Swish's DbD Tournament

Dino7281

Hi, so I like tournaments and I like data. I hope noone has done this before, but I didn't find it at least, so I watched tournament that was hosted by OhTofu and tried to get some data from it. Sorry, if I made a mistake somewhere.

So general rules:

Swish’s DBD Tournament V.2 (Prize not announced yet) Starting Jan 14th at 5pm EST Twitch.tv/Ohtofu Event info This will be weekly over the course of 6 weeks. Every Friday starting at 5pm EST. Casted by Ohtofu, Ralph, and Justyn. 16 teams Double Elimination Bo1 format up until finals...

-Perks can be used only once, except BT (twice)

-Number of items allowed is different for each killer

-9 banned perks for each killer and 19 banned perks for survivors + specific banned perks based on killer

-Team can't play same killer more than once

-Team can't play from same pool more than twice

Pool 1 - Blight, Nurse, Spirit, Twins

Pool 2 - Pyramid Head, Artist, Oni, Doctor, Hillbilly

Pool 3 - Huntress, Demogorgon, Plague, Hag

I wasn't interested in survivor's perks when they can be used only once, so I tracked just kills, hooks, killer perks.

Most used killer perks (41 games)

Corrupt 41

NOED 26

Ruin 23

No way out 18

PoP 11

Undying 9

Killer averages

Banned perks is just how many, you can see what perks in a link above.

Items are:

0 - no items

2 - Uncommon Flashlight + Firecracker

3 - Uncommon Flashlight + Firecracker + Common Medkit

4 - Uncommon Flashlight + Firecracker + Common Medkit + Common Toolbox

Games results

Average kill rate 2,17

Average number of hooks 5,05

Marc_go_solo

I don't get the whole banned perk idea. Seems completely arbitrary. I'd be more interested in a tournament which perhaps had perks put in groups of usefulness, and the players have to pick 1 from eaxh group and cannot reuse the perks until all have been done.

Looking at the killer setups, it's as if the same choices may have been made, whether the 9 perks were banned or not.

I'd actually like a tournament whereby players have to use a whole load of unique setups. Maybe that might be devised in future?

Dino7281

@Marc_go_solo said:

I don't get the whole banned perk idea. Seems completely arbitrary. I'd be more interested in a tournament which perhaps had perks put in groups of usefulness, and the players have to pick 1 from eaxh group and cannot reuse the perks until all have been done.

Looking at the killer setups, it's as if the same choices may have been made, whether the 9 perks were banned or not.

I'd actually like a tournament whereby players have to use a whole load of unique setups. Maybe that might be devised in future?

There was tournament like that hosted by Wwispy1.

Data from it are here:

@zarr said:

Over the last weekend, www.twitch.tv/Wwispy1 held a tournament in which every character, perk, item, add-on and offering was allowed. This makes it special because usually in tournaments there are a various restrictions for these things in place.

I have taken stats on a plethora of tournaments in the past, and every time without fail average kills settled slightly above 2. My experience with and analysis of tournament DbD has contributed to my opinion that the game is actually reasonably well-balanced at the top end, and to my understanding of why the game is statistically killer-sided in public matchmaking. Since killers can evidently compete even against survivor teams composed of some of the very best, most practiced players in the world that are also using voice communications (and do so in very efficient ways) and have highly coordinated loadouts, strategies and tactics, it is easier to understand why in the public matchmaking environment where only around 5% of the player pool even are 4-player SWF groups (of which even less use voice chat, and of which in turn only a tiny fraction is as skilled as tournament players, let alone as practiced and experienced as a team), killers overperform. I think in a scenario of a killer player facing four random survivor players (no voice communication, no prior coordination of builds, strategies or tactics) each equally as skilled and experienced as the killer player, the game is actually favouring killers. I would expect the average result to fall around 2.5 kills. Of course, different killer characters and maps affect things, but we know from public stats that none of them favour survivors universally (even the statistically weakest killer on the statistically strongest survivor map still averages above 2 kills), and so I would assume that scales into the theoretical realm of 5 completely equally skilled and experienced players facing off similarly to how the killer-leaning universal balance scales into the competitive realm. That is to say, I would expect most killers on most maps to still fall within a range of 2-2.5 kills.

One counter-argument against previous tournament stats is that there are various loadout restrictions in place for both roles. I have never taken real issue with that given that the loadouts in these tournament matches are still entirely comparable to those you will see in an average public match (if anything they are much more optimized), and argued that the overwhelming skill, experience, practice and coordination disparity would already more than make up for loadout differences compared to public teams anyway (meaning even with 10 Decisive Strikes the average public group would still not be anywhere near as competitive as these tournament teams are with their 1 Decisive Strike that they are usually allowed to have), but still, those restrictions do make the tournaments somewhat of an artificially balanced realm.

In the tournament I'm talking about now, all those restrictions were removed. So not only do you have some of the most skilled, experienced, practiced players in the world communicating with voice chat and coordinating their builds and plays, but they also get to use the most busted stuff, such as 4 Styptics/Syringes, multiple Brand New Parts, 4 Dead Hards and Decisive Strikes. This is a level of play and coordination and of sheer stacked loadouts that you will never see in public matchmaking. ...And yet, as the following stats show, the killers were absolutely able to compete.

https://us.v-cdn.net/6030815/uploads/D3QJSAMEUAP8/wispy-everything-allowed.png

So, we have an average of 2.16 kills, again as in all other tournaments a slightly killer-leaning result. We can see that 11 of the 30 matches ended in outright killer victories (3k, 4k), 13 in outright survivor victories (0k, 1k), and 6 in ties (2k). If we take into consideration that 3 times players decided to play non-S-tier killers (Pinhead, Huntress, Oni), we can say that outright victories are basically an exact split between both roles; in terms of kills, 40% of the time killers won, 40% survivors did, and 20% of the time they drew. That's a surprising level of balance given the ridiculous loadouts at play. It is also worth pointing out that killers overwhelmingly often lose less decisively than survivors, which is to say when the killer loses they usually at least still kill 1 survivor (of 13 losses, 12 (92%) were 1 kill losses), while survivors most of the time lose with 0 escapes rather than 1 (of 11 losses, only 3 (27%) saw 1 survivor escape).

Another argument people may have even if they accept the statistical fact that the game is balanced at the top end in terms of kills, is that it is not balanced in terms of hooks. I unfortunately did not count the actual amount of hooks while watching the matches, but I believe that hook stages can still be used to draw reasonable conclusions in that regard. So, if we assume the typical "the killer can just tunnel one survivor to death and then facecamp a second survivor in the endgame to achieve a "balanced" result of 2 kills" scenario that is often brought up to argue that an outcome of 2 kills does not necessarily mean the match was "balanced" in a "healthy" way, we would expect 6 stages to be the dominant result. Looking at the stats, we can see that of 30 matches, only 6 ended in 6 stages, or one fifth/20%. 22 matches ended with at least 7 stages, meaning 73% of matches were killer victories in terms of hook stages, and they were not achieved simply by tunnelling one and camping another survivor. This is also obvious by looking at the fact that none of the 6 2k matches ended in 6 stages, all of them had more. In fact, as you can see further at the bottom of the stat sheet, the perk No Way Out was used in 66% of the matches, and it obviously encourages spreading out your hooks on all survivors. On the flipside, only 2 matches ended with less than 6 stages, a mere ~7%. It is therefore obvious that in terms of hook stages (and in deduction hooks) the game leans even more toward killers.

Now it of course has to be noted that this goes for Nurse, Blight and Spirit, players didn't usually pick other killers for obvious reasons. Either way, at least those three can provably more than compete with even the absolute best that a survivor side can put up (which is way beyond what any public group ever will). This is something that surprised even myself. I always knew the game was much more balanced than many people make it out to be, and tournament results have always supported that, but I did still believe that given the ability to use the absolutely most stacked loadouts of multiple purple med-kits with Styptics, multiple BNP, 4 DS, etc., average kills would at the very least be slightly below 2 rather than still slightly above. Pretty crazy honestly.

In other tournaments even various of the other killers average out around 2 kills and 6-8 stages with 4-kill matches absolutely being possible, but tournaments do show that only roughly half of the killers are viable at the highest level, with only Nurse, Blight and Spirit viable enough to consistently be able to 4k, Twins being a close contender. It also has to be noted that the map pool in this tournament was limited, and although it included Wreckers' Yard, Dead Dawg Saloon, Rancid Abattoir and Gas Heaven which with few exceptions represent the state of balance of most maps in this game, I do think there are more survivor-favouring maps than killer-favouring maps overall. A variety of killers deserve buffs, a variety of maps should be further adjusted. But those buffs and adjustments are not strictly necessary from a public matchmaking perspective - not only can killer players simply choose to play S-tier killers every single time if they so desire, but good killer players win the majority of their pub matches even with the lowest-tier killers. Public matchmaking even with MMR is simply not able to provide good killer players with equal opponents sufficiently consistently. But likewise, the weakest killers could all be buffed, the problematic maps all adjusted, without it all too negatively impacting public matchmaking balance either.

The perks used here tell some clear stories too. There are only 19 different survivor and 15 different killer perks in the total pool. That shows how limited the pool of the strongest perks in this game is - 34 out of a total 185 perks seeing use. Not only that, but if we look at how often the individual perks are being used, it becomes clear that it's really only a handful for either role that dominate: Dead Hard and Decisive Strike alone already make for almost half of all survivor perks used, with Borrowed Time, Unbreakable, Deliverance and Prove Thyself making up another ~40%. So almost 90% of all perks used for survivors consist of only 6 different ones. For killers it's similar, with Corrupt Intervention, Deadlock, No Way Out, Lethal Pursuer, Ruin and NOED being 6 perks combining for 86% of all perks used.

I do think a ton of perks should be buffed to get them to a level where they are even approachingly able to compete for slots with perks the likes of those mentioned, but I don't think that the limitedness of perk usage at the absolute top level is necessarily all too problematic. In games there always are a fairly limited amount of S-tier characters and items, and it's no different in DbD. Ideally there would be more for a more varied top-level meta, but in public matchmaking you can already get pretty creative with your loadouts without giving up much. However, the perks do tell us a bit about the gameplay and parts of it could be seen as problematic: The prevalence of DS, BT, Unbreakable and Deliverance show that camping, tunnelling and slugging are the most effective killing strategies that survivors have to combat. Maybe the game could be better if it was reworked to concentrate more strictly on getting as many hooks as possible rather than revolve around these strategies at the top end. On the killer side, Corrupt and Pursuer highlight that the early game is crucial for killers, they have to get pressure up as soon as possible. It could help to introduce an "early game collapse" phase after all, such as simply by blocking all generators until the first chase has begun. No Way Out and NOED on the other hand highlight the fact that killers are at a disadvantage in the endgame - healthy survivors hit with a basic attack can regularly make it to a gate even over large distances, and killers regularly cannot defend gates unless they happen to spawn within close proximity of one another. I personally would not be opposed to the idea of having NOED be a basekit mechanic, without the movement speed boost, giving survivors a totem counter at base at the same time. To bring Nurse and Spirit in line with a NOED basekit reality, I think making Lethal Blink hits and Grudge hits not count as basic hits would be a welcome adjustment (it would be even without basekit NOED).

Some other notes: No Way Out's popularity shows that even the best killer players expect to get into an endgame scenario regularly in even matches. I think there has to be a change in mentality of players to not consider a game lost just because all generators have been completed. Given a desired goal of 2 kills and 2 escapes, a balanced match obviously has to see all generators get finished. Not only that, but since killers can almost always all but guarantee 1 kill in the endgame, a balanced match much of the time is expected to see all generators get finished with at least 3 survivors still alive at that point. Food for thought.

I was surprised that Circle Of Healing didn't see more play. It goes to show that other perks are more valuable if you just bring a med-kit. The standard build for most teams was: 4 DS, 4 DH, 2 BT, 2 Deliverance, 2 Unbreakable, 2 Prove Thyself, with everything else being "tech" variations, and Circle not even being a popular tech (only brought in for 3 matches of the 30).

The problematic add-ons are definitely still problematic: Nurse's range add-ons, Spirit's Mother-Daughter Ring, Blight's Alchemist Ring, Brand New Parts, Styptics, Syringes - these are all blatantly overtuned, and particularly the survivors' ability to bring 4 overtuned add-ons is ridiculous. I would even say it's precisely the ability to bring multiple that makes them a problem - if random survivors only bring 1-2 of these (which is already fairly rare), their impact potential is mostly alright. I think SWF groups should have perk and item limitations in place, such that they can only bring any perk or item/add-on at most 1 time. The killer add-ons should simply be rebalanced.

Moris are still the most balance-relevant offering in the game. Without map offerings, survivors in this tournament usually burned Petrified Oaks, and their impact on the matches was moderate at most. The Ebony Memento Moris that killers usually burned on the other hand consistently made much more of a difference. Bypassing any and all rescue mechanics (DS, bodyblock and sabo plays, pallet and flashlight saves) as well as saving time and holding positions particularly in the endgame turned out to be crucial at various points in these matches. That said, I still think moris are in a fairly good spot now.

Final note: The finals of the tournament are actually still to be played out. They will happen next weekend, so keep on eye out on the stream I linked.

Some killer perks banned would be used.

I am sure we would see Lethal Pursuer, Scourge Hook: Pain Resonance. Not sure about others, but Coup de Grace can work too for extra kills.

I am suprised noone used Deadlock, it was banned for pool 1 only.

Marc_go_solo

@Dino7281 said:

https://forum.deadbydaylight.com/en/discussion/comment/2728612#Comment_2728612

There was tournament like that hosted by Wwispy1.

Data from it are here:

https://forum.deadbydaylight.com/en/discussion/297338/statistical-evaluation-of-an-interesting-tournament

Some killer perks banned would be used.

I am sure we would see Lethal Pursuer, Scourge Hook: Pain Resonance. Not sure about others, but Coup de Grace can work too for extra kills.

I am suprised noone used Deadlock, it was banned for pool 1 only.

Thank you for that! It's a lot to go through and rest-assured I will read it when I get some more time.👍