard Dota map. There would be no reason to quit LoL. Even Dota2 isn’t doing that well and they’re backed up by Valve with Steam. Imagine. All the other MOBAS have to come up with something new to be noticed. SMITE succeeded at that by adopting an MMORPG style of control. HotS will be a classic MOBA but relies on completely changing the maps with sophisticated WoW style dungeon encounters that the teams need to battle each other around.
ETQW shook things up with extremely fun toys and unique map elements (Slipgate, EMP cage, huge underground facility) that really meshed well together. It rewarded creativity.
I agree this true for main objectives. But for side-objectives there definitely was a deep gameplay element in the class limits. In W:ET there was this recurring cat and mice game between covert ops and engineers around forward spawns. This worked mainly because of the class limits. It took a single class to sabotage and another single class to repair. There’s not even one ‘side objective’ class, there’s two for different tasks. That’s asymmetry in a game with (almost) symmetrical teams. It isn’t even tied to attackers and defenders because the forward spawns worked both ways.
And as far as side objectives go, there’s always the possibility to mix it up. Some side objectives work for all classes, some are exclusive, some work for a selection.
DB even has the opportunity to make a side objective respond differently depending on which merc interacts with it. The same objective could adopt a different function depending on whether it’s hacked or repaired. Same for damage. An objective damaged by air support could (mal)function differently from an objective damaged by a HE charge which could (mal)function differently from weapon damage, could (mal)function differently from being hacked.
A HE charge could obliterate an object. A hack could make the object swap sides, an air support could remove the top while weapon damage only takes away the cover but leaves the base skeleton intact (so that it remains impassable).
Lots of different outcomes. A nightmare to balance but incredibly satisfying and personal for the players that work hard at obtaining mercs.
I think a fun way to do a test run for that is the constructed bridge at the start of white chapel. It’s not a very important side objective so you can use it as a test-dummy for all kinds of different merc-sensitive interactions.
- Air support blows holes in the flooring making it difficult to cross over (like Fuel Dump’s bridge)
- An HE charge blows the thing entirely and permanently (the map isn’t supposed to last that long anyway)
- Weapon damage slowly withers away the cover plating (that can be repaired again)
- Hacking the bridge (from either side) electrifies it for the opponent team until the other team damages the electric device.
All these things connect the merc to it’s environment and make the player feel like they’ve earned themselves something more than just a new skin and a different weapon. They’ve actually acquired a specific key to the map’s geometry that not everyone else has.
That’s what makes F2P players tick.