This post has been de-listed
It is no longer included in search results and normal feeds (front page, hot posts, subreddit posts, etc). It remains visible only via the author's post history.
I know there are always suggestions about what kinds of mechanics can improve the game (and there are tons of really good ones I've seen on Reddit so far) but honestly this game desperately needs some kind of intelligence gathering mechanic, cause the current one is super broken.
The mechanic might work roughly like this:
You spend a resource (either time or money, but I'm kinda leaning toward time) on gathering intel on the contract you are about to undertake. The time you spend will determine the nature of the intel you get. From least amount of time required to greatest amount the intel gathered would look something like this, with 1 requiring the least amount of time and 7 the most:
1 Map layout, including major terrain features. (In addition to this we should know where we are going to deploy automatically)
2 Map hazards and cover (ie mineral fields, rough terrain, radiactivity etc)
3 Number of enemies and very general types (# of mechs vs vehicles)
4 Tonnage of mechs deploying
5 Types of mechs deploying
6 Enemy disposition on deployment
7 Armament of enemy mechs
I don't intend to suggest that this is a definitive list, but only a suggestion at how such a mechanic might work. The resource spent would either be time/money or maybe a combination of both. The idea is some missions are so easy you can go into them with little or no intel, but on others it would be imperative to research the battlefield beforehand. It might be reasonable to (after some kind of ship upgrade) let players gather intel on multiple contracts simultaneously.
This would let players decide what kind of mechs to bring, how to outfit them, and what tactics they think they'll use on the ground, or even if they want to attempt the contract at all, and would add a layer of depth to the management feel of the game.
To offset the obvious advantages that this brings, it would probably be reasonable to add some percentage chance that your attempts to gather intel are discovered, spoiling the intel and therefore changing one of the above variables (the most likely change being enemy deployment positioning, least likely the types of mechs/vehicles, not likely at all being the map or hazards). And let us choosde where on the map we can deploy (provided it's outside of enemy engagement range!)
On a different note - it's almost unthinkable that a professional mercenary corporation would go into a contract without attempting this kind of intelligence gathering. The skull difficulty indicator and Darius' little blurb don't do enough to address this. I've had 4 skull missions that were easier than some two and a halfs just because of where the enemy was in relation to me initially and how they were armed.
Lastly - while I definitely would like to see something like this one of the best things about the game is how all the mechanics work so well together. There's an almost seamless cross-impact between actions you take in the field and actions you take managing the corp, which is a an incredible thing and makes for some absolutely engaging gameplay. While I think this is a worthwhile addition, I'd be hoping that harebrain stays true to form and introduces it in a way that integrates with some already great mechanics (I realise this is much easier said than done, especially while working with variables that are difficult to quantify). But so far based on the we the current build plays I'm sure Hairbrained could pull it off.
Cheers
Subreddit
Post Details
- Posted
- 6 years ago
- Reddit URL
- View post on reddit.com
- External URL
- reddit.com/r/Battletechg...