Bloodborne speedrunner Qttsix recently proposed that hitting the brainsucker on his left side (the player's right side) with the initial ranged attack leads to better chances of drawing the favorable responses in which the brainsucker comes to the door. When I first heard this claim, I was a bit skeptical – not only because I found it surprising that non-boss enemies in Bloodborne would respond differently depending on what side of the body they were attacked on, but also because certain past claims about AI manipulation (e.g. picking up the bullets before Micolash to force him to the right, healing to lure the brainsucker to the door during his orb attack) had proved false when tested more rigorously. However, after testing out this idea, the difference it makes is undeniable. The strategy has already been put into practice for some time by both Qtt and Santzo in their speedruns.
How well does the strategy work? I think it's fair to say that it takes a skip previously feared as a late game run-killer (especially for no-quit-out speedruns) and reduces it to something routine. Sure, you can still occasionally get unlucky if the brainsucker backs out of range before you're able to get the kill, but the probability of drawing unfavorable RNG on his initial response becomes almost negligible. The results provided below are for provoking the brainsucker by means of a jumping attack with the Kirkhammer, and any instances where the jump attack was failed or missed were discarded. It should also be clarified that you must still follow the original advice of approaching the brainsucker at high speed (by sprinting down the left side of the hallway) for it to really matter which side you hammer him on; otherwise, bad RNG will come up frequently. Here are the breakdowns of attack probability:
As before, the brainsucker's responses are characterized as very favorable only if he comes within range of the unextended saw cleaver, and for the unfavorable magic orb x3 attack there's no chance to get the kill or even re-aggro him afterwards. Based on the charts, the likelihood of the dreaded orb x3 is estimated to be just 2.3% if the brainsucker is hit on his left side (player's right) with the initial Kirkhammer swing. This is much better than attacking his right side (player's left), in which case orb x3 occurs nearly 1/3 of the time (31.6%). That's really all there is to it. The preferred method for getting the kill quickly remains two charged attacks with an upgraded saw cleaver.
Notes
- Summary: New information about the Bloodborne Orphanage Key skip has recently come to light courtesy of speedrunner Qttsix. The idea is to hit the brainsucker on his left side (our right side) on the initial Kirkhammer attack, resulting in good RNG about 96% of the time. Note that we must still enter the brainsucker's awareness range at high speed in order for the new strategy to work.
- What about hitting the brainsucker exactly in the center? I ran some additional testing on this, and according to the results it works just as well. Of course, there's really no reason to risk a pixel-perfect center hit in a real situation – the purpose was only to prove that we don't need to aim so far to the right as to hit the brainsucker's left hand/arm to get the best RNG. The conclusion is that hitting him anywhere right of center (from our perspective) gives optimal results.
- Take as much time as needed to line up the initial hammer swing, just don't mess it up. After analyzing results in which we fail the jump attack or miss the target completely on the first attempt, this does have a very negative impact on RNG for this skip. Some all-bosses speedrunners have known about this for awhile already.
- How reliable is this skip when the new strategy is used? For this we'd need to know how often the brainsucker backs out of range after the first saw cleaver charged attack. Previously I'd estimated this would happen 20% of the time, but that's probably a bit pessimistic. Without going into any detailed analysis, I'd say the skip is probably 80–90% reliable with the new strats.
- How do the current projections for the orb x3 attack (31.6% or 2.3%) square with the 12.1% figure claimed in the previous blog entry? Well, since I wasn't aware during earlier testing that it made any difference which side the brainsucker was attacked on, the previous results are undoubtedly a mixture of hitting him on both sides. The numbers suggest I hit him about twice as often on the good side as the bad side. This seems reasonable considering the natural tendency is to turn just far enough to land the initial Kirkhammer hit rather than turning further than necessary.
- The same strategy works just as well if the Threaded Cane is used to provoke the brainsucker instead of the Kirkhammer. The Saw Cleaver is more of a mystery; even though the brainsucker can only be provoked by a cleaver strike to his right hand, the RNG I observed in (limited) past testing wasn't as bad as the current results would suggest.
- I was asked a couple times about the number of trials involved in previous experiments with this skip. It's a good question because it affects the statistical confidence of the attack percentage estimates. For the current tests of hitting the brainsucker on his left side, the sample size is n = 354. If we view the outcomes as binary in the sense of orb x3 vs all other attacks, the 95% confidence interval for the orb x3 attack is approximately 2.3%±1.6%. This means that it's very unlikely that the true probability of orb x3 could be any higher than 2.3%+1.6%=3.9% assuming proper execution of the door approach and hammer attack. See the 'Confidence intervals' section of the Wikipedia article on the binomial distribution for appropriate formulas.
