|
deetee posted:Does it really though, or is it hidden? When I was inactive for about 5 months, I got put back into a game with the same ranks as I left (Eagle Master) and when I won that game, I was assigned that rank again (my memory is a bit fuzzy, coulda be 2-3 games total). So it must remember. Most ELO systems I am familiar with operate a system of ELO 'decay' over time, but this one maybe not.
|
#
?
Mar 6, 2014 22:45
|
|
|
|
| # ? May 26, 2024 22:37 |
|
|
It just gets hidden, and it can come back within 1 game. Here is how I think the system works: You have your normal ELO bell curve, centered roughly around Master Guardian Elite/DMG. Each rank has a cutoff point where if your ELO number goes above or below that point you rank up or down. I'm also guessing that the reason that your rank can't immediately go up and down one match after the next is they calculate it in a method that is more similar to glicko, where you only recalculate ranks over a certain rating period (in this case ~3 matches). The results of all 3 matches factor into the calculation of your rank, but you won't see the results until after that 3rd match. edit: actually ELO already has a way to do results over a rating period instead of after every match so it doesn't have to be adjusted to look more like glicko astr0man fucked around with this message at 22:56 on Mar 6, 2014 |
|
#
?
Mar 6, 2014 22:50
|
|
|
DMG to Eagle seems to take forever. But it only takes two or three consecutive wins to get Eagle Master.
|
|
#
?
Mar 6, 2014 22:56
|
|
|
That would make sense because assuming your curve is centered on MGE/DMG then the range of ELO ratings you would need to get from DMG to Eagle would be a lot larger than the range you would need to go from Eagle to Eagle Master.
|
|
#
?
Mar 6, 2014 22:58
|
|
|
Yes yes standard deviations and other stats stuff.
|
|
#
?
Mar 6, 2014 23:00
|
|
|
astr0man posted:It just gets hidden, and it can come back within 1 game. Have we seen any statistics about the distribution? I'm surprised that it would be so high, I play regularly with a group of friends who are silver elite master at the very highest, and we always find matches quickly so there are tons of people in Silver. Maybe someday we will be Nova, but that's a ways off.
|
|
#
?
Mar 6, 2014 23:08
|
|
|
Twerk from Home posted:Have we seen any statistics about the distribution? I'm surprised that it would be so high, I play regularly with a group of friends who are silver elite master at the very highest, and we always find matches quickly so there are tons of people in Silver. Maybe someday we will be Nova, but that's a ways off. If there are 18 ranks, I guess the middle of the bell curve is at ranks 9/10, which are Gold Nova 3 and Gold Nova Master. edit: though I've literally never seen a player ranked lower than Nova Master, soooo Jeza fucked around with this message at 23:21 on Mar 6, 2014 |
|
#
?
Mar 6, 2014 23:19
|
|
|
I think that consecutive wins have an impact on elo gain. I mean given equal elos each game you would gain more elo if you win 5 in a row and lose 1 than if you won 2 lost 1 and won 3. not 100%, but it sure seems that way. Regarding the rank distribution, If you go in to your leaderboard, it shows a percentage for your rank. But I'm not entirely sure what the percentage means. I would assume it means what percentage of people are at your elo, but that doesn't seem right because my two badge accounts are at 22.5% and 7.3% and my eagle account is at 68.0%. I have a silver II account at 17.4% So I'm not entirely sure what the percentage means... Apogee15 fucked around with this message at 23:38 on Mar 6, 2014 |
|
#
?
Mar 6, 2014 23:28
|
|
|
I don't think the ranks are necessarily distributed evenly. A month or 2 after the game launched Valve adjusted the distribution because they decided there were way too many players in the top few levels. edit: Apogee15 posted:I think that consecutive wins have an impact on elo gain. I mean given equal elos each game you would gain more elo if you win 5 in a row and lose 1 than if you won 2 lost 1 and won 3. Even if everyone's ranks aren't the same, the order doesn't matter, just the results. I.E. if my ELO rating is 1500, and I play 5 matches against people rated 1450, 1475, 1500, 1525 and 1550 and win all of them except the one vs the 1550 guy, it doesn't matter if I played the 1550 guy first, or in the middle somewhere, or last. My ending ELO will still be the same. edit2: What would affect it is if my rating was recalculated somewhere in the middle of those matches, since the expected outcome of the match vs the 1550 guy depends on whatever my rating is, so yes it is possible for the order to matter, but there isn't an inherent bonus to winning or losing a bunch of matches in a rowt. astr0man fucked around with this message at 23:48 on Mar 6, 2014 |
|
#
?
Mar 6, 2014 23:29
|
|
|
http://www.twitch.tv/swampfox89/b/508794603 Probably one of the most personally rewarding games I've played in a while. They were kinda dickish in the first half (as T on Nuke no less) and so after we brought it to 6-9 before taking T-side for the second half it felt good to pull out that win. "AAAAHAHAHAHAHAHAHA YOU loving SLUTS" e: I didn't record it, that was someone else. FAUXTON fucked around with this message at 04:41 on Mar 7, 2014 |
|
#
?
Mar 6, 2014 23:37
|
|
|
Gristley Bear posted:DMG to Eagle seems to take forever. But it only takes two or three consecutive wins to get Eagle Master. I think part of this is who you're paired with/against too. At DMG you see a lot of AK/Double AK, and at Eagle you see far less, from my experience.
|
|
#
?
Mar 6, 2014 23:42
|
|
|
Mimetic posted:Just deliberately abandoned a game of MM for the first time. Someone ditched the bomb on the pistol round and we had to go back and get it. We ended up losing and the team starts talking about how we've lost the game already. Then once we get guns one guy starts smoking our side of mid every round to gently caress with our AWPers. The last straw was when I got a kill long, start to push up and die to find that my team has turned around and run off. Meanwhile our bomb carrier is on the other side of the map trying (and failing) to chuck the bomb out of bounds. I had a game a while back where we lost the pistol round, in a really unremarkable, completely standard way. Then we lose the eco round, again in a really standard way. Nothing stupid happened at all, nobody made any egregious mistakes, it was just run of the mill CS. After losing the eco round one guy on my team just lost his poo poo and started freaking out, calling everyone terrible noobs, and so on, and promptly went AFK for a few rounds and got kicked. If that set him off, I can't imagine what he'd be like in games where something legitimately stupid happened.
|
|
#
?
Mar 6, 2014 23:44
|
|
astr0man posted:there isn't an inherent bonus to winning or losing a bunch of matches in a rowt how can you know this?
|
|
|
#
?
Mar 6, 2014 23:49
|
|
|
astr0man posted:
I'm not saying it definitely does work that way, but you are wrong in saying "Elo doesn't work that way" because it most definitely CAN work that way. It wouldn't be at all hard for them to use a modifier based on wins in a row. It ALL depends on what formula they use, and no one knows what formula they use. Apogee15 fucked around with this message at 23:55 on Mar 6, 2014 |
|
#
?
Mar 6, 2014 23:53
|
|
|
No I do not know what valves formula is. But yes I do know how ELO works and that is what I was explaining. And ELO does not give a poo poo if you win consecutive matches. If you are always playing an opponent who is at the same rating as you, the expected outcome is always .500 (a tie) and your ending rank after a fixed record (like 4-1) will always be the same no matter what the order was.
|
|
#
?
Mar 6, 2014 23:57
|
|
|
astr0man posted:No I do not know what valves formula is. But yes I do know how ELO works and that is what I was explaining. And ELO does not give a poo poo if you win consecutive matches. If you are always playing an opponent who is at the same rating as you, the expected outcome is always .500 (a tie) and your ending rank after a fixed record (like 4-1) will always be the same no matter what the order was. Except not all elo systems are the same, and they could VERY easily make consecutive wins matter. It's not like there is an official standard which says "All elo systems must operate in this way". It wouldn't be at all hard to introduce a system which modifies your elo gain per game. For example: With 1 win, you could have a modifier of 1, with 5 wins you could have a modifier of 1.2. So if you would have gained 20 a win, you would gain 24 elo if you have a win streak of 5. It's entirely possible to introduce a system like that, and it would still be an elo system. Apogee15 fucked around with this message at 00:02 on Mar 7, 2014 |
|
#
?
Mar 6, 2014 23:59
|
|
|
down with slavery posted:how can you know this? I guess the question that comes to my mind is why they would implement a system that gives extra weight to consecutive wins. Like, that doesn't display superior skill at all. 5 wins in a row then a loss is no different to 2 wins, a loss, then 3 more wins...right?
|
|
#
?
Mar 7, 2014 00:01
|
|
Jeza posted:I guess the question that comes to my mind is why they would implement a system that gives extra weight to consecutive wins. Like, that doesn't display superior skill at all. 5 wins in a row then a loss is no different to 2 wins, a loss, then 3 more wins...right? To let accounts that are terribly mismatched to quickly adhere to the proper skill level? I'm sure this is what happens during the "provisional period" (your ELO is hidden and *probably* moves easier)
|
|
|
#
?
Mar 7, 2014 00:09
|
|
|
down with slavery posted:To let accounts that are terribly mismatched to quickly adhere to the proper skill level? I'm sure this is what happens during the "provisional period" (your ELO is hidden and *probably* moves easier) ELO should already account for terrible mismatches without needing to inflate any of the results. edit: there's also better and more common ways to handle the uncertainty of where a new player should be ranked, such as just using different K factors based on however much experience a player has. Apogee15 posted:It wouldn't be at all hard to introduce a system which modifies your elo gain per game. For example: With 1 win, you could have a modifier of 1, with 5 wins you could have a modifier of 1.2. So if you would have gained 20 a win, you would gain 24 elo if you have a win streak of 5. It's entirely possible to introduce a system like that, and it would still be an elo system. Also, again, I don't know what valve is doing! And they obviously aren't using a standard ELO system because a standard ELO system doesn't work for rating individual players in team games in the first place! astr0man fucked around with this message at 00:30 on Mar 7, 2014 |
|
#
?
Mar 7, 2014 00:18
|
|
|
I don't believe that there is a single 'legit' silver 1 out there. I've tried. Even I can't throw that hard.
|
|
#
?
Mar 7, 2014 00:51
|
|
|
ShittyPostmakerPro posted:I don't believe that there is a single 'legit' silver 1 out there. I've tried. Even I can't throw that hard. I have downtime at work, and lately I've been trying to get silver 1. I'm at silver II. It IS hard. I've been trying to reach silver 1 for weeks now. Sometimes I just can't throw hard enough(can't throw too hard or you get kicked, and too many kicks=temp ban... gotten a week long ban several times now). astr0man posted:Yes you could do that but in my opinion doing so would be really dumb because you are adding extra weight where you don't need any. For example, the system already expects a global elite player to push my poo poo in every match, so why would his rank need to be boosted even more for pushing my poo poo in 5 times in a row? Likewise, if a silver 1 gets lucky and beats a global elite player 5 times in a row that is great for him I guess, but the whole point is that over hundreds of matches the silver 1 will lose 99% of the time is already ranked in the right place. I think if you win 10 times in a row, it's a decent indicator that you aren't at the elo you should be at, whereas if you win a few, then lose one, then win a few, then lose one then maybe you aren't so far off what elo you should be at. But you're right that it's really not necessary to do that. Players will eventually reach the appropriate elo either way. Anyways, the only reason I suspected it was because I had a silver III account which I started playing seriously on, and ranked it up to badge. I won probably 15-18 games in a row and it seemed like I was ranking up faster and faster(starting around 4-5 wins for the first rank up, and I think I ranked up to MG2 in 2 wins). Then I lost one, and it took another 4 wins or so to get the double ak. Totally anecdotal, and doesn't prove anything at all. But it did make me suspect. It could have just been all a coincidence. Apogee15 fucked around with this message at 01:02 on Mar 7, 2014 |
|
#
?
Mar 7, 2014 00:59
|
|
|
Winning 10 times in a row may or may not be an indicator of anything because it depends on the other team's rank.
|
|
#
?
Mar 7, 2014 01:08
|
|
|
That's true, but I think matches with very uneven elo distributions are outliers more than the rule, so you can start with the assumption that on average you will be playing people roughly around your elo. Especially since the matchmaking system specifically tries to find games with players around your elo, and only widens the search after it fails to find one after a certain amount of time. You're not likely going to have 10 matches in a row of a bunch of AKs vs a bunch of novas.
|
|
#
?
Mar 7, 2014 01:13
|
|
|
MMing today with Rah, Beach, and a few others on cache. This conversation happened multiple times, except replace K/D with aim. Also nobody could manage to aim, that's an important detail:
|
|
#
?
Mar 7, 2014 01:14
|
|
This poster loves police brutality, but only when its against minorities!
|
If it matters I have ranked up after winning matches where I played like rear end
|
|
#
?
Mar 7, 2014 01:18
|
|
|
astr0man posted:ELO should already account for terrible mismatches without needing to inflate any of the results. What does E.L.O stand for? I'm just asking because you clearly know a lot about it.
|
|
#
?
Mar 7, 2014 01:24
|
|
|
Salt Fish posted:What does E.L.O stand for? I'm just asking because you clearly know a lot about it. It's named after Arpad Elo, who invented it for chess. It's not an acronym.
|
|
#
?
Mar 7, 2014 01:25
|
|
|
http://en.wikipedia.org/wiki/Elo_rating_system
|
|
#
?
Mar 7, 2014 01:30
|
|
|
my friend managed to get silver 1 through regular games
|
|
#
?
Mar 7, 2014 01:32
|
|
|
hello internet posted:If it matters I have ranked up after winning matches where I played like rear end If it matters, others who are higher ranked than me (Gold Novas for example) have ranked (to higher Gold Novas) after getting carried by me while they went negative/decimal and winning a game. In the mean time, my rank is at a stand still (Silver Nova Elite Crazy Person).
|
|
#
?
Mar 7, 2014 02:01
|
|
|
quote:- Dust2: gently caress OFF VALVE, DON'T gently caress UP YOUR FEW GOOD MAPS. I loved picking people taking A from Cat looking in between those boxes. Valve always changing poo poo that doesn't need to be changed and loving things up. Seriously gently caress off and stop trying to balance things you don't know poo poo about. Can't they release a patch that fixes KNOWN things without changing some other random poo poo too?  loving Valve, dude.
|
|
#
?
Mar 7, 2014 02:04
|
|
|
Dvsilverwing posted:gently caress OFF VALVE, DON'T gently caress UP YOUR FEW GOOD MAPS. That's an unnecessary balancing move. The slit is narrow enough that you can't really use it in a game-breaking way, and if someone sees you shooting through it then you give away your position. My favorite defensive spot is gone and for no good reason. 
|
|
#
?
Mar 7, 2014 02:31
|
|
|
Ernie. posted:That's an unnecessary balancing move. The slit is narrow enough that you can't really use it in a game-breaking way, and if someone sees you shooting through it then you give away your position. My favorite defensive spot is gone and for no good reason. My point exactly, what a bunch of gently caress-ups. Edit: I just don't understand, I don't get it. How can they not just fix poo poo that actually bothers people without trying to change other stuff (which 95% of the time winds up pissing people off and breaking something). I understand it's hard to fix certain issues, but it can't be that hard to not randomly break poo poo. Dvsilverwing fucked around with this message at 02:35 on Mar 7, 2014 |
|
#
?
Mar 7, 2014 02:33
|
|
|
Ernie. posted:That's an unnecessary balancing move. The slit is narrow enough that you can't really use it in a game-breaking way, and if someone sees you shooting through it then you give away your position. My favorite defensive spot is gone and for no good reason. It's still there, it's just less offensive, and more defensive now. You can't get killed there from cat anymore, and you can't kill anyone from cat anymore(at least not while you are crouching). I am going to miss that spot in some ways though. It was always fun to just spray that gap and get a random kill.
|
|
#
?
Mar 7, 2014 02:50
|
|
|
Apogee15 posted:It's still there, it's just less offensive, and more defensive now. You can't get killed there from cat anymore, and you can't kill anyone from cat anymore(at least not while you are crouching). It was defensive, it was a decent spot to play as a CT because you could occasionally get a kill or two through it, especially if you're using the M4A1-S. Lets see if they can make a T sided map even more T sided. They HAVE to be loving poo poo up on purpose at this point, I think they continually gently caress poo poo up in CS:GO so that people that don't own CS:S will buy it to play when they get fed up with CS:GO and it being ruined. That's why CS:GO is only $14.99 but CS:S is still $19.99. It's a god damned conspiracy, and we're getting the shaft.
|
|
#
?
Mar 7, 2014 03:06
|
|
|
Valve is one more box adjustment away from getting this thread its very own "they tore away my ability to feel human" moment. I'm pretty surprised to see people getting so passionately angry about a minor change to such an awful map strategy-wise. Isn't the whole reason that you queue up for de_dust2 is because you don't want to have to use strategies?
|
|
#
?
Mar 7, 2014 03:14
|
|
|
Maarek posted:Valve is one more box adjustment away from getting this thread its very own "they tore away my ability to feel human" moment. I'm pretty surprised to see people getting so passionately angry about a minor change to such an awful map strategy-wise. Isn't the whole reason that you queue up for de_dust2 is because you don't want to have to use strategies?
|
|
#
?
Mar 7, 2014 03:19
|
|
|
Challengers Stickers Legends Stickers
|
|
#
?
Mar 7, 2014 03:21
|
|
|
Coleman posted:...this post literally makes me want to hit things. Just because all the strategies for dust2 are well known and popular does not mean it's not a strategic map. The map is just as strategic as every other map (if not moreso, there's a reason it's so popular you know?), it's just that it's played so much that all the strategies for it are well known. Yet no one knows how to play it 
|
|
#
?
Mar 7, 2014 03:29
|
|
|
|
| # ? May 26, 2024 22:37 |
|
|
Dust2 is so popular because of all the decent maps that are played competitively it requires the least communication and teamwork. Most people don't CS play in organized groups and dust2 is by far the easiest map to play with 5 random shooters on a team.
|
|
#
?
Mar 7, 2014 03:30
|
|