Page 2 of 3 FirstFirst 1 2 3 LastLast
Results 11 to 20 of 23

Thread: My two cents on the performancebar change

  1. #11
    Basic Member mr_9sky's Avatar
    Join Date
    Apr 2012
    Posts
    270
    Quote Originally Posted by gpardini00 View Post
    its well analized and it makes sense to me
    Butt hurt.

    But seriously, getting on par with the ever changing meta of heroes in game due to trends is a pain in the ass.
    COMMUNITY REGULATED MUTE SYSTEM (CRMS) A Simple Idea with Massive Results
    UNOFFICIAL DASHBOARD POLL Dashboard - OLD vs NEW
    DOTA2 Items ICON SET Design Comparison CHART - Icon Cheat Sheet
    POLL: HEALTH BAR at TOP BAR? - Add health bar to top bar Valve!
    True Sight Indicator like dota1 - What happened to the true sight symbol?!?

  2. #12
    Basic Member
    Join Date
    Jun 2012
    Location
    London, UK
    Posts
    678
    Whilst these bars now do seem to more accurately reflect my personal skill with a hero, from my own perspective, it's all still rather nebulous when compared to a certain stat tracking site. The information we are provided on our performance is not detailed enough, we don't know what the metrics are, and making it only relative to your own performance across heroes with no distinguishing between roles and so on means it's inherently less useful than some kind of global system per each hero.

    Who cares if I can apparently play hero x as 'a 10' but my friend plays hero x like 'a 5' when he also plays heroes a b c like a '10' and I play them like a '2'? ...relative to ourselves. There's no useful comparison we can make there. All simply pointless stuff.

    edit: I will add that you have kept apparently the same ordering for heroes as it was before the change, which means that even if that is simple residual ordering and no longer updated, I can still say right now that although Axe is apparently my best hero (about a 4/5) compared to the rest of my skill bracket, he's also completely mediocre when compared to my own performances with other heroes (2.1/5). Making these kind of judgements is the most useful this area of the profile has been. If you actually showed two bars, one for global hero skill, and one for personal skill, you'd have a better way of presenting that information, if indeed you're still interested in per-bracket hero metrics.

    ie, something like this (blue bars being bracket rating)



    --

    I would like to see detailed, concrete, statistical information on performance per hero - this is not difficult based on the vast volumes of data provided during a DOTA match, and many people in the community (eg Bruno) are working to provide better tools for analysis already when they are completely absent in the client. For example I want to see which support hero with healing abilities I provide the most healing per match on average, then I want to see which support hero provides the most healing when items are also considered. I want to see which heroes I have played have given up the least amount of gold to enemies. I want to know which hero does the most damage for least amount of active ability uses. And who does the most right click damage. And so on. And I want graphs and pie charts and other sexy stuff.

    So overall in summary: points for effort, thank you for taking a look at this area in light of recent events, perhaps the chart is now more accurate per each person but it's now completely less useful as a grading system. Nice try, can do better. C-
    Last edited by cinephile; 01-31-2013 at 10:27 PM.

  3. #13
    Basic Member
    Join Date
    Jan 2013
    Posts
    2
    Dear Icefrog,

    Hire the Dotabuff guys.

    Sincerely,

    p00psicle

  4. #14
    Basic Member rockonrob's Avatar
    Join Date
    Apr 2012
    Location
    Wonderland
    Posts
    1,545
    Eh, the performance bars now are pretty much worthless as they are mostly based on wins, not the reasons why you won or lost. At least before you might have more of a general sense of where you could be, rather now its like seeing how many games I won with each hero and that's it. It would be nice for a revert or a more accurate ranking of skill, better than the joke they are at the moment.

  5. #15
    Basic Member
    Join Date
    Dec 2012
    Posts
    7
    Imo they should make the two available with a tab or something, I mean, everybody knows who is his best hero, but knowing how good you are with it compared to other people at your level its pretty great.

  6. #16
    Basic Member
    Join Date
    Mar 2012
    Posts
    105
    Well, I think any player who complains about the performance bars( about the old system and the new system) are right. There is no point in having a W/L based ratio to measure your skill in Dota, the game is too complex for just that.There are too many roles you have to fill in with different heroes that just K/D ratio doesn't do the trick. My view is that Valve has to drop the whole "performance bars" idea and start working on a system that tracks stats about every game and combine them at your profile, a unified "profile stats" if you like. The factors that will be taken into account must be in the range of "Gold gained, Experience gained, Last hits, Denies, K/D/A ratio, How many minutes you played each class (support, carry, durable) and what are your stats SEPARATELY with each class with sub-categories (example: Support---> wards bought, courries bought, courriers upgraded etc.). So imagine it like a statistics tree that the more you go down on it, the more analytical it gets. In my opinion this requires a lot of work but in the long run it would be profitable for the community and it will fix the broken matchmaking system too.

  7. #17
    Basic Member
    Join Date
    Feb 2013
    Posts
    4
    I'd say just have a table that shows each hero, and all the according stats. Kills/Deaths/Assists/LH/Denies/Towers/GPM/XP-MIN. have all that and base your chars off of that. An algorithm averaging them together for a common stat like "performance" would be good. Because even if you play support you should have a high assist number and a rather lower death number, and very few kills. LH and denies I believe is very key to performing a hero well. Denies for support LH for carry.

  8. #18
    Here's my analysis in similar form.

    Old rating:
    + none
    - was complete bullshit

    New rating:
    + Has some sort of connection to reality
    - Is basically your winrate with a specific hero in graphic form
    Quote Originally Posted by DeadlyGrim
    Is a man not entitled to view his own ranking? No says the man at Valve, it belongs to us!

  9. #19
    Basic Member
    Join Date
    Feb 2013
    Posts
    1
    Quote Originally Posted by Good old Grim View Post
    Here's my analysis in similar form.

    Old rating:
    + none
    - was complete bullshit

    New rating:
    + Has some sort of connection to reality
    - Is basically your winrate with a specific hero in graphic form
    Sorry, but I am able to do the math myself and see which heroes of mine have which winrate, I don't need a graphical system for it.
    I rather see comparision of my skill playing heroes like having warlock whom I play better than Omniknight but with both 50% winrate.

    Up with OP, I way prefer old system that actually made sense and reaching "4" score was a worthy challange and required to play good.

  10. #20
    Basic Member
    Join Date
    Sep 2012
    Posts
    470
    Quote Originally Posted by Whiterock View Post
    Sorry, but I am able to do the math myself and see which heroes of mine have which winrate, I don't need a graphical system for it.
    I rather see comparision of my skill playing heroes like having warlock whom I play better than Omniknight but with both 50% winrate.

    Up with OP, I way prefer old system that actually made sense and reaching "4" score was a worthy challange and required to play good.
    ^This. Seriously, it isn't that hard to do basic math. People like this only because they feel better when they have 10 with a few heroes, but basing it only on w/l ratio is completely retarded. It doesn't say how good you are with a hero. The old performance bar also didn't show much, yes, but the idea was much better and it actually tried to show how well do you play a hero. IMO, the performance bar should show how many people in your skill bracket play worse than you (or be limited to a certain amount of players above and below your position in the bracket), because having more wins with a hero does NOT mean you play him better.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •