bobc36
ACX Member
- Total Posts : 346
- Reward points : 0
- Joined: 2009/06/02 07:50:47
- Location: Pittsburgh, Pennsylvania
- Status: offline
- Ribbons : 2
Re:Chimp Challenge 2012
2012/02/24 05:33:17
(permalink)
What about a modified points total for each team? Instead of every team trying to reach say 200 mil or whatever, take the total monthly production for the past 6 months and the previous CC (for each individual team), average that all together and say "you have to make 75% of your total to win." for each team.
Each team would have a different target, but it should take roughly the same amount of time to finish. Like saying Lance Armstrong has to bike one leg of the TDF and I have to run a mile. It would take us both roughly the same time, but it would promote each team to ramp up production so that they could finish just a bit faster.
No goofy conversion factors based on arbitrary values.
Another idea would be to take the amount of points made in last years CC (per team) and multiply it by 1.25 (kind of a growth allowance) and say that is your target this year.
<--- See those Folding numbers? Ask me about how you can get them too! Rigs: 2P E5-2650 Xeon 
|
diplomacy
Superclocked Member
- Total Posts : 198
- Reward points : 0
- Joined: 2011/12/11 07:46:12
- Status: offline
- Ribbons : 0
Re:Chimp Challenge 2012
2012/02/24 06:27:01
(permalink)
I think we are moving in the right direction with simpler ideas.
the idea of having different goals but preserving the "all out race to finish idea" is my favorite so far. it does everything we need. its scalable, it is team specific, its fair and it goes back to the history of the challenge... we tried the goofy formulas last year, it didn't work, back to basics.
ASUS P9x79 Pro with Intel I7 3930k 4.2 Ghz, Dual EVGA GTX 580s SLI I use Bionic to help save the world while I'm sleeping, watching TV, or pretty much doing anything but gaming on my PC. 
|
troy8d
CLASSIFIED Member
- Total Posts : 2185
- Reward points : 0
- Joined: 2010/10/16 08:10:22
- Status: online
- Ribbons : 10
Re:Chimp Challenge 2012
2012/02/24 06:41:40
(permalink)
diplomacy
I think we are moving in the right direction with simpler ideas.
the idea of having different goals but preserving the "all out race to finish idea" is my favorite so far. it does everything we need. its scalable, it is team specific, its fair and it goes back to the history of the challenge... we tried the goofy formulas last year, it didn't work, back to basics.
What you probably don't realize is that this idea is the same exact formula as last year, just approached from a different angle. It will be subject to the same flaws and inadequacies if not corrected.
|
diplomacy
Superclocked Member
- Total Posts : 198
- Reward points : 0
- Joined: 2011/12/11 07:46:12
- Status: offline
- Ribbons : 0
Re:Chimp Challenge 2012
2012/02/24 07:30:46
(permalink)
true, i didn't realize that, because I wasn't here last year, so i have no frame of reference, but if thats the formula that has everyone so upset, then it seems like you guys are getting upset over nothing.
everyone was acting like someone created a system specifically to punish EVGA for being way out in front.
if the system was only designed to garner all-out folding participation from everybody, then i don't see the problem.
ASUS P9x79 Pro with Intel I7 3930k 4.2 Ghz, Dual EVGA GTX 580s SLI I use Bionic to help save the world while I'm sleeping, watching TV, or pretty much doing anything but gaming on my PC.
|
troy8d
CLASSIFIED Member
- Total Posts : 2185
- Reward points : 0
- Joined: 2010/10/16 08:10:22
- Status: online
- Ribbons : 10
Re:Chimp Challenge 2012
2012/02/24 08:06:39
(permalink)
diplomacy
true, i didn't realize that, because I wasn't here last year, so i have no frame of reference, but if thats the formula that has everyone so upset, then it seems like you guys are getting upset over nothing.
everyone was acting like someone created a system specifically to punish EVGA for being way out in front.
if the system was only designed to garner all-out folding participation from everybody, then i don't see the problem.
Last years formula was extremely flawed. Unfortunately, it took a completely, arbitrarily unbalanced contest to prove this to some people. Personally, I believe it was ignorance/incompetence on the part of last years captains. What got people around here upset is that we were left entirely out of the discussion last year and when we were finally clued in, no one was willing to listen to reason on a system that was extremely flawed. But other people have differing opinions and everyone is going to believe what they want to believe... The past, however, isn't whats relevant. We need to focus on making a fair and engaging Chimp Challenge this year...
|
zodac
New Member
- Total Posts : 73
- Reward points : 0
- Joined: 2009/07/10 08:58:25
- Status: offline
- Ribbons : 0
Re:Chimp Challenge 2012
2012/02/24 14:37:43
(permalink)
Can I just make a note of something? There are varying opinions on how the CC should be done this year: • People saying that we should split the CC trophy into several smaller trophies (for PPD, most users, etc). • People thinking we should have PPD brackets. • People who feel we should drop the team name altogether. However, we have just 6 weeks (realistically) to get team threads up and begin recruiting. Trying to get 9 different teams and their users to agree to such drastic changes is unlikely in that timeframe. One team will disagree with one facet, another will disagree on another, and then two/three teams won't reply until we're all decided, and we'll have to go back and readjust again. The formula is the one part that 100% does need to be changed this year. I feel we should all try and fix that ASAP, and come to a full agreement on the matter, and then consider changing the above issues, if time permits. We'll survive this year having to use the CC team names; we won't survive teams not agreeing on the formula used.
post edited by zodac - 2012/02/24 14:42:28
|
bowlinra
SSC Member
- Total Posts : 886
- Reward points : 0
- Joined: 2011/12/05 20:58:14
- Location: Virginia, USA
- Status: offline
- Ribbons : 5
Re:Chimp Challenge 2012
2012/02/25 00:36:57
(permalink)
zodac
Trying to get 9 different teams and their users to agree to such drastic changes is unlikely in that timeframe. One team will disagree with one facet, another will disagree on another, and then two/three teams won't reply until we're all decided, and we'll have to go back and readjust again.
What was wrong with the Olympics approach? Have each team propose 2 events, then agree to down select to the best 7 events by vote.
|
zodac
New Member
- Total Posts : 73
- Reward points : 0
- Joined: 2009/07/10 08:58:25
- Status: offline
- Ribbons : 0
Re:Chimp Challenge 2012
2012/02/25 00:58:39
(permalink)
bowlinra
zodac
Trying to get 9 different teams and their users to agree to such drastic changes is unlikely in that timeframe. One team will disagree with one facet, another will disagree on another, and then two/three teams won't reply until we're all decided, and we'll have to go back and readjust again.
What was wrong with the Olympics approach?
Have each team propose 2 events, then agree to down select to the best 7 events by vote.
Many people don't believe we should split the CC up into different events. You're first gonna have to convince the majority to agree to that, then get each team to submit two unique events, and then have them voted upon. And if one of those events is "best production increase", it's gonna be the same as last year, where it's biased towards smaller teams, and that's likely for most events; it'll favour smaller or bigger teams rather than being balanced.
|
diplomacy
Superclocked Member
- Total Posts : 198
- Reward points : 0
- Joined: 2011/12/11 07:46:12
- Status: offline
- Ribbons : 0
Re:Chimp Challenge 2012
2012/02/25 02:37:36
(permalink)
it'll favor smaller or bigger teams rather than being balanced.
I don't think its possible to be "balanced" in this instance. bigger teams have an advantage. pretending that they don't is dishonest. Doing something about that will favor smaller teams.
A simple points multiplier of .01234 across all the teams would be "fair" and would have the added benefit of knocking most of the teams down far enough that it would look like the other teams are close when in fact they aren't. a logarithmic equation could do that even better. it would look like the 1st, 2nd and 3rd place team were within .01 pts of each other.
Once you get into changing the order though, it will never be fair. you will have to favor one group or the other. So you may as well be blatant about backing one horse or the other. If a team of thirty has any chance of beating 1000 dedicated folders with 4000 dedicated rigs, then it cannot be a fair contest based on ppd. it just cant. It can be a fair contest in other categories like growth yoy, but we have to be honest that then we aren't having a straight ppd contest. We won't win. Most of our folders fold year 'round and can't up our ppd by 600% overnight, or triple our team's size. smaller teams can do that.
That said, I'd rather see us back the point system that is in our favor. Straight PPD, maybe even a multiplier bonus based on participation. If we don't get what we want, we can either not participate or we can be good sports and participate in the one contest out there that isn't a straight ppd race.
ASUS P9x79 Pro with Intel I7 3930k 4.2 Ghz, Dual EVGA GTX 580s SLI I use Bionic to help save the world while I'm sleeping, watching TV, or pretty much doing anything but gaming on my PC.
|
Re:Chimp Challenge 2012
2012/02/25 04:56:44
(permalink)
Sorry diplomacy, but what you're saying is contradicting your forum name  I am for a competition that gives a winning chance to all teams, not something that blatantly favors one or few teams. So far, the points-based system that favored large teams and the handicap system that favored small teams both failed. The formula proposed by OCN is too arbitrary with how it puts together different measures, although the measures there all make sense individually. The thing is, before I can start discussing or working on a formula, I need to know what exactly does measures are? 1) What is "Handicap"? I feel like it's the team PPD over a certain period of time divided by 10. Is this correct? If so, why divide by 10? 2) What is "Conversion"? Is it the CC-name points divided by team points, over the duration of the competition? 3) What is "Points"? Is it the CC-name points over the duration of the competition? For any formula to be meaningful and fair with these measures, I strongly believe that the competition should run at least 20 days. I strongly suggest 30 days. This way, we can nullify the effect of transient events such as 1) Priming WUs for bigadv, 2) Some donors changing teams for a few days just to be able to participate in the competition. Some people will still prime their WUs, there is no way to control it. But, we can make their affect close to zero by making the competition longer. Some donors will still change their teams to participate in the competition, and I believe it is very fine on a individual level. However, they do cause an imbalance in the competition and their affect would be reduced if we run the competition longer, because fewer of them would have incentive to change teams for longer stretches of time. If they really want to participate, of course they can still do that regardless of the duration. Now, for a fair system, I think rather than thinking "who should have won last year", we should think "what would it take to win"? Looking at the numbers, I agree that OCN should have won last year. The thing is I can come up with *many* formulas that favor OCN's profile last year but only the one answering the second question fairly matters.
|
troy8d
CLASSIFIED Member
- Total Posts : 2185
- Reward points : 0
- Joined: 2010/10/16 08:10:22
- Status: online
- Ribbons : 10
Re:Chimp Challenge 2012
2012/02/25 04:59:53
(permalink)
The only thing that makes sense to me in your post is that decreasing the point scale will do something mask the points difference for people that don't watch too closesly. That is not what we are trying to achieve.
I couldn't follow any of the rest of it.
|
Re:Chimp Challenge 2012
2012/02/25 05:07:20
(permalink)
troy8d
The only thing that makes sense to me in your post is that decreasing the point scale will do something mask the points difference for people that don't watch too closesly. That is not what we are trying to achieve.
I couldn't follow any of the rest of it.
I think you're talking about diplomacy's post, but it looks like you're talking about mine 
|
troy8d
CLASSIFIED Member
- Total Posts : 2185
- Reward points : 0
- Joined: 2010/10/16 08:10:22
- Status: online
- Ribbons : 10
Re:Chimp Challenge 2012
2012/02/25 05:10:18
(permalink)
You posted as I was typing 
|
zodac
New Member
- Total Posts : 73
- Reward points : 0
- Joined: 2009/07/10 08:58:25
- Status: offline
- Ribbons : 0
Re:Chimp Challenge 2012
2012/02/25 07:46:18
(permalink)
Just something I came across today. I've been testing some new equations - same variables, just manipulated differently. I actually had one which seemed very promising.
However, when I tested it with current team PPD and the baseline based off last year's productiom, the results were completely messed up. I saw the same with other formulae.
The conclusion I've come to is that using last year's stats for the 'handicap' is a bad idea - teams' PPD have changed too much. Using a more recent period should give us more accuracy when we do end up happy with a formula.
Feel free to check yourself, but most formulae with those 3 variables - total points/handicap, conversion percentage, and # millions - became really unpredictable when mixing last year's performance with current production.
|
troy8d
CLASSIFIED Member
- Total Posts : 2185
- Reward points : 0
- Joined: 2010/10/16 08:10:22
- Status: online
- Ribbons : 10
Re:Chimp Challenge 2012
2012/02/25 07:54:42
(permalink)
zodac
Just something I came across today. I've been testing some new equations - same variables, just manipulated differently. I actually had one which seemed very promising.
However, when I tested it with current team PPD and the baseline based off last year's productiom, the results were completely messed up. I saw the same with other formulae.
The conclusion I've come to is that using last year's stats for the 'handicap' is a bad idea - teams' PPD have changed too much. Using a more recent period should give us more accuracy when we do end up happy with a formula.
Feel free to check yourself, but most formulae with those 3 variables - total points/handicap, conversion percentage, and # millions - became really unpredictable when mixing last year's performance with current production.
You are completely missing the point.
|
zodac
New Member
- Total Posts : 73
- Reward points : 0
- Joined: 2009/07/10 08:58:25
- Status: offline
- Ribbons : 0
Re:Chimp Challenge 2012
2012/02/25 08:03:31
(permalink)
I wasn't replying to one of your points. Xavier has mentioned his preference for basing this year's system of last year's production. I'm just saying, while that could work for a simple points/handicap system, I don't think it works as well when you add more factors.
Which point do you think I was missing though?
|
troy8d
CLASSIFIED Member
- Total Posts : 2185
- Reward points : 0
- Joined: 2010/10/16 08:10:22
- Status: online
- Ribbons : 10
Re:Chimp Challenge 2012
2012/02/25 08:07:29
(permalink)
zodac
I wasn't replying to one of your points. Xavier has mentioned his preference for basing this year's system of last year's production. I'm just saying, while that could work for a simple points/handicap system, I don't think it works as well when you add more factors.
Which point do you think I was missing though?
First, that is the same thing that I am saying. Second, you assume that there is a direct correlation between monthly production and chimp challenge production. Just because you are comparing it the other way doesn't change the results. In fact, you only confirm why monthly points is so fail. Handicap = Monthly Points -> not correlate with CC points Handicap = CC Points -> not correlate with monthly points ZzZzZzZzZzZz....is there are 3rd way we can make the same comparison to realize it doesnt work?
|
zodac
New Member
- Total Posts : 73
- Reward points : 0
- Joined: 2009/07/10 08:58:25
- Status: offline
- Ribbons : 0
Re:Chimp Challenge 2012
2012/02/25 08:18:20
(permalink)
Honerstly, I don't understand what you're saying.
If you mean current production is a poor estimation of CC production, of course it is - it underestimates teams.
But if there are issues with current PPD, won't CC PPD just make it worse?
If that's not your point, I apologise, and ask that you dumb it down for me a bit, because I'm clearly missing something...
|
troy8d
CLASSIFIED Member
- Total Posts : 2185
- Reward points : 0
- Joined: 2010/10/16 08:10:22
- Status: online
- Ribbons : 10
Re:Chimp Challenge 2012
2012/02/25 08:20:05
(permalink)
I think we're both missing something at the moment. 
|
troy8d
CLASSIFIED Member
- Total Posts : 2185
- Reward points : 0
- Joined: 2010/10/16 08:10:22
- Status: online
- Ribbons : 10
Re:Chimp Challenge 2012
2012/02/25 08:25:03
(permalink)
I think the best thing to do would be for me to formulate what I'm thinking and see if there is an improvement. I've got some of the historic data, but probably not as much as you have. Any extra information you can provide would be greatly appreciated.
|
zodac
New Member
- Total Posts : 73
- Reward points : 0
- Joined: 2009/07/10 08:58:25
- Status: offline
- Ribbons : 0
Re:Chimp Challenge 2012
2012/02/25 08:25:37
(permalink)
If we agree that I'm right, we can stop and go relax instead. ^_^
|
zodac
New Member
- Total Posts : 73
- Reward points : 0
- Joined: 2009/07/10 08:58:25
- Status: offline
- Ribbons : 0
Re:Chimp Challenge 2012
2012/02/25 08:34:33
(permalink)
I think the best thing to do would be for me to
formulate what I'm thinking and see if there is an
improvement. I've got some of the historic data, but
probably not as much as you have. Any extra
information you can provide would be greatly
appreciated. I haven't really got more than the top section of the pic Xavier posted of the first formula, but send me a PM, and I'll pass on what I've got when I get home.
|
pgajr
Superclocked Member
- Total Posts : 188
- Reward points : 0
- Joined: 2008/12/04 10:55:35
- Location: Membertou, NS, Canada
- Status: offline
- Ribbons : 0
Re:Chimp Challenge 2012
2012/02/25 10:06:59
(permalink)
The only way I can see the Chimp Challenge being a fair race is if no one knows who they are folding for. All folders get put into a pool and evenly broken up into teams equal to that of teams that are entering. These folders then fold for a point amount over a period of time. the numbered team that get the most points or reaches the goal first wins. Then the team number is matched to a team name and they are the chimp challenge winner. This would make it all about the folding and not about who has more of whatever. No point calculating, no real number crunching just folding. Isn't that what the challenge is all about?
|
Xavier Zepherious
CLASSIFIED Member
- Total Posts : 4632
- Reward points : 0
- Joined: 2010/07/04 12:53:39
- Location: Medicine Hat ,Alberta, Canada
- Status: offline
- Ribbons : 15
Re:Chimp Challenge 2012
2012/02/25 10:18:20
(permalink)
if you go by team numbers OCN - 100M PPM bump since HPCS last 6 months only 2 months show the higher avg - so the other months drop your estimated PPD for the CC 390AVG - 520PEEK - 25% swing on EVGA PPD is about the same - no bump (avg 600M PPM over 6 months) however oct-nov contest months - are peek ...as well as march(MM) and june-CC 660 PEAK - 600AVG 10% swing so lets say you do you 520(your peak) and we do 660(our peek) for the CC you get a 15% advantage(over EVGA) because your average is lowballed by the off months if you use team PPD - use best peek weeks(2 weeks - same length of contest) that will give the best PPD a team can give period - barring improvements or new recruitment if we go by past CC values OCN had 311M PPM evga had 608M PPM even if HPCS is a factor for us - it won't be as much both [H] and us are not noticing the HPCS bump as much if any indication of out peak OCT month 660 we are only going to do about the same -even if HPCS a factor maybe 5-10% more where as OCN will get a 311M- 511M(520M - month not over) boost -40% maybe not so wise (seeing how HPCS will skew things)
post edited by Xavier Zepherious - 2012/02/25 10:19:36
|
Punchy
CLASSIFIED Member
- Total Posts : 2872
- Reward points : 0
- Joined: 2010/02/06 09:33:05
- Status: online
- Ribbons : 14
Re:Chimp Challenge 2012
2012/02/25 13:57:13
(permalink)
I have a few ideas, but I'm only going to release them one at a time to retain focus.
Idea 1: I don't think the Chimp Challenge should be run at a time when the free HP Cloud service is available, nor should any calculations be based on team output during times when it was.
Rationale: Since we don't know any of the dates when it will go public (still free) or when it will go paid, there will be major inflection points that could fall within the CC and totally destroy the value of any formulas.
Having it run while the free cloud is available simply would encourage abuse of HP's generosity. Does anyone really think beta-testing by running F@H is of any use to HP? And as an individual beta tester, if you really were beta testing, does running 6*N instances really tell you more than running 6 (the maximum you should be running).
I wonder how many cloud users are still exceeding 20GB per zone, even though HP on Feb 13 requested that users comply "as soon as possible". I guess we'll know in 2 days when they start enforcing the restrictions by terminating instances in excess of their limit.
|
Xavier Zepherious
CLASSIFIED Member
- Total Posts : 4632
- Reward points : 0
- Joined: 2010/07/04 12:53:39
- Location: Medicine Hat ,Alberta, Canada
- Status: offline
- Ribbons : 15
Re:Chimp Challenge 2012
2012/02/25 14:49:41
(permalink)
further
MPC output has barely changed in a whole year
Custom PC and Bit-tech's peek is during CC last year ...nearly Double of their yearly avg
TSC peek is 235M and avg is around 190M - 10%
HWC peek(without NCIX) is 84M and their 6 month best avg is 82M
a 10% drop of of NCIX during CC- and had HWC a 20% boost during CC
how do you propose a new formula will work if a concerted attempt is put thru- most of NCIX comes over - HWC would be double their last CC?
tech power up is currently at 24M(avg is 1/2 of peek month or 32) vs 62M during CC
so taking 6 month avg is not going to work
it benefits teams that have low PPD from peak months and rarely compete (or run internal challenges)
2 months would still be unfair
that leaves each teams best two week peek PPD
(no teaming up)
stragglers are fine(co opting more users) - as long as it's not a concerted effort(by whole teams or organized by teams) to beat the system we set up.
In fact Id rather we not steal each others folders but rather go out and get new ones (people who don't know what folding is)
It's not good to steal people from other folding teams but to bring new people in that never folded.
Fair play - and the cessation of bad feelings by stealing members or teaming up to defeat others will be put to rest..
This is what harbors bad feelings between teams
if you do it... you do it strictly on your own power and ability to get new recruits - fair play
NO organizing in a way that make the competition unfair -to beat the formula and/or rules
|
diplomacy
Superclocked Member
- Total Posts : 198
- Reward points : 0
- Joined: 2011/12/11 07:46:12
- Status: offline
- Ribbons : 0
Re:Chimp Challenge 2012
2012/02/26 02:30:42
(permalink)
Sorry diplomacy, but what you're saying is contradicting your forum name
ah, i created that when i was 13 or so, back when i was too young and inexperienced to wield an axe. over the years, the name has grown more and more ironic 0.o
ASUS P9x79 Pro with Intel I7 3930k 4.2 Ghz, Dual EVGA GTX 580s SLI I use Bionic to help save the world while I'm sleeping, watching TV, or pretty much doing anything but gaming on my PC.
|
zodac
New Member
- Total Posts : 73
- Reward points : 0
- Joined: 2009/07/10 08:58:25
- Status: offline
- Ribbons : 0
Re:Chimp Challenge 2012
2012/02/26 04:18:16
(permalink)
Xavier Zepherious
so taking 6 month avg is not going to work
it benefits teams that have low PPD from peak months and rarely compete (or run internal challenges)
2 months would still be unfair
that leaves each teams best two week peek PPD
(no teaming up)
Punchy
I have a few ideas, but I'm only going to release them one at a time to retain focus.
Idea 1: I don't think the Chimp Challenge should be run at a time when the free HP Cloud service is available, nor should any calculations be based on team output during times when it was.
Rationale: Since we don't know any of the dates when it will go public (still free) or when it will go paid, there will be major inflection points that could fall within the CC and totally destroy the value of any formulas.
Having it run while the free cloud is available simply would encourage abuse of HP's generosity. Does anyone really think beta-testing by running F@H is of any use to HP? And as an individual beta tester, if you really were beta testing, does running 6*N instances really tell you more than running 6 (the maximum you should be running).
I wonder how many cloud users are still exceeding 20GB per zone, even though HP on Feb 13 requested that users comply "as soon as possible". I guess we'll know in 2 days when they start enforcing the restrictions by terminating instances in excess of their limit.
An answer to both the above quotes. Last year, we took a 5-month average over the Oct-Feb period. Your complaints are that that might be too long (Xavier) and that HPCS will skew things (Punchy). So to avoid HP's influence, any period would need to be before January. So what's wrong with Oct-Nov? It's usually the best period for most teams (except OCAU, since it's summer for them). It's the most recent timeframe we're able to use (unless you wanna go halfway into Jan, since HPCS only came into play towards the end of the month), and I don't believe a 2 week period is long enough to get any meaningful information. Xavier Zepherious
how do you propose a new formula will work if a concerted attempt is put thru- most of NCIX comes over - HWC would be double their last CC?
.
.
.
stragglers are fine(co opting more users) - as long as it's not a concerted effort(by whole teams or organized by teams) to beat the system we set up.
In fact Id rather we not steal each others folders but rather go out and get new ones (people who don't know what folding is)
It's not good to steal people from other folding teams but to bring new people in that never folded.
Fair play - and the cessation of bad feelings by stealing members or teaming up to defeat others will be put to rest..
This is what harbors bad feelings between teams
if you do it... you do it strictly on your own power and ability to get new recruits - fair play
NO organizing in a way that make the competition unfair -to beat the formula and/or rules
That's really a separate issue to the stats we use. No matter what formula we have, or what timeframe we use, that's something we need to agree on. The HWC/NCIX issue might be the most obvious one, but most teams got extra users from teams not taking part in the CC. Aside from large collaborations as above (which we'll need to take into account, but as an exception), I think all we can do is ask that our users don't try and have people move over from other teams just for the CC. If those users themselves feel like they want to take part, fair enough, and we're not gonna get all our own users to listen to us, but if we can get the majority to focus on expanding our Folding teams through our own forums, I think that's as close to "fair" as we're gonna get. theGryphon
The thing is, before I can start discussing or working on a formula, I need to know what exactly does measures are?
1) What is "Handicap"? I feel like it's the team PPD over a certain period of time divided by 10. Is this correct? If so, why divide by 10?
2) What is "Conversion"? Is it the CC-name points divided by team points, over the duration of the competition?
3) What is "Points"? Is it the CC-name points over the duration of the competition?
Forgot to address these points yesterday; sorry. :o 1) For last year's CC, it was the average PPD over the period October-February. We've not decided on it this year (as you can see in the earlier part of my post), so feel free to mess around with different systems if you want. Something based off last year's CC and recent production might work out well. 2) Correct. 3) Correct again. theGryphon
Now, for a fair system, I think rather than thinking "who should have won last year", we should think "what would it take to win"? Looking at the numbers, I agree that OCN should have won last year. The thing is I can come up with *many* formulas that favor OCN's profile last year but only the one answering the second question fairly matters.
It sounds like we're thinking along the same lines, but you're wrong to think of it as two separate issues. See, I think both your questions are the same, just asked differently. "What would it take to win" is asking me what are the factors that matter in this competition. Only by using those factors can we come to our own, individual opinion on "who should have won last year". I mean, you won't find many people who think last year's formula was "fair", but very few have said why it wasn't. Was it unfair because your team should have come in first? Or was your team properly placed, but the teams in first and/or last weren't correct? Yes, the formula gave an advantage to smaller teams, but how big an advantage? What variables weren't included? Only by answering both of your questions together can we really come up with a system that is fair.
|
Xavier Zepherious
CLASSIFIED Member
- Total Posts : 4632
- Reward points : 0
- Joined: 2010/07/04 12:53:39
- Location: Medicine Hat ,Alberta, Canada
- Status: offline
- Ribbons : 15
Re:Chimp Challenge 2012
2012/02/26 06:34:39
(permalink)
I think you're not understanding me Zodac I said no to taking big sections of time and averaging PPD(not two months_ - take best peak 2 weeks only the time when every teams points were at their max(during contests) by averaging lower weeks with high ones you lowering their average PPD...this contest will be an all out race - not an average output one (yes this will benefit teams that don't run competitions) but its the most fair - and if you look at those teams their PPD didn't change much during the CC anyways and it will be more fair to teams that run more fullout (harder all year) why each teams has show their best peak PPD - hence we can establish a maximum output (where they had a team competition - running fullout). then the contest will be about beating your own best output and growing it also you include HPCS months (because you grew the team with it - now the goal is to estimate it and reflact it in your estimates no way we going down a road where other teams are gonna cry wolf because it's found out other teams greatly increased their output compared to others. (some might have none or little and others capitalized greatly)...and we threw out those months that reflected those numbers either we count it - or we tell the teams no using hpcs during the race if we ban it - then each team will have to enforce it - if there is any hint or suspicion of cheating you ban users (kick out cheaters) this would be harder hence I suggest we include the HPCS months to find the best two weeks By ignoring HPCS months zodac you give your team an advantage considering your team jumps by 20% from a month with no HPCS to one with HPCS (to your peak month) and now you want to throw that one away? you pick EVGA's best months(oct-nov) and not yours (jan-feb)? (BIAS again) - no way either we take every ones best months(or best 2 weeks) looking at all months including HPCS or it's not fair look at all the teams... most teams do not have quite that bump ------------------------------------- if we use (cc PPD/ best 2 week avg PPD) then are all going to be racing against a time we we achieved our peak for some it will be the last years CC for some like EVGA it will be TZC and for OCN it will be your last month(feb) or we do a straight point race based on our peak PPD sum( CC ppd - peak 2 week avg ppd) > 20 million or sum (cc ppd)- best two week output >0 or sum (cc PPD) > best two week total basically digging yourself out of your two week hole based on your best output
post edited by Xavier Zepherious - 2012/02/26 06:48:25
|
Punchy
CLASSIFIED Member
- Total Posts : 2872
- Reward points : 0
- Joined: 2010/02/06 09:33:05
- Status: online
- Ribbons : 14
Re:Chimp Challenge 2012
2012/02/26 06:44:43
(permalink)
Fixed.
post edited by Punchy - 2012/02/26 06:49:22
|
