Posts on Current Forum | Archived Posts
|
|
|
||
|
||||
|
||||
|
||||
| Frequently Asked Forum Questions | |
|||
|
||||
|
|
|||
| ||||
| ||||
| ||||
| Search Older Posts on This Forum: Posts on Current Forum | Archived Posts | ||||
| ||||
| ||||
| ||||
| Re: So you guys want specifics on game design eh.. | |
| Posted By: Louis Wu <halo@bungie.org> | Date: 11/15/09 7:03 a.m. |
In Response To: Re: So you guys want specifics on game design eh.. (Hawaiian Pig) : For example: I would wager that in playlists with AR starts, the median rank : is significantly higher than those of playlists with BR starts (it would : perhaps look like Figure 3 below, but flipped horizontally). Because games : with AR starts tend to hinge on the first few encounters (proceeding until : a team can assert map dominance and have availability to the more dominant : BR), and as a consequence of these encounters being easier to master , : the ground is evened between players (the skill ceiling is lowered). This is actually completely dependent on map size. For a while, there were multiple playlists in which Avalanche showed up with AR starts. Better players SERIOUSLY donged on less-good players, because early encounters could easily be avoided - you just headed for a longer-range weapon if you were smart. The current Team Mythic list has two different Heretic slayer games: Team Slayer on Heretic (AR starts) and Team BRs on Heretic (BR starts). In BR-start games, there's quite a bit more long-range fighting - folks standing at the opening of their base, shooting across the middle at the other base, teams holding one side or the other and taking out challengers as they climb. In AR start games, there's FAR more running-around-the-map play - and even though there are two BRs on the map (one beneath each base), they get minimal use - because the folks who pick them up rarely have time to set up a longer-range defense (and because not EVERYONE on their team has one). Summary of my experience, based on these examples:
AR starts on a small map: less-good players have more of a chance
One other note about your post: If you look at the skill distribution in a ranked playlist on LIVE (hit X at the main playlist window), you'll see that most of them look a lot like your Figure 3. The bell curve is skewed WAY down to the lower end of the spectrum. You'd think that means that there are plenty of casuals to play against, so that the matchmaking system would never pit you against people who were much better than you... but it's almost never true. The system uses in-playlist-levels only, not overall levels. So folks who work at getting a 50 (or a 45, or whatever) generally stick in a single playlist for a long time, in order to get that level - and then branch out. And when they branch out, they end up at really low levels in all other playlists, and dong on noobs a LOT. I used to think that this would be okay; that they'd blow past the early levels in a few games, and be out of my (and other casuals') hair in no time - but I see generals in almost every single game I play - and the frequency actually increases as time goes on. I don't think there's any way to fix this - I don't think there's any way to filter out the alt accounts, or complicate the pairing algorithm by taking overall skill into account - and in fact it will get worse and worse as the population shrinks. (That is: I don't blame Microsoft or Bungie for their 'lousy matchmaking system' - I think the system does as well as it can given the disparate community and ever-changing playlists. One of the few ways they COULD affect this is by decreasing the number of playlists (and the frequency with which they add or remove lists) - and this would be a bad thing, overall.) Bottom line, though: there IS a wide range of skills in most games I play - even in lists where purportedly 80+% of the players are below rank 20 - and any changes you introduce that widens the skill gap (or, in your terms, raises the skill ceiling) will adversely affect the casual population.
| |
|
| Replies: |
The HBO Forum Archive is maintained with WebBBS 4.33. |