Teams League Rev.4

[Massman]

As per request, an overview of the different alternatives proposed by HWBOT. I also added a (non-binding) poll that will serve as reference. Note that the voting options have vague names: this is to make sure people actually read the alternatives rather than judge based on a single name.

 

Alternative 1: "PowerTeams"

 

This is the teams league as proposed in the original Rev4 document, where team points are determined by the following variables (following variables describe "powerteam points"):

 

- best team score per ranking (global/hardware)

- amount of participating teams (determines weight)

 

=> Team Points = SUM(PowerTeam points)

 

+ : completely removes the benefits from hardware sharing on a team level

- : reduces the 'team spirit' as newcomers have very little to add to the team total

 

Alternative 2: "Two Leagues"

 

This suggestion came up after a few posts containing -ehem- positive criticism. It basically separates the team competition in two different leagues: one that uses the current algorithm and one that uses the algorithm proposed in the original document.

 

+ : current team spirit untouched

- : hardware sharing still incredibly beneficial

- : two leagues makes the working of the bot a lot less transparent

 

Alternative 3: "PowerTeam and User Points" (detailed explanation on why PowerTeam ranking is a good thing)

 

The latest suggested solution combining the two leagues of alternative 2 into one single algorithm, ranking teams based on the following variables:

 

- Team quality:

-- best team score in ranking (global/hardware)

-- amount of participating teams (weight)

 

- Member quality:

-- % of global, hardware and competition points attributed to team total.

 

Example was worked out here: link

 

=> Team Points = SUM(PowerTeam points) + [sUM(user points) / 10]

 

+ : illegal hardware sharing a lot less beneficial

+ : user effort continues to help the team

+ : one league is more transparent than two

- : the absolute value of a single score for the team decreases

 

Alternative 4: "MaxPointsRanking"

 

Another way of reducing the effect of hardware sharing. As example:

 

Max(GTX480,3DMark03,1xGPU) = 42.8p

=> Accumulated member points for 1x GTX480 3DMark03 =< 42.8

 

Looking at this ranking, it would mean that the following people are contributing to the team's total:

 

- Team Finland: SF3D and Maggaa

- Hardwareluxx: Benchbros and suicidephoenix

- OCClub: Smoke and Slamms

- ...

 

=> Team points = [sUM( MAX(effective team points in ranking, max team points in ranking) ]

 

+ : reduces the benefits of hardware sharing

- : not everyone on the team will contribute to the team total

- : very, very complex algorithm

 

Alternative 5: "Percentage Users"

 

Slightly different variant of alternative 3, where the team ranking is a following the same algorithm as the current league, but all non-best submissions of the team are downrated to a certain percentage. The best submission counts for the full 100%, the other submissions for (for instance) 10%.

 

=> Team points = SUM(points of all best team scores) + [ SUM(points of all non-best team scores) / 10]

 

+ : everyone contributes to the team total

- : big difference in contribution between best and non-best score of team

- : reduce effect of hardware sharing on a very small scale (see example: http://hwbot.org/forum/showpost.php?p=73945&postcount=11)

[Massman]

Please note that it's HIGHLY unlikely that the algorithm for the team ranking will not undergo any changes. So, please, even if you don't like to change at all, make a vote assuming something will change.

[Monstru]

Hard choices to make...

[TaPaKaH]

what's the point of this vote since you don't listen to forum feedback anyway?

 

I see an excuse like "hwbot has a zillion of users but only ~200 have voted in this thread therefore it does not represent the whole community" coming...

[zoro]

The "active" members will vote anyway, it is for their best interest. Whatever the outcome i believe the new bot will be beautiful

[BenchZowner]

what's the point of this vote since you don't listen to forum feedback anyway?

 

I see an excuse like "hwbot has a zillion of users but only ~200 have voted in this thread therefore it does not represent the whole community" coming...

 

One of the very few times that I can TOTALLY AGREE WITH YOU SAM.

Thanks!

 

List of excuses:

 

1. Multiple choice poll: The staff or somebody can say that the poll results are a mixed bag.

2. We want to keep HW sharing to the minimum, so alt #1 is the best option in our eyes, hw sharing is barely beneficial for the team rankings, so it's all good ( we don't care about the individual rankings ) [ good money-making opportunity for the first guy to open a binned hardware leasing company ]

3. Only 200 ppl have voted

Yes, we know that we could place the poll in HWbot's main page but it needs some serious coding.

Oh and it wouldn't be any different, most people wouldn't pay any attention to it or vote.

4. We could also do it in a way which would require every single member to cast his vote simply by making it a mandatory poll, any user who hasn't cast his vote would get to a poll page to cast his vote, voting page won't go away and the user won't be able to submit any new results before he casts his vote.

But that won't be ok again, because the Japanese & Chinese can't understand it as our beloved friend Pieter says.

 

[ The Japanese & Chinese are very smart and good people, and somebody can translate the poll page to Japanese/Chinese/Mandarin & the rest for the ones who don't understand English well or at all ]

[Massman]

Damned if I do, Damned if I don't

[knopflerbruce]

I voted for option 5, although 3 is also pretty good (it just sounds a bit more complicated, and basically the same as 5 anyway).

 

Option 1 looks nice, but it kills the motivation for alot of guys.

[Massman]

To those voting for Alternative 5: could you please explain why you don't like a specific teams league (as in Alternative 3)?

[knopflerbruce]

You mean one set of boints calculated for the team only, and one for the personal profile?

 

To me 5 just sounds easier, and the pros and cons from 3 are still valid.

[Massman]

You mean one set of boints calculated for the team only, and one for the personal profile?

 

To me 5 just sounds easier, and the pros and cons from 3 are still valid.

 

Not entirely.

 

The effect of hardware sharing is more prominent in Alternative 5 than in Alternative 3, as it affects the top score of the team. Maybe an example to show what I mean:

 

User Ranking:

 

1. OCA 1 - 50p

2. OCA 2 - 40p

3. OCA 3 - 30p

4. OCX 1 - 20p

5. OCX 2 - 10p

6. OCX 3 - 5p

7. HOT 1 - 3p

8. HOT 2 - 2p

9. HOT 3 - 1p

 

Assuming team/user ratio for this ranking of 1/2, PowerTeam ranking:

 

1. OCA - 25p

2. OCX - 20p

3. HOT - 15p

 

Alternative 5:

 

- OCA = 50 + (40 + 30)/10 = 57p

- OCX = 20 + (10 + 5)/10 = 21,5p

- HOT = 3 + (2 + 1)/10 = 3,3p

 

Alternative 3:

 

- OCA = 25 + (50 + 40 + 30)/10 = 37p

- OCX = 20 + (20 + 10 + 5)/10 = 23,5p

- HOT = 15 + (3 + 2 + 1)/10 = 15,6p

 

Fighting for a better ranking for the team will be MORE BENEFICIAL with alternative 3 than alternative 5. In addition, hardware sharing will be LESS of an issue in alternative 5. Beating the best score of another team is a lot more beneficial in alternative 3 ... isn't that good for the team spirit?

 

//EDIT: First case study - hardware sharing within top team//

 

Just as example, this is what happens when a OCA 4 comes in 4th in the ranking:

 

Alternative 5:

 

- OCA = 50 + (40 + 30 + 20)/10 = 59p (before: 57)

- OCX = 10 + (5 + 3)/10 = 10,8p (before: 21,5)

- HOT = 2 + (1 + 0)/10 = 2,1p (before: 3,3)

 

Alternative 3:

 

- OCA = 25 + (50 + 40 + 30 + 20)/10 = 39p (before: 37)

- OCX = 20 + (10 + 5 + 3)/10 = 21,8p (before: 23,5)

- HOT = 15 + (2 + 1)/10 = 15,3p (before: 15,6)

 

//EDIT2: Second case study - someone beating other overclockers and team//

 

Maybe a more relevant situation: what happens when HOT3 jumps from 9th to 3rd (and therefore HOT jumps over OCX to 2nd place in PowerTeam ranking):

 

Alternative 5:

 

- OCA = 50 + (40 + 20)/10 = 56p (before: 57)

- HOT = 30 + (2 + 1)/10 = 30,3p (before: 3,3)

- OCX = 10 + (5 + 3)/10 = 10,8p (before: 21,5)

 

Alternative 3:

 

- OCA = 25 + (50 + 40 + 20)/10 = 36p (before: 37)

- HOT = 20 + (30 + 2 + 1)/10 = 23,3p (before: 15,6)

- OCX = 15 + (10 + 5 + 3)/10 = 16,8p (before: 23,5)

 

In alternative 3, beating any given overclocker doesn't change that much to the ranking, but improving your team's PowerTeam ranking is highly rewarding. In alternative 5, the entire team total goes down quite rapidly if only ONE other person beats some scores.

[Lower]

damn, can I change my Voting to NR 3?

 

it sounds more better to me

 

I have choosed Nr. 2

[Patch]

Sorry guys, but option 6 was the only viable option for me (though I don't agree with it either).

 

You're using a sledgehammer to crack a walnut with these changes. Sure, you'll get the meat out. But it will be all squished and un-appetizing.

 

These revisions take the fun out of the team game. Pure and simple.

 

We bench hard for points. Globals. Hardware. Team points. Nobody, and I mean NOBODY, is going to appreciate the watering down or elimination of the significance of their efforts.

 

Right now the bot is a worldwide free-for-all where you can compete for individual glory AND band together for some tribal glory. It's fun.

 

Some of the awesome teams we've been chasing that are filled with armies of "little guys" are gonna be crushed by these revisions. That's sad.

 

Sure, hardware sharing sucks. But making the whole game less fun for the vast majority of participants sucks more.

[drnip]

Well said Patch.

[Gautam]

I think I prefer 3, but then again, massman might have just indoctrinated me.

[voidn]

Not trying to be a hater, but the Original Doc did not clearly define power teams. "- amount of participating teams (determines weight)", don't think that was clear until yesterday.

 

Can we get a "real world" example of Power Team points, from a popular category. Say i7 920 wprime 32m?

[Massman]

Can we get a "real world" example of Power Team points, from a popular category. Say i7 920 wprime 32m?

 

In comparison to the current situation? And in comparison to some other alternatives?

[voidn]

Feed existing data (submissions) into the purposed Alternative 1. And then Alternative 3 and 5 if you can.

 

Comparing to current team points would probably be a bit tough, but would give a better idea of what would change.

[S_A_V]

Hard to understand without 0.5L of vodka

For the first look I like Alternative 5 but need to think more before make final choice.

[KONAKONA]

Voted for Alternative 6: Remove points

[Massman]

'Feed' = manual calculations in excel, in this case

 

I'll give it a shot.

[voidn]

'Feed' = manual calculations in excel, in this case

 

I'll give it a shot.

 

You could query the database / write a report, instead of using excel . If you run mysql I can help. Just give me access .

[der8auer]

Pieter, thanks for this thread and vote Realy nice step!

Edited October 6, 2010 by der8auer

[Massman]

Getting the information is not the problem, it's the hwboint algorithm that is. I need to calculate the weight and individual points for each record as the algorithm does not support team rankings.

 

Anyways, I have an approx ranking now ... the question is whether I should make it look better or not. A simple trick with numbers (multiplying the end results x10) makes ALL the teams GAIN points instead of lose points in comparison to the current system. Some teams just gain more points than the others.

 

Algorithm as in Alternative 3:

 

[Team points = "PowerTeam points" + SUM(all points) / 10] is equal to [Team points = 10x "powerteam points" + sum(all points)]

 

Ah, I'll just add them both so you can see it with your own eyes

[Massman]

Pieter, thanks for this thread and vote Realy nice step!

 

At least I try, right? .

 

FYI, people voting for Alternative 2 are basically saying that currently there's no problem at all. Given the latest complain-threads and -articles, I find that very strange.