Hello All. Im sure everybody at some point has seen, or participated in a debate about CBMM vs SBMM. We have heard every argument, and most of us are stuck in one opinion or another. What I am keying in on today is the argument that SBMM is too laggy. We have seen this argument regularly, but some players say they had a perfectly fine experience under SBMM. So today I will cover exactly why both parties can be right about their own experiences. [u]SBMM Theory:[/u] [i][/i] Nobody truly knows how bungies complex matchmaking algorithms work, except those at bungie (and honestly id assume the spaghetti isnt even fully understood by bungie themselves) But in general, the assumption is that a numerical skill value is given to all players, (how this is calculated is the great mystery) These values are then likely plotted, in to a bell curve. The standard deviation is then calculated, and the algorithm then only matches players whose numerical skill values are a similar number of standard deviations from the median. [i]The Math:[/i] [u][/u] Assuming that Bungie SBMM only allows players within 2 standard deviations of each others skill to match together, here is how it breaks down: Destiny Hourly Populations: Steam: 60k Consoles: 120k Links: https://warmind.io/activity https://steamcharts.com/app/1085660 Hourly Players in the US by platform adjusted by pop % Steam: 15,600 Consoles: 31,200 link: https://charlieintel.com/destiny-2-player-count/133971/ Hourly Players in PvP by Platform, adjusted by pop % at time of writing Steam: 2,184 Consoles: 4,368 [b]Diving Deeper:[/b] [i][/i] [u][/u] Lets make some more assumptions. Lets say the game can feel healthy if we spend 80% of our time in match, and 20% in queue. (8 minute game followed by 2 minute queue) This means that at any given time, on average 20% of the active population will be in queue OR Steam: 436 Consoles: 872 These numbers can increase if players are all queueing at the same time (in theory but not likely up to 100% of the active population could be in queue at the same time) If you look at the population breakdown of a standard bell distribution, which the assigned numerical skill values would likely fall in line with. [url]https://en.wikipedia.org/wiki/Standard_deviation#/media/File:Standard_deviation_diagram.svg[/url] [u][i][b]THE EXPERIENCE:[/b][/i][/u] For top players, (3 standard deviations above average) the pool size would be 2.2% of the active population. Or, as low as 10 people on PC. Or 20 on console. So what can the game do? It can extend queue times, or extend regional boundries and start matching players from across the globe to boost the available pool size. Where as for players near average, the pool size is effectively 47.6% (max skill disparity in a lobby) or as low as 207 players on PC and 414 on console. Without expanding the regional boundaries or extending queue times. This is a massive difference in experience. Under the same system.