Dota 2 first release excel

Dota 2 first release excel

Hello, how does the ranking TBD work?

I’m requing my second account. Accidentally went into a ranked game today (my friend made a search for ranked even though I asked for a normal). And I was TBD.

Luckily my mmr was higher than I ever had it before (I calculated since I was the top 1 player/ average mmr)

I’m so happy right now. But the thing is:

What happens when I keep playing normal and do good? Win?

Will the TBD raise or not?

On what do they calculate?

Wins?
Lvls/ game experience?

Anyone that have insight about this?

Do you even know what TBD means?

Thanks, but how is it determined before I get my TBD?

I actually found a good thing on the Internet. Imma post it!

‼️Following post is copied from another site:‼️

The mmr you ‘start in’ is exactly the same as everyone else (the figure is not known but it is between 1500 and 3500). When you first start playing dota you actually just do hidden ranked calibration in your first 20-30 games where it massively scales up the swing from game to game in order to place you close enough to your skill level while also not taking 150 games to do so.

Ranked mmr is indeed influenced by unranked mmr in that youre starting point when you calibrate your ranked mmr is your unranked mmr. However this makes almost no difference as to how your ranked mmr will turn out after calibration because the same forces that moved you from a starting mmr of 1500 when you first played dota to a 2200 mmr player when you first started ranked calibration are going to be acting on you again. So the difference in mmr 10 games can make is potentially in the thousands. What makes people believe that it is “influenced a lot” by unranked score is because they often refuse to believe it is an accurate representation of their mmr. One thing about your unranked games that does greatly affect your calibration process is your average performance. For each game you play your performance in terms of kill participation/rate/time, death ratio/rate/times, creep score, GPM, XPM, hero damage, building damage and so on are all recorded. What the ranked calibration process does is compare the score you had on that hero in that one game against other players that also played that hero and what the average MMR was at the performance level you exhibited, this is then checked against your historical stats on that hero (if it exists) as well as your general averages. This is why it is said (and proven by observation) that your MMR will be increased if you play only core roles from the start of an account until you finish ranked calibration. The difference this will make however has been shown to be only around 200-300 mmr and well within acceptable bounds. This doesnt mean that you should always play carry and your mmr will increase, it simply means that the kind of things that are compared during the ranked calibration are slightly skewed to roles and heroes that excel in that area. Dota 2 first release excel

Lets say your unranked mmr was 2000 mmr when you started calibration and your played a perfect balance of all roles during your unranked games.

You decide to play enigma if your first calibration game. Your team loses this game pretty badly (down by like 40 kills and several towers as well as a huge gold/xp defecit) so it was clearly a one sided result. However despite this your average stats for this game were comparable to the average enigma player with 2400 mmr. So the calibrator says okay you are going to lose 50 mmr right now but if you play as well as you did this game (comparable to people of your current mmr) im going to give you 200 mmr if you win and only take 50 again if you lose. If you average stats are the same as a 1950 mmr player (+/- an accepted margin) then im just going to do +/-50.
MMR = 1950.

The second game you decide to play Tidehunter. This game is much longer and much closer but your team still loses (65 to 60). The calibrator looks at your stats and says well hey your stats were comparable to a 2500 mmr tidehunter player this time and decides that you are only going to lose 50 mmr again this time but if you win the next one like a 2500 player im going to give you +400 mmr, but if you only win it and you played similar to your current mmr we are going to use that +200 from before.
MMR = 1900.

Game three and you decide to play Troll Warlord. This game was pretty one sided and your team easily takes victory (50 kills to 10). The calibrator looks at your performance and says well look you won but you only played like a 2200 mmr player this time. It decides to risk it and bump you up by the full 400 it promised last round but because you are now above your performance would indicate it says well I’m going to take away half of that 400 if you lose and see how you do there and if you win im just going to give you the standard +50.
MMR = 2300.

Game four you decide to play Ogre Magi and your team wins. Another decently close game but a couple of mistakes in the late game cost them the win quite decisively (55 to 45). You played comparative to a 3k ogre magi player so the calibrator has a little think and says okay you are going to lose the 200 from before but because you played so well this game im going to give you a shot if you can win out your next one and decides that you will get between 200 and 700 mmr depending on your next win and only lose between 50 and 150 if you lose because your average performance so far has been comparable to a player of around where you are.
MMR = 2100.

Game five you decide to play Enigma again and your team wins really easily (31 to 8 and the enemy got no towers). Your performance was comparable to a 2450 mmr enigma which the calibrator knows was really close to the rating on the hero you got in your first game. So it decides to give you 350 mmr for the win and is going to reset to +/-50 in the next game and go from there.
MMR = 2450.

This goes on for the remainder of your calibration matches and leaves you within

100-300 mmr of your skill level which is easily acceptable. The ranked mmr calibration has VERY LITTLE to do with your team mates and much more to do with your individual performance which is why people who play cores during calibration get around 200 mmr more than the same person playing supports. This is not an avocation for playing carry it is simply a reflection of the process (which has been improved greatly since release as far as support calibration is concerned).

I have not proof read this post so if anyone needs elaboration feel free to ask for it

Oh I realised people might want to know how I have come to this information.

1. I have implemented an Elo-like ranking system in the past.
2. I have observed the calibration process 3 times.
3. I have read a report from a wonderful user on reddit (ill try and find the thread) where he details an experiment he did with some other redditors in order to see what effects final rating. They discovered a lot about the system when all of their MMRs were different despite only ever playing in the same games together from start to finish.
4. I have read all of valves releases on the subject and the debates that they bring.
5. I spent a lot of time modelling S2 games matchmaking rating system when it was first released (because much more of the data was exposed it was a little easier as well as it stuck more closely to the original Elo concept).

And my solo ranked MMR is 4500 and I exclusively play support (the position of that support varies heavily though).