Interested in volunteering to help moderate for the Forums? Please fill out an application here: https://dbd.game/moderator-application
Kill Switch update: We have temporarily Kill Switched the Forgotten Ruins Map due to an issue that causes players to become stuck in place. The Map will remain out of rotation until this is resolved.

http://dbd.game/killswitch

Machine Learning

HandsomeJack_049
HandsomeJack_049 Member Posts: 142
in Feedback and Suggestions

Let’s talk about Machine Learning, a perk (one of three) from a killer who disappeared almost as rapidly as he was introduced. Machine Learning is a failed perk. I am unaware of what official BHVR analytics demonstrate, but NightLight (a self-reported DBD analytics website, so take some of the numbers with a grain of salt) has Machine Learning’s pick rate sitting at a whopping .39%. This is a dead perk.

What is wrong with Machine Learning? I think the fundamental issue is the limitation and prerequisite for activation.

Machine Learning requires two unique gen kicks for the perk to be applied to a generator (thus compromising it) and this requirement resets if a compromised generator is completed. This is a completely arbitrary activation requirement and makes no sense compared to other compromised perks: Dragon’s Grip requires 1 kick and Call of Brine requires 1 kick for their effects to apply the generator.

Furthermore, Machine Learning requires a killer to lose the game. The perk’s effects only activate when a compromised generator is completed. And while that requirement alone is not unusual, the only other perk with a similar effect (Fire Up) permanently strengthens the killer for every completed generator. Unlike Fire Up, you must compromise a generator first and then that generator must be completed for the pay off (30 seconds stealth and haste). This mini-powerup is an okay return for a losing a generator, but you are likely not going to get more than 2-3 uses (if even that) per game. A perk that only activates a handful of times is not worth the limited perk slots available.

What can be done?

I think we should rework Machine Learning to be more in line with other compromised perks – apply an effect for a compromised generator actively being worked on.

For example:

Performing the Damage Generator Action on a Generator activates Machine Learning: While Machine Learning is active, the damaged generator becomes compromised, highlighting its Aura you in yellow, until it is completed. Receive the undetectable status effect while a survivor is actively repairing the affected generator.

[-]

This would address both the major issues of Machine Learning and bring the perk up to a more viable state in the current game. You can mess around with the effect, but the outcome should be to remove the 2-kick activation and the loss of a generator requirement (which means the effect should be toned down since it will be in play more often.)