Jump to content
HWBOT Community Forums

GeorgeStorm

Members
  • Posts

    1357
  • Joined

  • Last visited

  • Days Won

    13

Posts posted by GeorgeStorm

  1. When I said unrelated I meant the points given. Winning 50pts for a comp is a number pulled out of thin air which goes against how other points are calculated in my opinion.

    Removing points completely from comps, non Manu ones, would kill them. (And team/country as I don't think they'd be fine)

     

    Manu comps are done for prizes more than points in my opinion.

  2. 25 minutes ago, Matsglobetrotter said:

    Yes it is true there are more new and expensive hardware. Part of the sport i thought was that you could go out and also buy cheap old hardware to beat the C@#p out of through hard pushing.  It would mean if i never submitted scores before i will get nice good points in the season for going out and buy what i did not have of old stuff in my drawers. Ofcourse I would not go out and by a new 7640x or a 6700k if i have one but it would probably also mean it has degraded enough to not potentially give the peak score i had in  a previous year. Now I might be pushing the hardware with chiller last year and dice this year. In some of the last years team competitions this was exactly the case. Did not give me any scores but pulled a lot of points for the team. but can you then give advantage to the guys that went out and bought new (old hardware) over active users just because they did not get a new score on the same this year? if my score was actually higher while lower compared to previous year?

    Meanwhile I do think its a bit of a turnoff to see that my scores having finally come above 1000 points i will be back down to 73 points because i have only used my 9900k effectively in 2019.. I had an extremely intense period the last 3 months of 2018... probably spent 500 usd on dry ice and another 500 on the pots. now it would be more sensible to simply stop trying the last 3 months and start a major session beginning of the year... 

    Just ignore seasonal and care about career ?

    • Like 4
  3. Having fewer results doesn't stop people submitting, simply tries to put quality over quantity.

    Having a 'season background' would work, but I feel it would have to be treated as a comp, where users have to click to enter it as a season score, with a note saying the needed background is required etc, as otherwise the amount of incorrect subs would just be ridiculous.

  4. 8 minutes ago, websmile said:

    Doesnt it make you suspicious you have 2050 and 9k, and when you go to 2110 you have maybe 9500 at best at same timings, which is around 500 more, and that when you go from 2390 to 2450 you suddenly have around 10400? which is above 1k more even though you gained alot less frequency percentual and same in real frequency?? Not my business anymore after I quit benching but I think this is quite obvious to see together with other funny things, especially when you know benchmarks usually have baselines and it is percetage above baseline gain that makes the results.

    I thought it was because one of the gb3 subtests massively favours bandwidth, and is weighed heavily enough in the overall score that pushing that high gives you a better score overall even if you lose score in other subtests? (but I could be wrong as I've not tested myself)

    • Like 2
  5. 2 hours ago, Rauf said:

    No more compromise! Define a clear objective and make it happen! :)

    BTW, isn't "hw depth" what competitive overclocking is all about? Shouldn't the best overclocks be rewarded? I'm all for making it cheaper, but you can never factor out binning

    The objective is to allow people to take part in competitive overclocking/benchmarking in my opinion, which doesn't just mean the top 0.1%, that's just a WR tracker.

    Also to a certain degree yes, but one guy having one binned chip that gets a bunch of global 1st places/golds isn't any better or worse than someone who does well in a wide range of benchmarks and with a range of hw. So you need to have some kind of compromise in my opinion, as having too small a number of scores counting will mean those without highly binned chips will next get near the top, but on the other hand having every score count for example would turn it into just a spamfest of results hah (which would still possibly favour the top end benchers I think actually due to the points scaling)

  6. 1 minute ago, TerraRaptor said:

    These spamming results make categories popular bringing more points for the first place.

    Which undermines the algorithm in my opinion.

     

    Also @Rauf I can see where you're coming from, but by limiting it too strongly you're still making it a money game, just instead of hw breadth it's hw depth so to speak, with needing to bin/buy binned hw at a premium.

    As @richba5tard said, it'll be impossibly to keep everybody happy, but hopefully a compromise can be made :)

    • Like 1
  7. Results submitted to uat.hwbot.org won't be ported back to the main site as far as I'm aware. It's a test server, normally using a copy of the main database, hence why it'll often be a little out of date.

    • Like 1
  8. 10 hours ago, Fasttrack said:

    George, you are one of the most down to Earth members here on the BOT.

    So, question :

    Don't people need in every form of activity in life options and challenges in order to keep the spirit up ?

    To wake up and keep trying to get better ?

    If we start - cut this, stop that, restrictions, space getting smaller - then we wake up and say " what the hell, I might as well do something else "

    Yes I agree it wants to be inclusive and open to all. I was thinking of it in a way everybody gets to do what they want, with no pressure to do things they don't (so if they like comps but don't care about the 'normal' ranking they can still compete, and if they don't care about comps they don't have to do them, and if they like both then they get to compete in both etc.) Wasn't trying to divide things in a negative way.

     

    Rauf the reason I suggested either eliminating or greatly reducing both globals and WR was to improve the spread of points, as I feel it's currently a little extreme, since as you pointed out, when something like 03 doesn't scale with multiple cards, it means a single score can give over 450pts, which compared to a global first in a popular category (4x cpu r15) is over 200pts difference.

    It's possible that simply changing the benchmarks which are awarded them and reducing the ceiling of them will solve the issue whilst still rewarding those who want to push crazy setups that might not get many hw/globals. (what they were designed to reward in the first place)

    Edit:

    Unity one of the 'issues' with comps is that they're time limited, which means once they're over that's it, nobody can beat you, which goes against all the other points on hwbot, hence why looking to keep them to a system that resets, naturally counteracting that aspect.

     

    • Like 3
×
×
  • Create New...