I'd like to add that i tried to intentionally feed and still got calibrated almost 4.3k
how cute is that? System is broken af in my opinion.
WHO GIVES A FUCK AND DO YOU REALLY THINK WE DON'T KNOW HOW IT WORKS YOU SAD PIECE OF SHIT? WE KNOW YOU FUCKTARD WE KNOW ITS JUST YOU FUCKING SPAMMERS AND FAIL FUCKING SMURFERS THAT DONT KNOW SHIT NOW GTFO AND STOP SPAMMING OUR FORUMS!
Very helpful..
I am gonna calibrate soon..
So thanks...
Also dont be mad when someone tries to help... Thats just mean
Dude but how much mmr did u lose when u intentionally feed 10 calibration games in a row? A lot
Dotabuff
Esports
Heroes
Items
Players
Matches
Blog
Forums
Plus
mid or feed
General Discussion
General Discussion
My understanding of TBD & Calibration - related to 'getting out of th...
My understanding of TBD & Calibration - related to 'getting out of the 3k trench' in General Discussion
•)
•)
6 hours ago
First and foremost, everything said is my opinion. I've done a fair amount of experimentation to be able to draw some conclusions.
1. First X unranked games you play
Within your first ~30 games (I believe abandons do not qualify here - if you abandon with 0-20, it shouldn't influence your overall stats (KDA, GPM...)) your rating is more or less decided. You can calibrate upto 5k, and be matched in a 4.5k-5k average game. On this dotabuff there's an example of a 6k, 4.9k, 4,5k x2, and me (probably at 5000 virtual MMR at that point because I previously had a 24 KDA game at 4.5k average mmr). This means that there are brackets even in the 'VHS' and that keeping your account in good shape may result in having the first calibration game at the average of 5k mmr, which is the maximum.
"More or less decided" means that after your 30 games, your virtual MMR will stop spiking. Between my 6th and 8th game, my virtual mmr has dropped by almost a thousand (3900-4100) and in the 9th game the average mmr was again around 4.3k.
2. Stats
Everything is taken in consideration, even the sentries and dewarding, maybe even TP scroll usage. I think most important is 'kill participation' - even this stat I think is unavailable (?) on dotabuff, I believe is used as a multiplier in the formula. KDA surely is very important, as well as GPM/XPM compared to your teammates. Hero healing, tower damage, and LH+D per min.
3. Should you make a smurf to recalibrate?
I played some 50 games in solo matchmaking at 3300MMR (I play dota on and off, mostly party at 4k+) because I never took the time to boost the MMR, since it's what I calibrated at when it was introduced in beta(?, whatever).
I found that the game winning plays in that bracket are not necessarily the correct plays. I don't know to win a 3k game, other then feel I'm going to literally 1v5 them with a hero like slark or tinker, or just crush them in a teamfight with a massive AM or sven. But I like heroes like WR, QoP, TA, SF, weaver... and they don't win by themselves & they exploit tempo. You can't exploit tempo in a 3k game, because when you wipe their team, instead of doing dmg to a t3 your teammates go farm the first component of their fifth item or smth.
General lack of enthusiasm about winning is pretty much my feeling of the 3k. A round of CS:GO, literally - I don't want to play that game. I cannot improve in that game....
4. Things to know when playing unranked:
I like to think of them this way: I imagine that I'm playing with a hand of midas: 1/0/1 is much better than a 3/1/0 especially if I'm mid or even offlane. I'd rather spend time farming than searching for kills, but being in the right place for the fight is effectively the best place to farm.
By farming, I mean farming your stats: kill participation, hero damage GPM, XPM, LH, wards, dewards, sentries, camp stacks, etc. Kills are not better than assists in any sense other than that kill means more gold, but I think it's better to try do most most hero damage as possible.
Hit towers when you can, instead of running around. Tower damage counts too, especially if your team hasn't done much damage to towers. I think raxes count too but I can't be sure.
Apparently teleports count too, so tinker is a really good choice - if you're good on it, you can create a gap in your team kill participation and that's what will give your virtual mmr a boost. (if your teammates don't participate in kills, the game recognizes you as doing pretty much everything except for dewarding and herohealing (DP soul siphon counts as hero healing for example).
I'm 90% sure overall hero versatility (the graph as a whole too) counts a lot, and any 'red points' on the graph might negatively influence your mmr before calibration, and a well balanced graph might boost it.
5. Scrubs shouldn't recalibrate, but climb their way out of 3k?
TBH, if the max rating of calibration is 5k, then whoever calibrates at 4k can't be more terrible than some of the terrible players who are at 4k already. I just think that a lot of people who suddenly fall into the 4k bracket get too cocky, want to play mid all the time and eventually lose because of miscommunication and flaming. I also believe that unless you belong to a higher bracket, you should be able to play 3-4 positions in all types of strategical variations if you are to maintain your mmr. If you can play 20 carries and 20 mids, but 3 supports and 5 offlaners youre fucked, and you shouldnt be wasting your time anyway.
.
.
5 hours ago
id like to add that calibration games can swing ur average mmr games by over 1000 points.
Good Riddance
Good Riddance
4 hours ago
I'd like to add that i tried to intentionally feed and still got calibrated almost 4.3k
how cute is that? System is broken af in my opinion.
Twitch.tv/benao_noruega
Twitch.tv/benao_noruega
4 hours ago
WHO GIVES A FUCK AND DO YOU REALLY THINK WE DON'T KNOW HOW IT WORKS YOU SAD PIECE OF SHIT? WE KNOW YOU FUCKTARD WE KNOW ITS JUST YOU FUCKING SPAMMERS AND FAIL FUCKING SMURFERS THAT DONT KNOW SHIT NOW GTFO AND STOP SPAMMING OUR FORUMS!
King of Low Prio
King of Low Prio
3 hours ago
ROFL, game winning plays that don't win the game. Spoken like a true 3k ape
RoadTo50Reports
RoadTo50Reports
3 hours ago
Someone's mad, I found this one new helpful tho so ty
Chemistry Knight
Chemistry Knight
2 hours ago
Very helpful..
I am gonna calibrate soon..
So thanks...
Also dont be mad when someone tries to help... Thats just mean
MEDREPBOY007
MEDREPBOY007
2 hours ago
@OP you will calibrate at 3.3-3.6 kiddo. Your not a 4k material. 5 k dream? :-)
some things are just inev...
some things are just inev...
2 hours ago
GIT GUD OP
beefballs
beefballs
an hour ago
Dude but how much mmr did u lose when u intentionally feed 10 calibration games in a row? A lot
mid or feed
WritePreview
Remember to keep your post friendly and within the Dotabuff Code of Conduct
English (English)▴
About
FAQ
Support
Privacy
Hiring
Service Status
Help Translate!
Facebook
Twitter
TrackDota
Copyright 2016 Elo Entertainment LLC. Dota 2 is a registered trademark of Valve Corporation.
Premium advertising by GameRevolution
"Dude but how much mmr did u lose when u intentionally feed 10 calibration games in a row? A lot"
idk, about 300-400, who gives a fuck anyways
Будь ласка, увійдіть до системи для написання коментарів.
First and foremost, everything said is my opinion. I've done a fair amount of experimentation to be able to draw some conclusions.
1. First X unranked games you play
Within your first ~30 games (I believe abandons do not qualify here - if you abandon with 0-20, it shouldn't influence your overall stats (KDA, GPM...)) your rating is more or less decided. You can calibrate upto 5k, and be matched in a 4.5k-5k average game. On this dotabuff there's an example of a 6k, 4.9k, 4,5k x2, and me (probably at 5000 virtual MMR at that point because I previously had a 24 KDA game at 4.5k average mmr). This means that there are brackets even in the 'VHS' and that keeping your account in good shape may result in having the first calibration game at the average of 5k mmr, which is the maximum.
"More or less decided" means that after your 30 games, your virtual MMR will stop spiking. Between my 6th and 8th game, my virtual mmr has dropped by almost a thousand (3900-4100) and in the 9th game the average mmr was again around 4.3k.
2. Stats
Everything is taken in consideration, even the sentries and dewarding, maybe even TP scroll usage. I think most important is 'kill participation' - even this stat I think is unavailable (?) on dotabuff, I believe is used as a multiplier in the formula. KDA surely is very important, as well as GPM/XPM compared to your teammates. Hero healing, tower damage, and LH+D per min.
3. Should you make a smurf to recalibrate?
I played some 50 games in solo matchmaking at 3300MMR (I play dota on and off, mostly party at 4k+) because I never took the time to boost the MMR, since it's what I calibrated at when it was introduced in beta(?, whatever).
I found that the game winning plays in that bracket are not necessarily the correct plays. I don't know to win a 3k game, other then feel I'm going to literally 1v5 them with a hero like slark or tinker, or just crush them in a teamfight with a massive AM or sven. But I like heroes like WR, QoP, TA, SF, weaver... and they don't win by themselves & they exploit tempo. You can't exploit tempo in a 3k game, because when you wipe their team, instead of doing dmg to a t3 your teammates go farm the first component of their fifth item or smth.
General lack of enthusiasm about winning is pretty much my feeling of the 3k. A round of CS:GO, literally - I don't want to play that game. I cannot improve in that game....
4. Things to know when playing unranked:
I like to think of them this way: I imagine that I'm playing with a hand of midas: 1/0/1 is much better than a 3/1/0 especially if I'm mid or even offlane. I'd rather spend time farming than searching for kills, but being in the right place for the fight is effectively the best place to farm.
By farming, I mean farming your stats: kill participation, hero damage GPM, XPM, LH, wards, dewards, sentries, camp stacks, etc. Kills are not better than assists in any sense other than that kill means more gold, but I think it's better to try do most most hero damage as possible.
Hit towers when you can, instead of running around. Tower damage counts too, especially if your team hasn't done much damage to towers. I think raxes count too but I can't be sure.
Apparently teleports count too, so tinker is a really good choice - if you're good on it, you can create a gap in your team kill participation and that's what will give your virtual mmr a boost. (if your teammates don't participate in kills, the game recognizes you as doing pretty much everything except for dewarding and herohealing (DP soul siphon counts as hero healing for example).
I'm 90% sure overall hero versatility (the graph as a whole too) counts a lot, and any 'red points' on the graph might negatively influence your mmr before calibration, and a well balanced graph might boost it.
5. Scrubs shouldn't recalibrate, but climb their way out of 3k?
TBH, if the max rating of calibration is 5k, then whoever calibrates at 4k can't be more terrible than some of the terrible players who are at 4k already. I just think that a lot of people who suddenly fall into the 4k bracket get too cocky, want to play mid all the time and eventually lose because of miscommunication and flaming. I also believe that unless you belong to a higher bracket, you should be able to play 3-4 positions in all types of strategical variations if you are to maintain your mmr. If you can play 20 carries and 20 mids, but 3 supports and 5 offlaners youre fucked, and you shouldnt be wasting your time anyway.