Last Friday we wrapped up a round of community voting for The Hackaday Prize. The theme? Most Likely To Save The Planet. Now it’s time for some results:
The projects voted Most Likely To Save The Planet by the Hackaday.io community are, in order:
- Junkyard Hydrogen Fuel Cell
- Smart Dew Point Water Harvester
- Convert Farm Fields To NFT Hydroponics
- Soil Moisture Sensor For Agriculture
- Biohand – Low Cost 3D Printed Hand Prosthesis
- Grey Water And Urine Recycling Earth Boxes
- Project EDEN
- Urban Kitchen Garden
- NIRGM – Non-Invasive NIR Glucose Meter *
- Medical Plasma Sterilization
Congratulations to everyone who has a project that was voted up to the top. Even though these rounds of community voting don’t decided which projects make it into the Hackaday Prize semifinals, you’ve earned the respect of your peers and a nifty Hackaday Prize t-shirt.
Since NIRGM – Non-Invasive NIR Glucose Meter won last week, we’re moving down the list to #11 and awarding SciPlo a Hackaday Prize t-shirt as well.
A New Round Of Voting!
This week, we’re asking the Hackaday.io community to vote for the most Amazingly Engineered project entered into the Hackaday Prize. To entice everyone to vote, I’m going to pick a random Hackaday.io user next Friday around 22:00 UTC. If that person has voted, they get a $1000 gift card for the Hackaday Store. If that person has not voted, I’ll be giving a few prizes away to people who have voted. Last week, we gave away a SmartMatrix, an Analog Stepper Gauge, and a Simon Says kit. We’ll probably change that up this week; I don’t know what it will be, but someone who votes will get something. Imagine; giving away stuff just for clicking a button. How magnanimous can we get?
As with every community voting update, it is requested that you vote. If you need a nudge to understand how this works, here’s a video tutorial on how to vote.
I’m really glad to see all of this community involvement!
However I don’t really like this unfair A-vs-B system for voting. If you’re going to stick to this format, can we at least get a “neither” and a “both” button?
For “most Amazingly Engineered project” I’m sorry but most of the ones there simply don’t qualify. Some aren’t even actual tangible projects yet. If you give me two choices, and neither match the description, I should be able to vote “neither” and somehow have that counted.
Similarly if I happen to see two amazing projects, then I’d like to give my vote to both! Use two of my votes, that’s fine, but they’re both deserving, and one shouldn’t lose out because the other is a bit better.
Or better yet, just go to a “hot or not” independent voting system. Every project is isolated, and you can vote “Yes”, “No”, or “Abstain”. Maybe “Abstain” will use your vote (so that you can’t just skip ahead and game the system) but if “No” is counted -1 and “Yes” +1, then sometimes I’d rather vote “0”.
Yes- I like the hot or not idea.
The hackaday voting system doesn’t measure popularity, it measures correlation.
Popularity voting (ie – vote for the best, or vote for the top 10) measures popularity, and can be unfairly skewed by people who have access to a lot of friends (hey – everyone in my hackerspace, please vote for me!), or who spend more money/effort on advertizing and polish, or who simply make a lot of user ID’s and vote for themselves. (Consider a 4Chan or reddit voting block gaming the votes.)
Hackaday voting avoids this problem. By making people consider all projects in pairs, it’s much harder for someone to game the system. It also makes people look at and consider projects which aren’t as much in the public eye: projects that have only recently been started, projects entered by people with poor command of English or presentation skills, and so on.
Hackaday voting measures correlation to a specific description, and can be positive, negative, or zero.
A fair coin flip will return no information, so flipping a coin to vote says effectively “these two projects don’t correspond to the description”. At the end of the voting round, those projects with no correlation will register around 50%.
Projects which are contrary to the description will have negative correlation, and register less than 50%. For example, “new and improved ebola virus” would have negative correlation to “most likely to save the world”.
Projects which match the criteria will have positive correlation, and show up as greater-than 50 percent on the voting. These are the projects which are most closely described by the criteria in the voting round, and that’s what Hackaday wants.
You only need to look at the list of prize projects sorted by skulls to see this in action. Skulls are effectively popular voting: consider whether the best hackaday prize projects are the ones with the most skulls.
Link to Hackaday prize entries, sorted by skulls:
https://hackaday.io/projects/tag/2015HackadayPrize
The Hackaday voting system has some big red flags, but not having a “skip” option isn’t one of them. You’re not wasting a vote, you’re contributing to the correlation.
Flip a coin, it means that neither project correlates with the criteria very well.
It would only measure correlation if every project were presented randomly against every other randomly presented project.
The rate that the projects show up is not random. There is a strong tendency for the project(s) that I did not vote for to show up again and again. I have even had projects show up again that I had already voted for.
And, with only 50 votes, you will never see all the possibilities. I looked through the list of projects that could be voted on, and saw quite a few that never showed up while I was voting that were “Amazingly Engineered” and I would have voted for.
Projects are divided up by IP or something apparently, so, the projects in your “block” won’t change.
Again, it’s not a popularity contest. It’s not about you specifically selecting your winner. It’s about the community as a whole registering preferences on projects and amalgamating *that* into a list of winners. Something it does far better than a popularity contest.
That, and others, would be the “big red flags”, as mentioned.
There are currently 588 projects under the tag “2015HackadayPrize”. That would be 588 x 587 = 345156 unique pairs of projects which can be presented for vote.
The image above states “22541” votes. Let’s assume that means *voters* and not “total votes”. That means 22541 x 50 = 1127050 total votes were cast.
If the projects were uniformly paired and voted we would expect, on average, about 3.2 votes per project. The list above implies that each entry was presented 83-ish times.
Does this make sense?
(Also – did a full third (22541) of Hackaday.io users (77000) actually vote?)
Projects list:
https://hackaday.io/submissions/prize2015/list
As they can’t possibly be doing a full cross-pair voting, I wonder if it’s better to think of it as a massive sorting system where previous votes are used to influence subsequent votes to improve the result of the ‘sort’
Bad UI design and major staff level bias in the way the staff hand picked project gets to get featured. When they show a topic, they don’t show every projects related to that and more often that the better one in that are ignored in the post. This has happened last year and again this year.
Take a look at the way the skull/view count stats on the first few pages and you’ll see that there is a huge gap that the “lesser” projects can never fill up as the first few pages of prime real estate are the one the average readers get to see first every time.
Now that SPAM bots start showing up, the current voting system can be abused too.
Press F5
22,541 votes cast? If that is the total number of votes cast with each(assuming) casting 50 votes each, does that imply there are only 450 people here actually voting? Did I see somewhere there are 77,000 io.users? Yikes…….
Yep, that’s what it means. But remember, not all the people that voted used all 50 votes, so it could be more like 600 or more people voted.
So it could be possible that even people who are in the contest itself are not voting? You would think given the prize for this thing(gift card) just about everyone would go up and click at least one vote off to be in the drawing. I am shocked and amazed. Is there that little interest in the HaD contest itself, or just this portion of it? Inquiring minds want to know.
I think the fact that a $1000 gift card has never been given away put a bad taste in the mouths of potential voters.
I still voted due to the potential for getting other free goodies.
But if the above mentioned 77,000 Hackaday.io user number is true (I don’t know where that number came from), and they are randomly selecting one of those for a chance at the $1000 gift card, and they will only give the gift card to a Hackaday.io user if they voted (I would guess 600 of them voted), then I think the odds would be something like 1 in 77,000. Or maybe it is 1 in 128? Probability was never my strong suit..
Maybe they should randomly draw from the people whom voted for the $1000 gift card next round?
yeah, it’s closer to 1 in 128, 1 in 128 & 1/3 by my numbers.
Which to me suggests that the editors need to stop whinging that people aren’t voting enough. I can’t be the only person who registered once, forgot that I registered, and then did it again. If I vote on both accounts, that’s cheating the system; so I only use one account; the other is dead. Even a 10% dead list, with no other problems, is a bad start to contests like “$1000 to a random person if they voted”.
Then you have the bot accounts; the people who pay someone in India to register 10 or 100 accounts to vote for them last year. Or the ones who think it’s a social network and just send friend/follow requests to everyone.
So, if you actually want to give away prizes, and not whing about “oh no not enough people voted to give away $1000” while the stat folks out here hang our heads in shame and laugh about it because we read “ahh, another chance to advertise a big prize and give away cheap stuff”, then only count people active in the last X months. People who have logged in in the last 11 months, or only people who have logged in since account creation and have done something in the last 15 months . . . anything at all is better than the (at best 1000 voters? Why not publish that number) 1 in 77 chance that you cry about every week.
I’m sorry. I don’t have anything to add about voting, but why do so many people think that hydroponics would be that high on the list to save Earth? Do any of you know what kind of scale farming is done on? Earth will not be saved when you convert your backyard garden to hydroponics. There families that farm 6,000 acres around here, and corporations that farm much more than that.
They don’t know any better, they live in cities and have never done any farming.
But it is more likely to save the earth than the NIR blood glucose meter or 3D printed hand prosthesis.