Verifying  
Help Desk
MLG Store
2009 Pro Circuit
Pro Circuit Landing Page
Online Pro Circuit
Broadcast Personalities
Old Spice Rising Star Award
Season Calendar
Pro Circuit Archive
Photo Gallery
Standings
Halo 3
Gears of War 2
Call of Duty 4
Rainbow 6 Vegas 2
World of Warcraft
Pros
Pro Teams
News
Latest News
Featured Articles
Transactions
Timeout - Player Profiles
CaLANdar
Video
Dr Pepper Ultimate Gaming House
MLG/ESPN Top 10
MLG Championship Sunday
ESPN Saturday Night
Old Spice Report
Stride Championship Series
Get Better Fast
User Submitted Video
Tournament Gameplay (VoD)
Pro Player Features
Top 10 User Videos of the Week
More Video
Blogs
Pro Player Blogs
Forums
Compete Online at GameBattles
GameBattles
More
FAQ
Help Desk
MLG Store
MLG on XBox Live
Social Networking
MLG Wallpapers
MLG on ESPN
Friends of MLG
Sponsors
MLG Corporate







X
Username
Password
Remember Me
Forgot Password 
Home > Walshy <3's math
By Walshy | Published: September 13, 2009 9:35pm EST
I've been thinking a bit lately about stats after blogging about who I thought should (and ended up getting) Dallas MVP. What I've been thinking about specifically is if there is a better way in how they are calculated for determining MVP and just how they are compared to another individual in the tournament. It is impossible to ever make the stats perfectly show what someone contributed, they are without a doubt accurate but I feel like they could be shown much more accurately. A few ideas I thought of to possibly improve for example would be. First, I feel that calculating team slayer stats separate from objective. I think someone going hugely positive or negative over a tournament in team slayer would carry much more weight when separated from objective. Just being able to look at that stat alone I feel would be very interesting to know someone k/d ratio in ts compared to objective or being able to see average amount of kills, assists, and deaths in ts for each individual would be very comparable across the board since every team slayer game for the most part ends in 50 kills or close enough to that. Objective on the other hand I feel like wouldn't be as comparable across the board since the difference in game times. If you compared two different Pit CTF games stats you could easily end up comparing a 15 minute game where the highest kill count for one person is in the low 30's to a 7 minute game in which the kill count for person with the most is less than 15. A fix for this which I haven't tried out but thought it could be interesting/more accurate in comparing objective games would be taking the kills, assists, and deaths for the game and multiplying they by (15/minutes the game took). A 15 minute game would end up being the same stats, however a 7.5 minute game would end up doubling everyone's +/- ratio. Tell me what you guys think or what you think could possibly improve or more accurately compare stats across a tournament?
» Walshy's blog | login or register to post comments


Facebook | Tips N4G | Digg this
Post a comment
Posted by: Harmanicus
I'm pretty sure they ignore (09.13.09 9:43pm)

I'm pretty sure they ignore assists too...
login or register to post comments
Posted by: GlynTW
Everything- (09.13.09 10:01pm)

I would want to see every stat recorded for every player and every team. So that would be flag caps- grabs- objective time- assists- deaths- kills. YOu should be able to compare any player or any team on any gametype in the tournament. I hear what you are saying about the stats mutiplier, but i dont think that would work at all. But what you could have is recording the time it takes for teams to win an objective game and a slayer game. So like carbon would possibly have a reputation for closing out a narrows flag game extremely quickly every time on average.
login or register to post comments
Posted by: Seven_X
I believe MVP should be (09.13.09 10:01pm)

I believe MVP should be based on a multitude of things. 1. By watching VOD (someone might make game changing plays but it won't show up in the stat sheet.) 2. Stats (Assists, K/D Ratio, K/D spread, Amount of deaths, assists, kills during objective because sometimes although you didn't go most positive you were more in the action pushing up putting shots in and being disruptive especially during onslaught) 3. Compare stats to teammates 4. Establish a certain placing criteria (EX: Only players from top 3 are eligible to become MVP, but winning is not necessary)
login or register to post comments
Posted by: onsitecs
Would be nice to better define "MVP" (09.13.09 10:16pm)

Hey Walshy, I can see where you're trying to go with this and I'm a fan of getting a more "true" picture of a players performance. Not sure that the doubling of Ks and Ds would help ratios though. For example 1k/1d = 1 and 2k/2d = 1. The number of kills and deaths would be a more accurate representation of the pattern shown in the first 7.5 min but the overall ratios would still be the same. The ratio actually levels the playing field in a way for each players k/d performance but I'm still with you where it feels we are missing something when nailing down the players overall contribution based on the situation/gametype. Unfortuantely I'm not a huge fan of math but I'm always a fan of problem solving and improving something that appears to be sub-par. If a solution to the problem is not embraced by all maybe we'll at least be a little more satisfied. I'll think about it and let you know if I come up with anything. Let me know your thoughts.
login or register to post comments
Posted by: Walshy
Glyn- That's actually a (09.13.09 10:18pm)

Glyn- That's actually a better idea, having the average time taken on each gametype's separated by win/loss. Like Team A on average wins Onslaught flag in 8:30 and on average loses it in the time limit. The gametime could even be used for TS although I don't know what significance it would show, would be sweet to see though.
login or register to post comments
Posted by: GlynTW
YOu could even go further (09.13.09 10:28pm)

YOu could even go further with this uncle david, say for example on king and ball games. They could record the average win gaps for each of the gametypes. So if Team A wins 250-0 for the 3 times they play pit king (lol) their average obj/spread would be 250. SO teams could see that Team A was even more dominant than a 1-0 would show.
login or register to post comments
Posted by: GlynTW
SOrry if this is a double (09.13.09 10:56pm)

SOrry if this is a double post. But walshynator I have thought of even more stuff for objective games. YOu could record the objective -time gap spread for a team. Or do an average. So if a team dominates a certain ball game type, their average objective spread would be far higher than a team that manages to just scrape past or even loses.
login or register to post comments
Posted by: StrikerForce
Formula (09.13.09 11:36pm)

    (k + a) / d = ?
      With this formula it shows how much DAMAGE you actually did to the other team before you died, whether it be by assisting or killing. Lets say for example Defy had 400 kills, 450 assists and 500 deaths. Defy - 400 + 450 / 500 = 1.7 Now lets say Walshy went 375 kills, 500 assists and 500 deaths too. Walshy - 375 + 500 / 500 = 1.75 This shows that walshy, despite getting less kills proved too put more damage on the other team by posting a 1.75 enemies taken out per death as opposed to Defys 1.7 I think this is better then just the simple K/D spread MLG is using right now.
      login or register to post comments
Posted by: craneteam
[(k*2 + a*1.5) * % of team's (09.14.09 12:40am)

[(k*2 + a*1.5) * % of team's objective] / d this way kills take a more important role over assists since 40% of a series is TS, but assists are still given a good amount of weight
login or register to post comments
Posted by: FuzzyJAM
I think they should invest (09.14.09 12:43am)

I think they should invest in a good program to allow everyone to quickly compare stats between different players and work out various other things - kpd versus a single team, objective time for that gametypes, whatever. Raw data is nice but it's rather unwieldy. Another thing that would be really nice would be the ability to split stats from wins and losses. I want to see who really has the biggest impact on whether a team wins or loses, at least statistically. One thing I'd love that obviously can't be implemented is Shadowrun's damage counter rather than the simple assists and kills, that's an awesome aspect of the stats page and one I've never seen in another game. Maybe MLG could talk to Bungie about that for Halo Reach, perhaps even asking for LAN support of the bungie.net program that runs their stats, though I'm guessing that's stretching it. Of course, there's nothing that can be done to get over the ultimate problem which is that simply looking at someone's kills, deaths and objective time doesn't give you an excellent indicator of who's doing great, especially not in objective where, say, it doesn't even matter if you get out-slayed or even out-played for most of the game if you can hold on and pull off a great cap at the end.
login or register to post comments
123Next Page >
  • Win $10,000 in the Dr Pepper Collegiate Championship!
  • Orlando Online Qualifier Play Now Open -- Sign Up Now!
11.20.09 -
Shibby Got his MLG Hat at Lids -- Did you Get yours? (Photos Inside)
11.20.09 -
APG Featured in ESPN SportsNation Chat -- Read the Transcript Here
11.17.09 -
MLG Anaheim Halo 3 Statistics Released
11.16.09 -
The Chairman's Office - MLG Chairman and Co-Founder Mike Sepso Discusses MLG in Lids Stores
11.16.09 -
Major League Gaming Launches Licensed Headwear Line With National Retailer Lids
11.13.09 -
Watch the First Wave of MLG Anaheim Video on Demand
GameBattles is the largest online destination for competitive console and PC gaming. There are dozens of free ladders and tournaments running right now!
TOP ARENAS
Xbox 360
Call of Duty 4: Modern Warfare
Call of Duty 4: Modern Warfare
Xbox 360
Call of Duty: World at War
Call of Duty: World at War
Xbox 360
Gears of War 2
Gears of War 2
PlayStation 3
Call of Duty 4: Modern Warfare
Call of Duty 4: Modern Warfare
PlayStation 3
Call of Duty: World at War
Call of Duty: World at War
Xbox 360
Halo 3
Halo 3
2008 MLG Pro Circuit
Powered By: XBOX 360 Bungie
Subscribe to the Official MLG Newsletter here!
MLG Pro Circuit
  • Pro Circuit Info
  • Online Pro Circuit Info
  • Pro Teams
  • Pro Players
  • Official MLG Halo 3 Settings
  • Pro Circuit FAQ
GameBattles
  • GameBattles
  • GameBattles FAQ
  • Online Tournaments with Cash Prizes
  • Compete Online in Halo 3
  • Compete Online in Gears of War 2
  • Compete Online in Call of Duty 4
  • Compete Online in Rainbow Six Vegas 2
Video and News
  • MLG Video Homepage
  • ESPN/MLG Top 10
  • Dr Pepper Ultimate Gaming House
  • MLG Championship Sunday
  • ESPN Saturday Night
  • MLG Pregame Show
  • Tournament Game Footage
  • User Videos
  • Latest Features
  • Latest News
My MLG
  • Create Account
  • View Profile
  • Edit Profile
  • Upload Video
  • Post a Blog
  • Forums
  • Site FAQ
  • Help Desk
MLG Corporate
  • About
  • MLG Milestones
  • Media
  • Sponsors and Partners
  • Calendar of Events and Programming
  • Executive Team
  • Contact Us
  • Careers
  • MLG Gives Back
Get More MLG
  • MLG on ESPN
  • MLG on Xbox Live
  • MLG Merchandise
  • MLG Facebook
  • MLG MySpace
  • MLG on YouTube
  • MLG Playlist in Halo 3
© 2009 Major League Gaming. All Rights Reserved Terms of Use|Privacy Policy|Contact Info|About Us|Info for Parents|Newsletter