I think my winrate stats are low. I am stuck in the 1600-1700 bracket for a few months despite being ver likely the best player there.

I don't think it is because of the way I play or my skills. It is maybe a component of my SKILL. You can otherwise blame the elo for letting players into 1800s or 1900s only after 10 lucky wins against the right players. And later holding the position by playing player that would give 0,1 or 2 points in order to occupy the place permenantly.

Any ideas on how improve stats and winrate ?

Edit : I replaced my obsession of awesomguy with alhazi.

Edit 2 : I am top noob of 1600 so I am mediocre. My only goal is to become bottom noob of 1700s for progression purposes.

Edited 8/18/2015 04:38:20

Any ideas on how improve stats and winrate ?

Easy! Play against weaker opponents, or even create an Alt that you can beat repeatedly!

Now, if you meant improving your skill, it is an entirely different game.

Edited 8/18/2015 02:48:21

For starters, drop the ego.

Accept that you're mediocre and that the reason you're rated 1600-1700 is because you deserve to be. Until you do so you're damning yourself by fooling yourself into thinking the reason you do poorly somehow isn't your fault.

Additionally, you'e not rated at an elo of 1600-1700 because the RT ladder doesn't use elo ratings.

Elo ratings are used on the 1v1 ladder, and you'd likely be more like a 1300.

First, realize there is a reason you are rated lower than you'd like. You say you are likely the best player rated 1600-1700 but it's unlikely. If you are rated the same as someone else, there is a strong probability that you are around their skill level. If you have stayed 1600-1700 a while, it's where you belong with your current skill. If you want to get a better rating, get better.

Make games vs good players and ask them to help you improve. Study games by high level players and see how they played. Think more in your games and study every game you lose. These methods all work to some degree.

It is just my winrate is slightly equal now on wins/losses. At least more acceptable. There are days where I lose most of my games and other days where it is otherwise. I am not farming players about the same skill level as me. I wondered if I need to avoid players as some do. Dom365 or alhazi are maybe good bets for pts because I have satisfying runs.

Playing less gams can help too. It may be that a third or fourth ladder game is not as good as first. I may otherwise make an alt and play against top 20 players I know their play styles well to maximze win rate and also choose days where they perform less well. It is impossible to get into 1800 or very near 1900 in 10 FIRST games if you are not an alt. Players losing 300-400 pts from peak rating seems strange. Good players keep thei positions.


I rarely play wih people on my bracket.

And tommy15 is the only player I remember that is super good and also 1500-1600.

I will play some more games against alhazi or review more. He is better than me therefore my play must be very above standard. Quite a while I played him.


How do I improve rating if I absolutely want to play against Dom or CharliBlue when they are on ladder.

Edited 8/18/2015 04:18:42

Unless I improve as much as others in the ladder, it makes a lot of sense.
Because I did improve while gravitating around the same rank 1650.


Interesting take on things. How much a 1700 would be worth 6 months ago in rating ? How about first ? I heard higher than 2300 is impossible in chess world.

Edited 8/18/2015 04:54:50

Bayeselo is pretty accurate, actually. If your win rate in recent games is significantly higher or lower than 50% (say <30% or >70%) and your rank doesn't seem to be changing quickly, it could be due to opponent quality or due to playing fewer games at once than you used to (or taking longer to finish them). If your skill level has changed significantly during the period for which your ladder games remain unexpired, then you might have to wait for your old games to clear out before you see any jumps in ranking. It's not just your most recent games that count, after all.

If you want to test your hypothesis that you're significantly better than your bracket, start inviting players in that bracket to strat ME 1v1 games (or PM me on my main with how many games you want to be in at once, and I can automate something) and see how significantly your results deviate from Elo-based predictions (if at all)- i.e., how often you pull upsets and how significant those upsets are.

As for how much a certain rating would've been a certain number of days ago: hard to calculate but it all depends on how many points have entered the system. Leagues have set means (the # of points new players come in with) and therefore the total # of points is always (# of players whose points are in the system * mean # of points). You can see how ratings will get inflated over time as more points enter the system.

Why not try to improve your skills rather than the gimmicks you mention? Your goal should be to play well and wins will come.

^ +1

And if you feel that the 1v1 ladder isn't giving you competitive enough opponents, just join another ladder or one or more of the CLOTs. You can also just create games against good players yourself, or buy a Membership and set up an automated system to do so for you.

i dont really know hwo the ranking works like in the rt ladder, but the old games dotn expire and that miht make it harder to go up, but im not quite sure

Thread closed. Will use clots and I will be creating tourneys. :)

I just glanced at your seasonal games (maybe 4 or 5) since you don't have 1v1 ladder on your profile. I noticed your picking could use a lot of work.

Sometimes you pick decent locations (you are not picking terrible spots), but your order is off. For example, in one you picked Aus, Indo, SEA where SEA/Indo is a FTB. What would happen if you get 1 and they get your 2-3? They get FTB and the counter to Aus, so that pick is useless.

My tip would be to think about position within your pick areas. For example, if Central America is a FTB with Baja and a spot in South America, take SA #1 and CA #2. SA is more desired there because Baja must take out 2 neutrals to get to SA, and also SA would have access to Africa and AA. Picking the correct order within your pick set can matter a lot.

I think you are miles ahead of a noob, but still need some work on picks to put yourself in good position to make good moves.

um... win???

Hostile is trolling, guys.

@Pink: TrueSkill is actually really good at detecting changes in skill because of it's mu-sigma model.

Drop out of the rt ladder. That place sucks.

Its true that after you got your standard deviation down on the real time ladder, games give less points. However im pretty sure you lose less too. I was hovering around 1600-1700 for a while too, but if you just keep winning games against people rated the same or higher your rating will go up.