A Proposed Rework to Skill Based Matchmaking
NOTE: It's only lengthy because I tried to answer as many questions as possible It NEEDS to be much more dynamic, adjusting your skill rating after every game, either raising it slightly or lowerin...
Forum Discussion
I think Season K/D, accuracy and achievements. You should not meet anyone outside 0.5 of your K/D. So if your K/D is 1.5 your opponent window will be 1 K/D to 2 K/D.
Putting players with a K/D of 1 or below in a lobby with players who have achieved the 20 kill / 4K badges is absolutely pointless. These could be paying customers being used as fodder.
Noob stomping is not good.
But for me, I think the SBMM being addressed is as important as implementing Two Factor verification for accounts.
- @SquishyVader if someone is putting full clips into people or near full clips, then they should rank up till this stops. simple sbmm then.
all the 100 percent accuracy is sweating in the high ranks. simple to implement too.
if theyre hacking then it won't be 100 percent accuracy but it will be the same accuracy from every person they fight. or a high percentage or persons.
just on server side see the statistics for clips and bullets that hit then move the people around. sure some people will feel left out by not being ranked high but they will have fun and be rewarded for their skill by competing with similarly skilled people. @SquishyVaderPush it even further I say. Only put people in lobbies who have the exact same stats. This way everyone wins and loses at exactly the same ratios. Everyone should be equal and no one should be enjoying themselves more than others. Skill shouldn’t matter. Everyone deserves a trophy.
All I did was try and create a fresh new idea and we're back to the same old tired argument of K/D = skill. I know quite a few sweats with 4ks and 20 bombs with tanked K/D's. K/D's that range between 1 and 3. If someone has a tanked K/D for whatever reason they're still going to be in your lobbies with much more skill then the rest. I wrote out all of that to really analyze a player realistically in real time and honestly I'm just disappointed with the people of this forum for being so one track minded all the time. This is why nothing ever gets done around here
@ErectRainbowHere's my point of view.
There are many things in your post that can be easily circumvented. For example, if I want to stomp noobs, I just need to do an awful performance in the placement matches and there I am.
Accuracy doesn't mean much in this game, there are moments where you can put up some cover fire to keep the enemies behind cover and thus gaining your teammates time to push, flank, heal, escape, revive, etc.. Also, the fact that Apex uses projectile based weapons (implying bullet travel time and fall off) makes accuracy even less relevant, as some prefer to aim or track the target a bit longer to be sure they land most of the shots, while others start firing early and make adjustments based on bullet traces and hit marks.
Who engages first doesn't matter, I've actually seen many skilled players firing later, because they prefer to do the bullet dodging dance to force their enemy into emptying the mag while also "locking the target" then start firing back as the enemy tries to reload or switching to their secondary. Some even risk taking a few bullets as they pop a full shield battery, knowing very well that all the shield damage will reset in a few seconds, giving them control in the situation. By the way, things like these are really hard to quantify, but they make a huge difference.
K/D doesn't mean that much either. I've recently seen someone posting his stats after extensively looking for lore stuff. Almost 1000 games, K/D was 0.17. Now imagine this guy going up to a k/d of around 1 in the next 300 matches. He'd seem average judging by the stats, but for those 300 games he would actually have a k/d of 3.76, which is waaay above average, I'd dare to say even pred/ master worthy.
I could probably keep rambling on about other things, but there are key factors that matter much more than fine tuning of SBMM. First is the queue number. The SBMM formula can be the best of the best, it will have almost no effect if the queue is barely above 60 people. Judging by how the percentages of players are distributed per rank, SBMM would need thousands of people in a queue at a certain moment to make sure no wolf gets among sheep. That roughly translates to having many hundreds of thousands, if not millions of concurrent players consistently, not just at peak times.
Then, there the waiting time. Let's say you have the ideal formula and enough active players in an area. if it takes 10 minutes for the queue to fill enough so that SBMM creates ideal brackets, the time spent actually playing would be less than the time spent in the lobby area. This is a problem in multiple ways. Generally, players don't like to wait, and some of them would break from the queue by cancelling, while the beginning of the match would catch others away from the keyboard, mouse or controller. Then, grinding the season battlepasses and limited events would be slowed significantly, resulting in a much higher failure rate, so a lot of unhappy people. Another huge issue caused caused by long waiting time is the broken flow. Long pauses can make most players lose focus and perform below their own expectations. Worse than in those matches where you only loot and move around for 15 minutes and, when you inevitably get into the fight, you're too slow to the draw and lose with an awfully small amount of damage. All in all, a long waiting time would eventually result in large numbers of players abandoning the game, basically early death, so SBMM needs to be compromised often because of this.In reality, you can circumvent ANY SBMM formula if you tank your stats. Not just mine. Let's remember that. That's the whole reason Respawn doesn't reveal how it works. Because then people would tank the needed stats to exploit it.
I've already addressed cover fire with aim assist. Respawn measures accuracy through when your aim assist goes off on a target. In case you're wondering, on PC, or with aim assist disabled, they use the same aim assist boundary framework, you just don't get aim assist. I said using weapons at different ranges. How does that make accuracy less relevant? But yes you could track individual weapon performance and pair that against where the 50th percentile is at to decide whether to raise or lower the skill rating of that player. All of that literally had to do with accuracy and someone aiming. Someone with a lot of skill at aiming will do that and have higher accuracy. And if you're saying the first 10 shots are missed to gauge the range, imagine the other guy only misses 2 shots and you die in the other 8 before your bullets hit him. That says something about the skill of the 2 players doesn't it?
If someone engages second and still wins that says a great deal about their ability for whatever reason. Dodging bullets has to do with movement, and the movement is factored in like I described. I guess you could implement a system to factor in live health percentages and how they coincide with accuracy and who gets the knock in the end.
I'm glad you tore down K/D so I didn't have to.
You also tore down SBMM and how it is essentially flawed. There's an entire study about how queue time has a negative impact on players within skill based matchmaking systems. I can't remember the university that published it. You would certainly have to widen the skill brackets enough to allow enough players to queue. Which you're right, would defeat the point of all of this. Like I said. SBMM is essentially flawed. Or you could just remove SBMM entirely. I hate SBMM. I'd love that.
