Ranked Game Pie Chart

Hey All,

Anoek was kind enough to accept a contribution I made just recently: a “Ranked Game Pie Chart”.

You can see this on your Profile, it looks like this (this is from my own Profile page):

To help make sense of it, you should note that:

  • Greens are your ranked wins, Reds are your ranked losses

    • So you can see I have almost exactly 50/50 wins and losses
  • Bright colours are vs stronger opponents, dull colours are vs weaker opponents

    • So you can see that I have played more games against strong than weak opponents, overall

I was motivated to produce this pie chart because:

  • Ranking works best when you are playing evenly stronger and weaker players
  • Many people find Go most enjoyable when they are are winning and losing evenly

… so if you are like this, now you can see at a glance whether whether you’re due to play up or down :slight_smile:

(It also has the nice (IMO) effect of increasing the vertical aspect ratio of the rank graph, which I find more pleasing than the long flatish line right across the screen, and it adds the somewhat-requested total number of ranked games played).

GaJ


PS: it may also be worth pointing out that you can select a portion of your record to examine, by dragging on the lower plot:

I dragged from beginning of the year 2018 to the right, and now the pie chart shows only my games from this year, and shows that I have an even split of playing up and down this year (50/50 bright vs dull), and have more wins than losses (bigger green than red …very soon to be dramatically corrected in the Fast Correspondence Tourney games finishing soon!)

16 Likes

@GreenAsJade: thank you for your contribution.
I assume that these changes do have fixed also the reliability problems. :wink:

I don’t know if it’s really possible to control the balance of both win/loss and weaker/stronger. When I tried to do that, one indicator out of two usually were going in the wrong direction! :smiley:
So I’ll keep my chart unbalanced and see what happens…

Also increasing the vertical ratio makes rises more pleasing, but declines more depressing. :smiley:

It’s not scientific, but FWIW, I tend to be guided by the win/loss. If I am on a losing streak, I play down. If I am on a winning streak, I play up. I think it should be the case that if your rank is relatively slowly changing, then the weaker/stronger balance should follow, and if you are improving, then you will end up playing more strong people this way, which also seems right.

GaJ

Thank you for the nice contribution to OGS, Green. And thank you, anoek, for adding it. I enjoy it as a visual kyu! (Get it…kyu…cue…oh, nevermind. Lol).

GaJ, if one does as you state above, doesn’t their rating change less and therefore not reflect their progress as accurately. If all I played was 1 dans and 25 kyus, back and forth, my rating would be stuck; that’s an extreme example, but you see the point that there might be an unintended flattening or middling effect to picking opponents in that way. I do the alternative, which is to generally play players a touch stronger than me whether I win or lose. But, there is not right or wrong, it’s personal preference. I imagine that, using your method of opponent selection, you sleep better at night than I do…lol

Aside, am I accurate in thinking that OGS ratings are based only on the most recent 15 games played? I ask because I’m thinking that, if the 15 game rating calc is true, I can use the slider bar to look back at my pie chart and get a good feel of how I’m doing on recent games that reflect my rating (rather than the sum total of all my games played which reflects long term progress but not my current rating). Please advise. Thanks.

2 Likes

I’m not clear on why playing even would make your rank go up slower than always playing against stronger players.

Actually, the fastest immediate boost you can easily give to your rank is to play only down for a burst, so you win all the time.

If you play up all the time, then your rank will go up by bigger steps when you win, but the steps will be further apart, and interspersed with little down steps each time you lose.

Again, not scientific, but I have a feeling that this is another reason why the middle ground is likely best. YOu get the benefit of lots of small boosts to your rank when you play down (and ego :wink: ), and you get the occasional big boost when you win playing up, interspersed by little steps down.

Gut feel also says that if you are on the improve, then playing up more would be more beneficial, since you have a better chance of winning those than when you are reaching the inevitable slowing of rank.

If I am on a losing streak, I play down and lose against 15k. :smiley:
True story! :joy:

5 Likes

Just following up my earlier comment to @lysnew

“if you are improving, then you will end up playing more strong people this way,”

I think this backs up my point:

Being a beginner, I have been on the improve, and I have been playing up when I win and down when I lose. As a result, I have a 50/50 win loss rate, and play more games up than down.

I imagine that now I am plateauing in ability ( :frowning: ) I will start tending to play a higher proportion of down - when I am no longer improving, then I will be playing both 50/50 win/loss and 50/50 up down.

All that being said, you can obviously use the numbers for anything or nothing. I’ve just shared what motivated me :slight_smile:

1 Like

Many cases may occour, obviously. I just wanted to tell my story, just for chat.

Now I rarely can choose my opponent, because I mostly play the main ladder and some tournaments.
In the ladder usually I am challenged by stronger players going up and also I’m forced to challenge stronger players because weaker ones are already taken! :slight_smile:
In tournaments it depends on pairings…

It also happened many times that me and my opponent changed rank while the game was still ongoing. Sometimes I was challenging a stronger player and end up the game winning agaist weaker player, or vice-versa. In correspondence games it can occour.

Actually my pie says that I play most against weaker players, that’s not what I sense. But records don’t lie, I guess.

Cool that we have a pie chart now, but I don’t like that the shades of colour were reversed. I think weak (light) colours should continue to stand for weaker players and strong (saturated) colours for games against stronger players.

(Even beside the point of what makes more sense, that’s what everybody’s already used to.)