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: Hawaiian Pig | Date: 11/14/09 9:01 p.m. |
In Response To: Re: So you guys want specifics on game design eh.. (RC Master) Ok. For starters. Wow, someone gets it. I was wondering if I was being too vague or incoherent. The topic is very abstract and hard to pin down in meaningful ways, but the mode of thinking seems to have specific application when you inspect individual cases (as you did with ODST). That said, I read Narc's post in the wee hours of the morning last night and wanted to get to it today. Really cool to see someone else hop into the fray. There are probably a few things I'd like to flesh out with the matchmaking side of things, as Narc clearly spent a considerable amount of effort trying to demonstrate what he thinks would happen if the skill ceiling was raised. : Only when there is a mismatch, and/or a player is still on their way to
: In practice the matchmaking system does not always work this way, due to
This is the idealized matchmaking system. However, the most important variable, as Narc fairly claims, is the size of the player population. Granted a rich normalized population, the idealized system should work perfectly. Let's assume that the current population is perfect and normalized. It would look something like this:
Now Narc suggests that a higher skill ceiling would "spread out the distribution and drive median skill down." If I understand him correctly, this essentially shift the whole curve down and then flatten it, like this:
While I agree that the median skill may go down, I'd contend that the spread of the distribution would not be even. With a high skill ceiling, I'd contend that the distribution would not be normalized, as not everyone plays games competitively. Still, I agree with Narc that the median ranks would fall. This fall, however would probably be a more accurate descriptor of player ability. 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). N.B.: I can see an argument being made here in defense of the merits of AR play. Such an argument may claim that these encounters require skill as well, and ranks in playlists that center around AR play are accurate. To dispense of this fairly quickly, consider what gametype you would rather play against a group of Generals, ARs or BRs? There is significantly more complexity involved in BR play than AR play. Due to the quick (and harder to master) execution of 4 shot kills, the employment of other strategies (such as grenades, melees, snipers, powerups, etc.) become extremely important in tipping the scales. As a result, since games tend to rely upon these encounters more often, lesser skilled players win slightly more than they normally would. This results in mismatching of players; players attain trueskill ranks that do not reflect their abilities. The existence (and commodification) of BRs in AR playlists, and their intense domination of other strategies available, confounds the outcome of games with regard to determining the player of greater skill. BRs become more like a power weapon to control, and, as AR/BR encounters are inordinately imbalanced, the weapon players spawn with becomes of utmost importance. With BR starts, the dominant strategy is equally available to all players and does not hinge on the outcomes of an AR encounter. As the outcome of games then require more from the player, I'd imagine that the median ranks in a playlist with BR starts would be lower. It should be noted that actual numbers regarding this, using existing playlists such as MLG and Social Slayer, may not exactly reflect this proposed phenomenon due to the intentions of players entering each playlist; there may be a significant amount of players in the 0-10 rank range for Social Slayer due to sheer apathy toward competition. : Personally, I'm failing to see how and why you believe that having the cap on
: While top players are moving up the True Skill ranks, perhaps then their wins
As Narc mentions, "the majority [of players do] not take your approach to playing the game." Indeed, this is why I would imagine a skewed skill distribution. With more accurate matching due to more accurate ranks (due to sufficiently complex, skill based gameplay), the higher skilled players do pull further away, but their numbers will be fewer. Essentially, you'd end up with this:
In fact, I'd suggest that more mismatches would occur at the highest ranks because of this; with fewer players at rank 45-50, the more concessions the system would have to make in matching players. This phenomenon actually occurs on the private, highly competitive, gaming servers of ICCUP. Here's a quick rundown of player ranks with respect to player population: A+ - A- (322 players (only 21 at A+))
Ranks are determined in on a simple scoring system, so they can't be directly compared to rank distributions in Halo (they are not devised to match players, but to distinguish them based solely on win rates). Still, the staggering differences in the in the amount of players at the top level as opposed to the bottom is telling, players of higher skill are much fewer in number. Just for a visual, the distribution looks like this...
It should be noted that this is a niche community of players who deliberately seek out (download and install) this private server for the sake of competitive play. These numbers represent the total players logged into ICCUP (in a 3 month season). That total is currently 66,639. On the other side of things, Battle.net logs anywhere between 60 to 80 thousand players online at a single moment during peak times, and the calibre of player is much different. With the majority of players being at a lower skill, the issues that result in mismatching players occur at the margin; players of the highest skill would have more trouble finding matches and would occasionally be matched up with lower skills. This would really only be an issue for players pushing the limits of the ceiling. It seems the donging-on-noobs scenario Narc envisions seems to assume that the population is so small that an A or B rank on ICCUP would be bumpin' uglies with the C and D ranks. This, however, won't occur, because the majority of the population still remains at the lower level. Casuals will still get matched with casuals, as there is more than enough of them to go around. With a higher skill ceiling (resulting from an emphasis on skill-based gameplay), the top players will move further up, the TrueSkill ranks will become more accurate, and players will be matched more accurately. Sounds like a better matchmaking experience to me. All of this aside, the development of playlists really does well to separate the sections of the community looking for casual relaxation and those looking for a competitive fix. They effectively minimize the amount of uglies being bumped. Now that I think about it, it's similar to how ICCUP and B.net operate (although, at peak times in Korea, this gets complicated). (I'd just like to add here that players can do both of the above. I'm not averse to hopping into Social Skirmish, holding a trigger down and running in the direction of the enemy in a straight line) The rest of this here is just lovely: : ODST on the other hand, I (and 13 others) have already completed the campaign
: Now, obviously many would say, in this situation, go play on legendary,
: Other maps too, because of the map design and volume of enemies, gameplay
: From what I've read so far, HP's game design paradigm seems admirable (make
Yes, I do admit that it's very abstract. But I do hope that some can see how it can be narrowed down for specific applications, as that was the whole intent of the post. : Chess was one example given of a game that excellent in that: it is
: Going from this there are, of course a myriad of other reaons why people
: - wow, this way longer than intended and probably won't even be read by many
I did! I'd have thought the same thing about my original post, but it was only ever intended for the type of poster/lurker who would read it.
| |
|
| Replies: |
The HBO Forum Archive is maintained with WebBBS 4.33. |
