DUST Search - General Kill Stats

MlDDLE MANGEMENT wrote:

Thank you for telling me the mean, median and mode of the the analysis.

How bout you tell me the power of the study, how about whether or not your conclusions should lead me to accept or reject the null hypothesis, what bout that alpha and the beta of the study. When you are done doing these basic parameters would you gauge your analysis as a retrospective or prospective and what will you use cohort or case control analysis meaning is this subject to an odds ratio, chi square, weighted t-test, maybe an anova?

All you have shown is that the AR is the most widely seen weapon on the killfeed, this actually tells us nothing on wheter or not its actually the most widely used weapon in the game. Logi LAVs could in fact be the most used weapon in the game, or Remotes or Even Scramblers, it doesnt take into any account dmg done over time by all other weaponry it doesnt even tell us wheter those AR kills were from pure AR dmg or a stolen kill.

Even if it did, its still based off of pub data which is a whole other set of fail.

Please stop or dont and allow us to continue to point and laugh at you.


Author	Thread Statistics \| Show CCP posts - 0 post(s)
Django Quik R.I.f.t 920	Posted - 2013.07.13 23:09:00 - [1] - Quote Firstly, I want to apologise to anyone who played a game with me in today because I was AFK while I was doing this for 10 games - I don't do it often but at least this time I had good reason. Okay, so what I did was fairly simple: I just tallied up every cause of death for every kill of anybody [/b]on either team over the course of 10 matches. This was during a Saturday evening GMT (I plan to repeat the process at various other times during the day/week in the future to get a wider overview). These were the results and I've given some analysis below: Assault Rifle - 379 Sniper Rifle - 136 HMG - 112 LAV - 83 Mass Driver - 73 HAV - 67 Scrambler Rifle - 52 SMG - 51 Precision Strike - 46 Forge Gun - 39 Locus Grenade - 37 Swarm Launcher - 26 Shotgun - 17 AV Grenade - 12 Flaylock - 10 Installation - 9 Scrambler Pistol - 7 Remote Explosive - 7 Plasma Cannon - 6 Dropship - 6 Laser - 1 Melee - 1 Flux Grenade - 1 So, it's not great surprise that ARs were top of the list by quite some margin (and by far the most common AR kill was with a GEK-38) but personally I very rarely get sniped, so I wasn't expecting that value to be quite so high. Then there was a big surprise that the HMG was so high up because heavies are constantly complaining about how terrible it is to be them these days (on a related note the FG was also surprisingly high up the list). The LAVs is a big issue really (although a handful of those kills were actually with the LAV turret but most were by running people over). Seriously, the 4th biggest killer on the battlefield over 10 games was hit and runs!? Also of note is how high up the SMG is and how low the Flaylock is - seeing as how many people have been complaining about the 'Godlock' pistol being OP, I was expecting this to be high up if not only the best sidearm. Finally, another couple of expecteds at the lower end the shotgun, laser rifle and plasma cannon. Fortunately, I know that CCP collect this sort of data for all matches and so are very well aware of the kinds of numbers on a much much greater scale - so with any luck they know the problem areas and are planning on sorting them all out fairly soon (TM)... - I'll add more data as and when I collect it -
Django Quik R.I.f.t 920	Posted - 2013.07.13 23:17:00 - [2] - Quote ladwar wrote: so i killed you with my laser? sounds great! Not me - I was tallying everyone's kills over the 10 matches. I was AFKing while counting all of it, so had no deaths/kills at all.
Django Quik R.I.f.t 924	Posted - 2013.07.14 08:41:00 - [3] - Quote Master Jaraiya wrote: This is all nice and fine and all, but really proves nothing as to what weapon is over powered or under powered. It simply tells us what the most popular weapons were in the matches that you recorded these statistics. There are simply too many variables to conduct a test as simple as this. For more information see my thread here. For the record i was not trying to use there stats to show what is op or up. This was just to see if there were any interesting correlations with expected results and there were. The fact the ar came out on top means nothing really but the hmg also being quite high is significant and shows it's possibly not quite as useless as people have been making out - possibly. Also 10 games is quite a lot and covered every single map at least once but like i said i plan to redo this process at other times to get broader results and avoid the possibility of having the same people in every match skewing the data. Also, the data has to be analysed and interpreted to make any useful conclusions and that's the hard part. You are quite correct that this shows what weapons are most popular but the next step is to figure out why they're most popular. The ar we might saw is because it's the most instantly familiar to fps players but it could also be op - the stats alone can not tell us that. However, 1 kill with a laser rifle indicates that almost no one uses it and the only conclusion must be that it's useless. Even the plasma cannon got more kills and people claim that that's terrible too! As for the flaylock, that was my most intriguing result and I'll be very interested to see if the trend continues in my next few sessions. Thanks for the constructive feedback though - i look forward to discussing this further soon.
Django Quik R.I.f.t 930	Posted - 2013.07.14 16:34:00 - [4] - Quote Updated for today's stats and added new analyses. GET ATMESON wrote: I-Shayz-I wrote: I have done this a couple times and have gotten similar results. However I did mine before the 1.2 update so shotgun was much higher on the list. I'm thinking that the reason flaylocks and HMG's are skewed on this list is because of a few reasons. First of all, HMG's and flaylocks are only good for players that have skilled way up into them and are either stacking damage mods, or suing some other method of making the weapon more useful. Also, if you get a good HMG player in the match they can easily get 20-30 kills and skew the chart substantially. Flaylocks aren't used as most as other weapons because they require a special setup for them to be effective. The problem is, they're too effective and annoyingly so. Ding ding ding. This guys a correct. +1 Unfortunately I don't think so. At least about the HMG thing. The HMG kill count was pretty consistent in most matches with only one of the 20 matches so far reaching over 20 kills. I'm still unsure as to why the flaylock gets so much OP whine but it certainly isn't getting the kills to warrant it
Django Quik R.I.f.t 933	Posted - 2013.07.14 16:48:00 - [5] - Quote gandalgrey wrote: I love stats guess im jus a stat geek would love to see my own Hehe yeah, love me stats too but god damn it gets dull just reading the kill feed and counting! The analyses are the fun part :)
Django Quik R.I.f.t 935	Posted - 2013.07.14 17:04:00 - [6] - Quote Banning Hammer wrote: 2293 kills in 10 matches ? It is something wrong with you math, or you been playing a different game. By the way... that is over 200 kills per battle .... Well it's kills on both sides, so that could go up to 300 but also I've updated for a 2nd set of 10 matches now, so that's 20 matches.
Django Quik R.I.f.t 935	Posted - 2013.07.14 17:05:00 - [7] - Quote RevoL Frog wrote: Do you even know how to count? When has 41 ever gone between 32 & 21? Maybe an unconscious bias on your part? Ah you are quite correct - no need to be an ass about it though.
Django Quik R.I.f.t 935	Posted - 2013.07.14 17:07:00 - [8] - Quote Rynoceros wrote: 87.3% of all stats are false. Only 4 Knives? Not that effective, is it dip****s? Get a real weapon. My scouty buddies have been telling me that there's hit detection problems with the knives lately, so that may be why they're not featuring much.
Django Quik R.I.f.t 978	Posted - 2013.07.14 23:29:00 - [9] - Quote Banning Hammer wrote: pegasis prime wrote: Banning Hammer wrote: I know what he means .. still doesn't adds on. That is an average of 4 kills per person in both teams, all 32 players. Take away the AFKers, the noobs, and snipers.. which is about half of the team... That leave you with 8 people in each team getting an average of 8 kills each. Most battles ends in red-lines, with the top players with only about 8-12 kills... so.. you see.. doesn't adds on. Your experience of dust seems crap. I can vouch for severall of tge kills django has counted as he was in several games against my corp mates and myself . I was there for several games and none of them were as you describe. And yes the kill feeds were going crazy. Possibly due to mordues chalange I havent seen anyone afk appart from django and he was logging kills. Even if the top 3 players in a team gets, lets say; 30 kills / 15 kills / 10 kills ( 55 ) I seriously doubt that the rest of the players will get another 55 between them... maybe in a VERY good match 70-80 kills in total, but definitely not over 100.. and specially not 10 battles in a row. You must be trolling, right? 100 kills total each match is easily understandable. That's 50 per team. Even if 6 people from each side get no kills, that's still am average of 5 per player. And actually i did also keep track of the state of each game - around half were red lined by the end, a quarter comfortable victories and the rest fairly close. Of the red lined games, the red line was seldom early and even once red lined in not one game did the losing side not at least attempt a comeback. Further to that i remember the number of people afk with me on all there occasions as it was also interesting to see how that affected the games. During the first 10 matches there were very very few afk, only 1 in a couple of matches. The 2nd 10 had a lot more afk with up to 3 on my side in a couple of games but not as many as to stop plenty of kills from still happening please play some games and let me know how many you actually see with less than 100 kills.
Django Quik R.I.f.t 978	Posted - 2013.07.14 23:33:00 - [10] - Quote Wait a minutes - i just re-read your last post and realised something - you know Im counting kills from both teams, right? So 50 per team on average. Not just my team getting 100 on average. Read this all through really carefully and then think before you post something silly again.

Django Quik R.I.f.t 989	Posted - 2013.07.19 13:46:00 - [11] - Quote Updated for the 3rd set of 10 matches I recorded data for today - that brings the total to 30 games worth of kill stats.
Django Quik R.I.f.t 998	Posted - 2013.07.19 18:53:00 - [12] - Quote I wish to completely distance myself from Buster's comments in this thread, as they do not reflect my own views or conclusions. The purpose of this data was never to show that anything was OP or not and I don't believe any of this data alone could possibly show this. It is quite obvious that the AR is the most used weapon and gets the most kills - that does not infer it is OP. Likelihood is that many people use it because it is the most familiar weapon to them after playing other FPSs. howard sanchez wrote: Django, you've done our community a great service by attempting to provide objective statistical data for fair minded analysis. Thank you. Your efforts should be recognized, if not rewarded. And while I agree that it is likely CCP gathers and maybe even tracks this kind of information, the fact that they choose not to publish or share it as the basis for reasonable discussion and transparency with the player base only serves to keep opinions, hyperbole and emotionally charged rhetoric as the guiding force behind player feedback. My point was only ever to provide some data. I was a little irked by a few people recently publishing lists of the weapons that kill them over a number of games because these all relied on a single player's playstyle - this data is a generalised version of those attempts and no one else has provided this information previously. Dr Allopathy wrote: You say you included the third data set, but you failed to include the 50 flux grenade kills. Link to thread: https://forums.dust514.com/default.aspx?g=posts&t=95744 Edit: As I stated before, you are not recording enough matches to have enough data to record accurate stats here. 10 games out of thousands is not enough I realise my dataset is limited but I will continue to expand it over time. 30 matches and thousands of kills is a good start. The reason I don't have those 50 flux grenade kills is that none of those happened in any of the matches I happened to be watching - i.e. I can't collect data for things I don't observe. bethany valvetino wrote: nice stats... But honest, this either the worst excuse to cover up AFK MCC camping, or you just have WAY TO MUCH time on your hands... It's summer, might I suggest you go out side or perhaps... Get a fking job, you tramp. Heh, yes, this does mean I spend a few hours doing nothing in MCCs but damn it's frustrating and dull. I sometimes wish I could go down and help out, especially when I can see on the map exactly where that sniper on the killstreak is hiding or when I know I could get that LLAV buzzing around murder-taxiing everyone but I resist, so that I in no way skew the results. If I was playing, you'd see a lot more pistol kills. But it is boring and I don't get to do anything else whilst doing it because I literally have to be watching the killfeed the entire game or I miss stuff - it's hard as it is when there's 5+ deaths within seconds and I have to mark them down before they disappear because only the last 3 show at once. I've been unemployed for well over a year now and have been desperately seeking work, whilst doing my own games design projects. Fortunately, I have an interview tomorrow and another 2 on Wednesday, so I won't be in this state much longer (fingers crossed). I hate feeling like a bum :(
Django Quik R.I.f.t 999	Posted - 2013.07.19 19:02:00 - [13] - Quote Further to my above comments - I also realise the limitations of this data in that it only represents pub matches and so can not and should not be taken as representative of high level or competitive play. This data should only be taken as general information on pub matches and not a complete overview of the game or weapon balances. With that said, some important conclusions can be drawn this data that are soundly backed up by many anecdotal reports from players throughout these forums and ingame chat channels - notably the low kill rates of previously well used weapons, such as shotguns (though this isn't looking as bad with this latest additional data) and laser rifles and weapons that clearly are not filling any role, like the plasma cannon.
Django Quik R.I.f.t 999	Posted - 2013.07.19 19:08:00 - [14] - Quote MlDDLE MANGEMENT wrote: Thank you for telling me the mean, median and mode of the the analysis. How bout you tell me the power of the study, how about whether or not your conclusions should lead me to accept or reject the null hypothesis, what bout that alpha and the beta of the study. When you are done doing these basic parameters would you gauge your analysis as a retrospective or prospective and what will you use cohort or case control analysis meaning is this subject to an odds ratio, chi square, weighted t-test, maybe an anova? All you have shown is that the AR is the most widely seen weapon on the killfeed, this actually tells us nothing on wheter or not its actually the most widely used weapon in the game. Logi LAVs could in fact be the most used weapon in the game, or Remotes or Even Scramblers, it doesnt take into any account dmg done over time by all other weaponry it doesnt even tell us wheter those AR kills were from pure AR dmg or a stolen kill. Even if it did, its still based off of pub data which is a whole other set of fail. Please stop or dont and allow us to continue to point and laugh at you. I'm assuming this isn't aimed at me but if it is, please understand I only serve to provide the data and some very light analysis - there is clearly not enough and not deep enough information here to be able to draw any deeper meaning. This is however the deepest study publicly available on these forums and no one else has provided anything near this depth of data and those other things you ask for are simply impossible for anyone other than CCP to gather or to publish.
Django Quik R.I.f.t 1002	Posted - 2013.07.19 21:30:00 - [15] - Quote Terri Ashra wrote: Darth-Carbonite GIO wrote: Quote: I know what he means .. still doesn't adds on. That is an average of 4 kills per person in both teams, all 32 players. Take away the AFKers, the noobs, and snipers.. which is about half of the team... That leave you with 8 people in each team getting an average of 8 kills each. Most battles ends in red-lines, with the top players with only about 8-12 kills... so.. you see.. doesn't adds on. I think you may be underestimating our noob community. Put a few monkeys with typewriters together and they will "eventually" write the Iliad. Simple probability. AFKing has gone down sharply in the past few days, so I wouldn't put much stock in that either. We've heard that a million monkeys at a million keyboards could produce the complete works of Shakespeare; now, thanks to the Internet, we know that is not true. Robert Wilensky Lol - I enjoyed this comment greatly. The important point though is that you should not remove the above listed players from the stats (except for AFKers) - noobs still get kills or more aptly get killed (still contributing to total kills!) and snipers, well just look at the stats; snipers are the 2nd top killing weapon in the game!
Django Quik R.I.f.t 1066	Posted - 2013.07.26 07:17:00 - [16] - Quote Zsiga What wrote: A better more effective statistic would be the amount of kills for each weapon type divided by the number of people who are carrying the weapon. This will better inform the OPness of the tool. Everyone carries grenades so comparing flalock kills to grenade kills is like comparing the number of murders in the US to the number of murders in Brazil. Your logic isn't logical. Yes, that would be a better statistic but that's also impossible for me or any other player to get. Also, read my other posts in this thread responding to similar comments - this data is very limited and was never intended to be used as a tool to show OPness . My analyses also make no effort to imply that anything is OP and simply state areas where these basic facts show possible issues, such as lasers and plasma cannons and that it's pretty stupid that LAVs get more kills than most other weapons in the game. My logic is sound but you are implying bad logic in my post that has never been stated by me.
Django Quik R.I.f.t 1066	Posted - 2013.07.26 07:30:00 - [17] - Quote Cere Harkens wrote: I am in the market for a new car and want the best I can get. Inspired by this thread, I went to the street and counted the first 800 cars that passed by. (I was going to count to 1,000 but I got bored.) In order of frequency, the cars were: ... My buddy has an Audi R8, but that thing didn't even make my list. It must be a piece of crap! Ford Focus is clearly the best car on the road, so that's what I'm buying. That's great. Good for you. However you are missing some important points. First read my response above this and stop trying to imply things in my stats and analyses that were never even remotely mentioned. Secondly, in real life there are thousands of car makes, in dust there are only a couple of dozen types of weapon. In your example you've also counted specific car models- a better comparison would be car manufacturers. When collecting data like this it only becomes useful when combined with other data or knowledge. We know the shotgun has been a good and popular weapon in the past, so if it doesn't get many kills, it suggests there's a problem making people either not use it or struggle to kill with it. If you don't see a lot of ferraris on the street, you don't just assume it's because they suck - you have to look deeper to figure out that maybe the reason they're rare is that they're really expensive or especially niche a bit like oh let's say tanks in dust.