Voting systems on parties
category: general [glöplog]
Quote:
while the newbie group with the most buddies at the partypalce won 1st place
Reminds me of "raum, klang und design" which wasn't very great but had great successors. ;)
Quote:
As for "intelligent sceners, who can appreciate quality entries" - well, some can and some seem to be able only to vote for the friends (or even against their enemies). Just look at BP04 results.
Do I read here some "Why has Metalvotze won the pc democompo?" once again? If yes than ask yourself if it is ok to vote for entertaining demos rather than for tech demos which show you for the 1000th time a rotating cube or some 3d-stuff without any sense.
The problem with grading each demo is that you're vote counts more if you use the high end of the spectrum (like over-optimistic people would vote) - so it would probably be a good idea to "normalize" each voters result so it comes to an average of 3 before counting the totals...
This means that if one people gave demo1 a 5 and demo2 a 4 and another 5 and 1 their votes would still count the same.
This means that if one people gave demo1 a 5 and demo2 a 4 and another 5 and 1 their votes would still count the same.
Do I read here some "Why has Metalvotze won the pc democompo?" once again? If yes than ask yourself if it is ok to vote for entertaining demos rather than for tech demos which show you for the 1000th time a rotating cube or some 3d-stuff without any sense.
The question remains the same:
Do you vote for the best demo or your favorite?
The question remains the same:
Do you vote for the best demo or your favorite?
Sasq: that would basically make the system the same as voting for top3, except that you would have the choice to rank either none or all.
mmm st(art)2004 with the bracelets..... >:(
No matter what system you use, there will nevertheless always be people releasing fun, crap and entertaining productions. Can't you consider that there are still sceneres that see it more important to participate than to win? Some might convince a lot of friends to vote their crap-prod. Others get their votes just because of their famous name. Both facts divert the result. So there is no need to complain. Preselection is only good if you have too many entries where it would just be impossible to show them all.
The questions is why discussiong about voring rules at all? What is the aim of such a discussion? It's a feeling of being handled unfair, right? Well that is just the difference when you show your prod in front of experts or a wide spread audience.
It's like with the c64 graphics competitions, there are often workingsteps that don't really convince me, but those productions get the upper places. That is what the audience nevertheless is voting for, basta. Important for me is not this voting feedback, but the feedback you get from the sceners personally, and having a chat with the other competitors, that is also the best way to improve and get critics.
So in the end, is it really the fault of the audience, or is it the fault of the competitors who hand in the same and same style (converted) productions again and again.
And now i suggest another thread: 'What is killing the scene' :-)
The questions is why discussiong about voring rules at all? What is the aim of such a discussion? It's a feeling of being handled unfair, right? Well that is just the difference when you show your prod in front of experts or a wide spread audience.
It's like with the c64 graphics competitions, there are often workingsteps that don't really convince me, but those productions get the upper places. That is what the audience nevertheless is voting for, basta. Important for me is not this voting feedback, but the feedback you get from the sceners personally, and having a chat with the other competitors, that is also the best way to improve and get critics.
So in the end, is it really the fault of the audience, or is it the fault of the competitors who hand in the same and same style (converted) productions again and again.
And now i suggest another thread: 'What is killing the scene' :-)
At Forever we have scale 0-9 for every production. While the productions are shown people have their votesheets and in short few-sec breaks between prods they give the points to the shown production. For results we count sum of points for production from every votesheet .
Actualy with this system it doesn't matter if a votes ir optimistic (giving points scaling from 6-9) or pesimistic (1-5). This system also gives possibility for very close wins (usually few points) and also even for undobtedly wins (first entry 30 and more points away from the other). It simply depends on how much impact the production have at the time when it's shown. Of course if something stands out, it is usually clearly visible in the differences between the minimal and maximum scores given by the signle visitor.
Also we have internal rules which disqualify votesheets which don't have all shown entries in the cathegory ranked, or which have manipulative voting like 0,0,0,0,0,9,0,0,0,0 .
This kind of voting has only flaw, it's realtively time consuming to count all the votes especially when you have 30 and more entries in some of the 12 cathegories on about 70-80 votesheets. But IMO it's fair even with mixed audiences from 3 different 8 bit plattforms (Commodore 64, Atari & ZX Spectrum).
Actualy with this system it doesn't matter if a votes ir optimistic (giving points scaling from 6-9) or pesimistic (1-5). This system also gives possibility for very close wins (usually few points) and also even for undobtedly wins (first entry 30 and more points away from the other). It simply depends on how much impact the production have at the time when it's shown. Of course if something stands out, it is usually clearly visible in the differences between the minimal and maximum scores given by the signle visitor.
Also we have internal rules which disqualify votesheets which don't have all shown entries in the cathegory ranked, or which have manipulative voting like 0,0,0,0,0,9,0,0,0,0 .
This kind of voting has only flaw, it's realtively time consuming to count all the votes especially when you have 30 and more entries in some of the 12 cathegories on about 70-80 votesheets. But IMO it's fair even with mixed audiences from 3 different 8 bit plattforms (Commodore 64, Atari & ZX Spectrum).
Of course it matters if you vote optimistically or pessimitically.
Lets take an example:
At a party, the large crowd of old-skool guys considers the not-so-impressive contributions to the compo and grades them thoughtfully and correctly, averaging at about 3.
Meanwhile, a smaller crowd of young gamers really loved the Scooter-inspired techno tune in one of the ugly and technically under-acheiving demos, and happily puts down 10s for that demo.
Result: The bad demo with the loud techno tune wins even though more people prefered another demo.
Lets take an example:
At a party, the large crowd of old-skool guys considers the not-so-impressive contributions to the compo and grades them thoughtfully and correctly, averaging at about 3.
Meanwhile, a smaller crowd of young gamers really loved the Scooter-inspired techno tune in one of the ugly and technically under-acheiving demos, and happily puts down 10s for that demo.
Result: The bad demo with the loud techno tune wins even though more people prefered another demo.
democracy is still the less worst voting system...
I can suggest another voting system:
- each person gets 2 (random!) productions on the screen, and must choose which one they like better;
- this process continues to the first step as long as the person likes, that is one can vote as often or as rarely as one likes;
- it loops only after all pairs have been tried - tracking should be done per machine (like with cookies);
This system has some advantages:
- it forces one to decide which production is better, and since you are always comparing 2 prods "face 2 face" cheating is very unlikely, and even a drunk person is somewhat more likely to choose "wisely" because they must meet a very simple decision;
- people without a computer can use a public machine to vote without the need for a key;
- works well for big parties and many visitors, where statistics starts to work.
For smaller parties, i agree that a system which requieres to rate each prod is the best. It also works well with drunk visitirs since it requieres a set of fairly simple decisions. As opposed to that, picking up the best n is always very complicated.
I think also authors should be forced to submit a screenshot with their prod to be used in a vote, because names are often hard to remember and "professional voters" start keeping a list with keywords to each demo. The system i propose requieres either a screenshot, or at least a detailed description of contents, otherwise it would be highly frustrating! If both screenshots and descriptions are allowed, prods without a screenshot are likely to be voted down.
- each person gets 2 (random!) productions on the screen, and must choose which one they like better;
- this process continues to the first step as long as the person likes, that is one can vote as often or as rarely as one likes;
- it loops only after all pairs have been tried - tracking should be done per machine (like with cookies);
This system has some advantages:
- it forces one to decide which production is better, and since you are always comparing 2 prods "face 2 face" cheating is very unlikely, and even a drunk person is somewhat more likely to choose "wisely" because they must meet a very simple decision;
- people without a computer can use a public machine to vote without the need for a key;
- works well for big parties and many visitors, where statistics starts to work.
For smaller parties, i agree that a system which requieres to rate each prod is the best. It also works well with drunk visitirs since it requieres a set of fairly simple decisions. As opposed to that, picking up the best n is always very complicated.
I think also authors should be forced to submit a screenshot with their prod to be used in a vote, because names are often hard to remember and "professional voters" start keeping a list with keywords to each demo. The system i propose requieres either a screenshot, or at least a detailed description of contents, otherwise it would be highly frustrating! If both screenshots and descriptions are allowed, prods without a screenshot are likely to be voted down.
nice idea but it's very timeconsuming for more than >10 releases :/
and if all those votesheets contained a screenshot in addition to the title/group - that'll do great! especially at parties like BP, where many demos == many forgotten treasures :-)
can't we just keep it as is...
its way less of a hassle to complain afterwards how fuckin' stupid the audience was again ;-)
its way less of a hassle to complain afterwards how fuckin' stupid the audience was again ;-)
Gargaj: the idea is, that it would give fairly correct results for a big number of visitors even if not every visitor goes through every combination. Which is, by itself, nearly impossible, because i think for n productions it would take n*(n-1)/2 comparisons to compare all prods with each other. However, a complete set of preferances can be created probably with logarithmical complexity (not sure, have to think about that).
Could someone with better knowledge in statistics say how feasible this system is, that is compute an error depending on the number of productions and comparison steps? Any heuristics on how to select the productions for comparison? How to compute the total result?
I'm not saying it's any good, nor do i have enough understanding on how to make it work right, i just thought it was interesting so i post it. I cannot decide whether it could perform better than "rate each prod" scheme for large parties like BP, but perhaps it's better than "top 3", which forces many people not to vote for good prods because they might think that other prods would become underrated (shameful 0 points!) otherwise.
I think we must also distinguish 2 types of potential voters - some voters want to rate each prod, others are too drunk or uninterested for that. Perhaps these 2 systems could somehow be made to compliment each other, with e.g. results of "drunkvoting" making out 20% of the result, or in some other manner. For example, 2n decisions (n being the number of prods) could make up for one list-rating voter.
Could someone with better knowledge in statistics say how feasible this system is, that is compute an error depending on the number of productions and comparison steps? Any heuristics on how to select the productions for comparison? How to compute the total result?
I'm not saying it's any good, nor do i have enough understanding on how to make it work right, i just thought it was interesting so i post it. I cannot decide whether it could perform better than "rate each prod" scheme for large parties like BP, but perhaps it's better than "top 3", which forces many people not to vote for good prods because they might think that other prods would become underrated (shameful 0 points!) otherwise.
I think we must also distinguish 2 types of potential voters - some voters want to rate each prod, others are too drunk or uninterested for that. Perhaps these 2 systems could somehow be made to compliment each other, with e.g. results of "drunkvoting" making out 20% of the result, or in some other manner. For example, 2n decisions (n being the number of prods) could make up for one list-rating voter.
discard the voting sheets that have vomit/alcohol marks on them ?
how about letting the pre-selection crew rate the demos with a weight of 50% and let the crowd have the rest?
this might be a bit harsh, but taking the example from assemblys separated music-competitions; "vocal music" and "instrumental music". now i'm not suggesting that we separate the demo-compo into "funny entries" or "tech entries", but more in the way of finding a way of judging them more fairly - like the scene.org-awards.
say, for example, that one had to select top five demos in different categories - graphics, code, music, atmosphere, etc.
one would of course have to figure out a clever way of calculating the different scores, but.. what the hell, it was just an idea. :)
say, for example, that one had to select top five demos in different categories - graphics, code, music, atmosphere, etc.
one would of course have to figure out a clever way of calculating the different scores, but.. what the hell, it was just an idea. :)
