Sybreed, on 07 October 2014 - 09:38 AM, said:
Actually.
I think I would have taken some inspiration from mobas to design a Mechwarrior game. By that, I mean that different mechs have different roles, all based on their loadout, and the skills mostly come from the fact that you play well and take the best tactical decisions to win the game. The maps in MWO don't allow for such tactics, so of course everyone will want to max weapons as much as possible. But in scenarios where you need stealth, detection, long range artillery and multi-purpose mechs for a variety of scenarios, stock builds would shine in that you would need your team to build in accordance to the map and objective.
Then, having pinpoint accuracy, max weapons, matter less. It becomes a test of tactics and positionning and good reaction time, a bit like in mobas. Right now, MWO is to battletech like a dota 2 game with only carries is to an actual, good dota game. Supports win games people
I might be speaking chinese right now, but I hope some understand.
I tend to disagree that the maps are the issue, let me explain:
People want to progress
To progress in MWO you need CB and XP/GXP
To get CB and XP/GXP you focus on rewards in game
The current reward system is extremely limited and favors damage and kills
You can build the best role warfaresystem in the world and no one will ever play scouts if they dont get CB and XP for doing their job.
**** NOW, hopefully there are major changes for the reward system upcoming. and hopefully those rewards will make running a scout worthwhile. Only time will tell. ****
My system (proposed many many times since before closed beta):
1) You choose a role when you ready up (scout/attack/support/Commander)
2) MM takes role into account as part of the MM process (1-3 scouts, 3-6 attack, 2-4 support, 1-2 commander)
3) Rewards are based upon role, (scouts get points for spotting/UAV, attack for kils, support for assists, command for ??)
Do that and people will be far more plikely to play other roles, even if the mechs involved are non optimal.