richba5tard

Just a hypothesis, but what would be the effect if no points were awarded for wr/gl, only hw? It would significantly reduce the money factor, while skill is still involved in getting high hw points. People who go for world records have less "points" awarded, but I doubt they do it for the points anyway. For shits and giggles let me try this out on uat.

I obviously can't get rev 7 fully working due, do I? Despite hundreds of automated tests, a test environment with a full production database copy, we are not able to reproduce the issues with rev7. Rankings just get fucked up on production and nobody seems to know why or how. If I can't reproduce it, and i've got 685.000 lines of 14 year old code to maintain, well guess what, I can't get it bug free. Even worse, all my spare time is spent on finding issues with the points and dozens of different rankings, I can spent time on improving other stuff. One major reason for rev7 was to free up calculation power, but in the end I had to write a tool which checks for broken rankings and repairs them, causing high load... I need it easier to maintain, otherwise HWBOT will never improve quality wise.

Indeed. In this case it would not be a burden on the code/maintenance, but I see it as a big drawback if the point algorithm can not be explained easily, without dozens of exceptions.

You guys are exhausting. Current revs are no good. Old revs are no good. New revs are no good.

If "accumulated mistakes" is implementing feedback from crew and community, yes that caused a huge complexity. In current rev8 proposal i was able to remove a large chunk of this accumulated code.

If we have an easier algorithm to maintain, i can actually spend some time polishing other parts of HWBOT. Please keep this on topic, not bug reports of unrelated stuff.

Ah I see. Does it really make a big difference though? Whether the total amount of points is used instead of hw only? I can increase it to top 40 (total), but i'd like to avoid the "sum of top 15 global/benchmark points and top 40 hardware point submissions" approach, as it is only clear to people who have been using hwbot for a long time.

Can you elaborate? I'm not sure whether I know what you mean with "dedicated hardware slots".

Yes, that is a possibility, and it is not hard to explain neither hard to implement so I'm not against it. It does mean that the team ranking is nearly static. Why not do the same as member points and have both?

Thanks for the feedback Jon, i feel your pain. It is a very reasonable suggestion competitions such as the country cup have a big impact on the team ranking. It shows skill and team participation as no other. I will take it under account!

No, hardware points still contribute for your member points. In rev8 the suggestion is to use the "total points" of your submission, which is hw+gl+wr points.

If needed be we can make it part of the rules, and auto-ban people taking advantage of it. You would not want to spend moderator time on this though.

HWBOT is not hurting for cash, and rev7 is computational wise already OK. What is not OK is that we have such a huge code base to maintain for all the specific kind of rankings, all with their own logic, all influencing each other. It's nearly impossible to keep stable with limited time to be able to spend on it and for what? I agree that staying on top of the team/country ranking would require yearly effort, but that is exactly the point. The team and country rankings are almost static, nothing interesting going on. If you reset yearly at least there is a yearly fight for the top. We can do the same for teams as for members though: show both a seasonal as an all-time ranking. Team and country can also be based on top xxx members, instead of all. That is not too complex to explain.

Goal is to have 2006 complexity in rankings. I suspect there are only a handful of people understanding how the rankings work. Each ranking has it's own logic (especially those based on competition points and team power points). In rev8, ALL rankings (except the hardware/global/mobile masters) are based on member points, which is top 30 submissions + top 10 competition entries. Good point if we stop with user competitions. It levels the playing field for the seasonal ranking but it does have an unfair advantage on career ranking. We can always make competitions ourselves though. But hardware points are a contributing factor! Member points = top 30 submissions (global + hardware + wr points) + top 10 competition entries. Why the idea it is no longer relevant? I am dead set on having all rankings based on same logic. Difference in career and seasonal should only be the scope of the year, not different logic IMHO. Hardware points still count. Why do you think it does not? Deciding which apps get points and which not is an related discussion.

Moving to public section to gather some more feedback.

Summary is the first post of this topic. UAT does not have it's own wordpress server, hence broken links. In the long term I want to get rid of wordpress completely.

Edited to labels to make it more clear where the rankings are based on seasonal results or career results. What do you guys think of the top 30 for your memberpoints? Is the UAT version polished enough to give a dryrun in the forums?

Ok, made the newspost. Big thanks Albrecht! ?

Congrats to OGS!

uat now uses top 30 for career and season rankings.

Is there any other GPUPI submission except the one of bob(nz) that needs to be added to the competition due to the v2.3 / v3.x mixup? bob(nz)'s submission has been added manually: http://hwbot.org/submission/3985668

Sorry you have reached your allowed amount of comments for this month. Please cool down until dec 1st 2018, or become a pro member(tm) for just 0.99$ and post now!

Allow people to submit up to 12 hours after competition ends for only 4.99$? ?

The intent is to provide overclockers with a sense of pride and accomplishment for unlocking different points. As for cost, we selected initial values based upon data from the Revision 7 Open Beta and other adjustments made to milestone rewards before launch. Among other things, we're looking at average per-player points earn rates on a daily basis, and we'll be making constant adjustments to ensure that players have challenges that are compelling, rewarding, and of course attainable via overclocking. We appreciate the candid feedback, and the passion the community has put forth around the current topics here on HWBOT, our forums and across numerous social media outlets. Our team will continue to make changes and monitor community feedback and update everyone as soon and as often as we can.

I tend to agree. If I remember correctly the reasoning of massman for the high cut off percentage was that only high effort submissions should be awarded significant amount of points.