Mechanism re-proposal: Preshold

A separate issue is the whole quadraticity of crowdmatching. It doesn’t necessarily sit well with me that the more popular projects get proportionately more funding per donor compared to the more under-the-radar projects.

Sort of like in this XKCD image of a small, unknown project really doing a lot of good.

1 Appreciation

There’s indeed some orienting toward popularity, but it’s not exponential, it’s quadratic. Preshold still has a bit of this. Popular projects hit their thresholds and less popular do not.

As to the XKCD example, we can’t rely on democracy and votes and coordination to deal with upstream needs. We need to have the downstream projects take care of their dependencies. I do think downstream projects should more boldly and publicly list their dependencies.

1 Appreciation

Welcome back @sandra.snan ! Your perspective and input is highly appreciated, here and in these ongoing conversations. Just as an FYI, the current steps we are making towards a launch-ready mechanism include the potential for multiple formulas running side-by-side for different projects, in a sort of lab, as we figure out which steps to roll-out for the next iteration. Your earlier responses have absolutely influenced some of our recent discussions and I am excited to continue exploring them as we move forward.

2 Appreciations

With the XKCD example I didn’t mean only dependency issues. Just that small can be important.

Preshold still has a bit of this. Popular projects hit their thresholds and less popular do not.

This is a problem with all funding other than UBI.
Preshold tries to mitigate it somewhat by making the amount of money a project receives less of a factor of how many patrons it has. Of course, the monthly re-settable thresholds factor then partially mitigate that mitigation back into some amount of popularity contest.

Crowdmatching is a form of leverage a.k.a. gearing. Leverage/gearing can be good in some applications and bad in other contexts.

It can reduce or increase precision; in the case of crowdmatching it is reducing the precision.

I have a radio where when I turn the tuning wheel, it’s geared so that the real tuning wheel turns much more slowly than the one I move by hand. The gearing gives me more precision to find the right radio station.

Crowdmatching is geared the other way around. When a project is struggling because there are few patrons, well, with crowdmatching it will really struggle.

As a child I remember an explanatory comic book why insurance exists. Because you don’t know which house is going to burn. We share the burden. Fire insurance is a way to gear towards each separate home making a smaller investment, an investment that’s hopefully not even needed.

  • X, Y and Z all pay $33. Y’s house burns down. The $99 helps rebuild it. Contrasted with the “un-geared”:
  • X, Y, and Z all pay $0. Y’s house burns down. Y spends $99 out of pocket to rebuild it.

The gearing in that example is meant as a democratization effort, a fairness effort, a mutual aid effort.

The gearing in crowdmatching is tuned in the other direction.

  • X, Y and Z are projects in a non-crowdmatched, non-preshold, simple donor-ware world. X has ten patrons, Y has twenty patrons, and Z has thirty patrons. Neither can really make rent and all need day jobs, but the cup of coffee a week that ten patrons can buy is appreciated even by X. That’s ungeared. Now, the backwards-geared:
  • X, Y, and Z are projects in a crowdmatched world. X has ten patrons, Y has twenty patrons, and Z has thirty patrons. Y’s situation is the same as in the other world. Z is really blossoming and can buy luxury items. X got basically nothing. Nickels and dimes.

I would ideally want it to be tweaked a bit, adjusted. To give X, Y and Z the same amount of coffee cups (or whatever, I don’t even drink coffee, I don’t know where this “I’ll buy you a drink” euphemism for donationware came from) per month. X and Z pooling their efforts as a sort of, uh, insurance-style cooperation system to average out their support to reach Y levels. Or, if not the same, then maybe Z can give some of its coffee cups to X, maybe not to even out completely but a little bit.

Preshold doesn’t do that. Preshold is unskewed. This is a flaw in preshold, but, the “overshoot refund” (a.k.a share-the-burden mechanism) at least mitigates that by making sure that patrons to popular projects get a refund (which they can, optionally, give to other projects, evening things out overall).

Crowdmatching skews in the other direction.

The basic value mismatch between traditionally “left” politics and traditionally “right” politics is question “What is fair?”. Left says “From each according to ability, to each according to need.” Right says “Quid pro quo.” Now, it’s hard to reconcile those two groups, but in today’s day and age there are many issues where that particular “What is fair?” question isn’t very relevant—climate issues, diseases, externalities, exploitation, other obvious bugs in market capitalism that should be in the interest of all humans [even if not all corporations] to address, before coming to blows about “the fairness question”.

Left says “From each according to ability, to each according to need.” Right says “Quid pro quo.”

Insurance is a way to set up a “from ability to need” system.
Traditional donation/coffeebegging/busking–ware keeps it “quid pro quo”, kind of.
Crowdmatching skews it beyond quid pro quo. The less need you have (because you have a lot of patrons), the more you get. It’s the anti-insurance in that regard.

That’s not really what I want.

1 Appreciation

I appreciate the depth of thinking you’re sharing here. It seems to me to be still coming from a different set of base assumptions.

If the premise is a fixed set of patrons and costs, and it’s just about distribution, I’m not supportive of crowdmatching at all. Crowdmatching is not designed for that premise. Applied to that premise, crowdmatching has a popularity-contest effect. Or like the pareto principle: if 80% of the value comes from 20% of the projects, crowdmatching works against the 80% of lower-value projects that are nonetheless super important for a healthy ecosystem. We don’t want that result.

The entire foundational premise of crowdmatching is that we need to motivate latent patrons, currently freeriding. That the point is to demonstrate to the cynics that people aren’t entirely selfish and would freeride unless forced otherwise, but instead it’s the game theory coordination problem etc. If we can change from X Y Z projects with these small number of patrons to orders of magnitude more funding in the system, that’s transformational. We’re aiming to grow the pie. If anything is zero-sum it’s the desire to move resources from proprietary to FLO. But that itself would mean the FLO side isn’t zero-sum, it’s growing.

If we can get the amount of resources to FLO that the whole ecosystem actually needs, crowdmatching would be the wrong tool for allocation.

Put another way, my political views are in the from-each-by-ability-to-each-by-need direction. Crowdmatching exists in a reality that I assess as getting much less than people are able (and even willing, with the right motivation) to contribute, and the vast majority of projects are underfunded.

All that said, if we can motivate the same number of patrons to make the same resources available, within that, I want our mechanism to allocate according to need as best as it can. The way I’ve thought of that is outside the funding model. I’ve emphasized requiring projects to express transparency in their funding and build a social norm of supporting projects by need. So, I want to encourage patrons to pledge to projects with the most need and most value. If patrons come to the platform and look at the state of projects, we should be orienting them to care about where their pledges are most needed. That doesn’t have to be part of the mechanism itself though.

Anyway, we can and should continue these discussions when we have the mechanisms ready for clear comparison. We can have charts and examples with as close to comparable scenarios as possible. I think that will help. Something is off in the current discussion when two of us who seem to share values and politics are somehow disagreeing. The story in my mind is that the disagreements are not based in differing politics or values or worldviews or goals, those all seem to line up to me.

1 Appreciation

Yes, that seems like a more concise, and better, way to say what I tried to say with the whole gearing/insurance analogy.

Each month, the amount of patrons and each patron’s available donatable money is a number. The next month that number might be bigger or smaller, but in this moment the number is what it is.

That number will be skewed by the Pop-Con Pareto effect.

You want to increase that number, or, rather, shift money that people are “investing” in Beanie Babies, Charizards, and tulips, into growing FLO works.

That’s a fine goal. That’s why “which setup is going to motivate people to participate in the funding” is important.

I think freeriding is a good thing.
It’s good that people can participate freely in FLO culture.

As Amelia Earhart said in 1935:

Candle dipping, weaving and crude methods of manufacturing necessities are things of the past for an increasing majority. Today, light, heat and power may be obtained by pushing buttons […]
One hears a lament that a mechanized world would not be a pleasant one in which to live. Quite the contrary should be true. And it can be true if the fine minds who have accomplished so much in the realms of applied science will unite with the same enthusiasm to control their creations against social misuse.
Obviously, research regarding technological unemployment is as vital today as further refinement or production of labor saving and comfort giving devices.

That’s also why I like the share-the-burden mechanism.
We pay for the rides because we want the road to be clear enough to ride on, but, payers can get “cheaper” rides if more people pitch in.

In the giving/​crowdfunding/​Kickstarter/​Patreon/​gofundme/​bug-bounties/​Twitch-subscription/​YouTube-superchat/​Itchio/​DriveThru/​Bandcamp world, I see two kinds of money.

“Outside money” (money that comes from outside of the internet), and “community money”. “Community money” is money earned inside the online community (artists, journalists, musicians, many other things) via things like Patreon.

Alice makes fan art and has a Patreon up (which earns her “community money”). She also has a dayjob digging ditches (which brings in “outside money”). She spends some of her combined money on rent and some of it to her own favorite creators’ Patreons, let’s say she gives to Bob’s and Carol’s. Alice dreams about living on her fan art alone and not having to do her day job anymore.

A problem with this setup is that “community money” is never actually generated. It all starts as outside money, from people doing outside work, and becomes “community money” as it enters the online ecosystem. (Of course, some of it becomes “outside money” again as it’s being spent on rent and food and drugs.) The community money only sloshes around and is trickled up in a big popularity contest pyramid scheme. The more popular creators and projects get the most money.

I also really love some of those bigger projects and popular creators, don’t get me wrong, I’m not slagging them. I’m just not eager to continue to building on this pyramid structure. Or even ascerbate it, which to me crowdmatching seems to do.

FLO goods&services are de jure freeridable. That’s a feature. You record a song once and it can be enjoyed by millions of times. A dish at a restaurant doesn’t have the same property.

What we should build is a system that really respect the burden of generating outside money, digging ditches and such, for the working class. That’s why I put such an emphasis on trying to create a system where they, if they decide they want a particular FLO project to exist, only should have to pay as little as humanly possible for that project to exist.

Now, most “outside money” that exist come from the gambling halls of Wall Street and similar corrupt and gross shenanigans. Market capitalism’s two biggest bugs are externalities and exploitation and heaven knows that a lot of value is concentrated by those two mechanisms. That’s not a good thing and not something I want to depend on. I dig trying to fix that issue radically (for example favoring worker coops), or to trying to alleviate it via things like taxation or via really motivating donation schemes. And what’s the most motivating donation scheme? The one where you have to pay as little as possible but enough for the project to exist.

Some mechanims and systems can influence some people’s behavior to some extent some amount of the time. The “rules of the game” of the original snowdrift game is that the earlier you get out and dig the wetter, tireder, and miserabler you are gonna get, and the game theory is that that might influence some people into being less likely to go out early and shovel.

That is exactly the phenomenon that the original treshold pledge system (including, but not limited to, preshold) is designed to alleviate.
You sign up, but you don’t have to do any actual shoveling until there is enough sign-ups to clear that road.

“Safer giving.”

So let’s try to apply that to a situation where “flexible funding” (or, as Erudition calls it, “half-assed funding”) for some reason would be desirable. (Now, I think I read somewhere there is a lot of data out there that shows that flexible funding campaigns are less successful compared to fixed threshold campaigns. But that kind of data-gathering isn’t my area of expertise exactly.)

You sign up to shovel one hour, but, if fewer people show up, you only have to shovel a few minutes. “Symbolically”…? Because you are making a significantly smaller dent in that road, Pareto style.

The purpose of share-the-burden is, of course, that you sign up to shovel one hour but if more people show up—enough people to really start working on that road!—you only have to shovel a few minutes.

You summoned me and, after catching up, I see that the opportunity to clarify the mechanisms in this level of detail is now, and that’s what I’m trying to do here.

All I’m saying is that if today someone could give me convincing info that crowdmatching would fail, we wouldn’t proceed.

I wouldn’t necessarily go that far. Testing things out can be valuable.

However, it’s better that I write about my thoughts and ideas and analyses now while I have them, rather than wait until I’ve forgotten them.

That can help you both with your “patching/tweaking crowdmatching” efforts (which I don’t support) and your “implementation of alternative (‘better’, in my mind) mechanisms” efforts. After all, a lot of the difficult implementiation work made would be similar. Handling payments is hard.

The reason I brought up the question of “what is fair” is not to put your values in that regard into question. I just want to establish the larger context here.

All throughout yesterday, and continuing here, you’re approach comes across “We have different stories in mind and if Sandra could only adjust her story in her head, we would agree”. Something I opened the door to by, on IRC, saying:

Talking to smichel17 and wolftune about it is interesting because there is this original seed of an idea, the eponymous “crowdmatching”, that I never got onboard with and you both are very committed to and it’s become a lens for you, something you want to apply to every other protocol.
I never bought into that idea in the first place.
So crowdmatching seems to be a for-its-own-sake motivated thing for you.

And now you are serving up my own medicine to me. That’s fair, but, it makes me realize that we’re not moving forward, which was my hope by opening up that avenue of reasoning.

As you try to associate crowdmatching into various narratives, such as “people who are currently not donating but could afford to donate would donate”, please don’t abbreviate the narratives by eliding other facts and consequences about crowdmatching.

On IRC, I wrote:

Isn’t it correct that the core pitch of crowdmatching is that N (N rather than X since N here is a non-static, variable number) people get together?
They all wanna donate $10, but, they pledge that the fewer they are, the more they will withhold.
As the group grows, the less of the $10 they will withhold.

You responded:

[T]hat pitch certainly doesn’t get a “that’s right” from me. It’s definitely not the story in my head.

Is the issue that you hope that crowdmatching will motivate more people to join, or something else?

3 Appreciations

I think this is the key point of distinction. Crowdmatching is supposed to reach both freeriders and people who are currently not involved in the community. The N you discuss is real, but there are other groups that never considered themselves as a part of that group, but, we think, would join if provided the opportunity differently.

Anyhow, still split in many other places, but love seeing this active discussion and having everyone debate/question each others perspectives and assumptions. Thanks!

2 Appreciations

agreed there too!

And there

That’s because it has no mutual assurance. If you pledge and nobody else does, you’re out 100% of your funds with nobody else helping.

There have never been any campaigns that work that way. Flexible is you sign up for an hour, and you still do your whole hour alone if nobody else joins you. Crowdmatching is the new mechanism where you do less than your full pledge and there is a mutual-assurance for you to do more.

Let’s make this clear: everyone agrees with some form of share-the-burden as soon as we’re in the state where projects are adequately funded. Our statement all along is that our hope is to get to where that is our problem. How to spread the burden is the problem we wish for. But right now, we have the problem of getting projects funded enough to sustain or reach potential in the first place.

That’s not the story in my mind. My story is: “if Sandra could fully get the perspective I have in mind, and I get the one she has in mind, then we would understand each other fully”. I don’t know what we would agree on and certainly wasn’t expecting that you fully understanding my position would mean that you would agree with it.

Getting to “that’s right” is about listening and understanding, all directions. It’s a potential step to agreement. Reductio ad absurdem: We could listen to some bigot and get to “that’s right” in our understanding of them, and that would have value in itself, but we still wouldn’t agree with their bigotry.

Yes, at its core, I do hope that crowdmatching will motivate more people to join. I also hope that it will be a sustainable motivation. I know that with all-or-nothing thresholds like Kickstarter, it motivates people to join but when some portion of projects fail (or worse, succeed in their campaign but it turns out the threshold was set wrongly and they don’t have enough after all), those people become cynical and discouraged. I’m not concerned only about what motivates an initial willingness to pledge. I’m concerned about what builds a sustainable long-term FLO ecosystem. And I see a large portion of projects not having the reality of an all-or-nothing state. I see roads that could be cleared to the point where there’s one lane you can slowly go over with risk of getting stuck all the way to roads that get cleared fully but with a lot of delay, the clearing is much slower than it could be.

Core tension: I don’t think thresholds reflect reality

This is why I want to get concrete with real projects and situations. The story in my mind is that most all-or-nothing threshold points are somewhat arbitrary guesses we use for practicality but 90% of the funding actually would make a difference almost no matter how you set the threshold. Setting it to 90%, 90% of that will still be useful. And getting zero until the threshold is met is something I imagine will both (A) discourage patrons and project over the long-run and (B) lose the actual progress that would be achieved with partial funding.

I believe there are projects where all-or-nothing thresholds exist, but I think we impose the idea on situations because it is easier to think about. I believe the vast majority of situations do not actually fit that model, estimates are generally widely off, and nearly every person and project and community could do more or less with more or less resources, and the thresholds are out of touch with reality and are a blunt tool we use just for conceptual simplicity. Thresholds are like a hammer, they make us see every situation as nails. And they will only accommodate the subset of projects where we can use that hammer effectively, and they force projects to do the extra work of conforming to that model even when it’s an ill-fit.

1 Appreciation

I can pretty easily be convinced that flexible giving without a threshold has advantages especially in situations where projects is far from being funded. The whole “You don’t need a penny just to hang around, but if you’ve got a nickel, won’t you lay your money down?” classic busking mentality.

I do believe that thresholds is the best-idea-yet to get cynics to contribute, but, I don’t see thresholds as as core as the following:

The whole “mutual-assurance for you to do more” thing. The quadratic curve shape. This doesn’t appeal to me. It feels “unfair” to me in ways that strike deep and intuitive. Like, when we were kids we hadn’t figured out the “one person divides, the other person splits” protocol so it was always a stingy measuring-tape-and-rules type situation. Learning about that pie-division protocol was like an immediate “Oh, that sounds fair!”

And the idea here that you commit to match each other by donating more and more is… uh, that’s not even the protocol as proposed. You don’t commit to matching each other’s donations; you commit to bumping your donation as a factor of head-count.

Even without a threshold it should’ve been the reverse:
You commit to donating $10 but you get to donate less if more people join. Let’s say two people donate $10 each, three people donate $9 each or whatever: the sum is still going up, so we get the benefits (and the drawbacks of “half-assed funding” that I and Erudition has pointed out) of thresholdless that you requested, but each person is “shoveling less” the more people join in. We get burden-sharing right away. I’m gonna call this the “many-hands” protocol. The scaling factor for each donation decrease with headcount N but the scaling factor is kept bigger than 1/N. It’s X/N where X is a number higher than 1. That way the sum will always go up even as each hand’s work is lightening.

2 Appreciations

smichel17, I disagreed with the multi-goal version of Preshold (the one that ended up being isomorphic to a stepped/quantized crowdmatching), but a multi-goal version of Preshold that ends up isomorphic to a stepped/quantized version of many-hands I could get behind.

This sounds to me like feeling a mild sense of disgust / revulsion / discomfort and then seeking intellectual justification for the feeling. What’s most insightful to me is that your intuition exists and is surely not unique to you.

That is one variant of the mechanisms which @mray has been emphasizing and proposing. It’s not the one we originally started with or the framing we have decided to use. The mechanism in which you give in proportion to the total funds compared to a goal, as @msiep originally suggested for our next iteration, that one is indeed matching each-other’s-donations. In that, two people could decide to bump one another higher by increasing their pledges.

The overall concept is that I do not want to start out paying all my taxes while the rest of my neighbors cheat on theirs. I want to express that I’m willing to pay my taxes as long as others are paying theirs… but I’m not willing to wait for some threshold to hit and live until then in a community that is totally broke and can’t pay for basic services at all. We who agree to already be paying our taxes are telling others that we will be willing to pay more fully the more we know that the rest of the community is paying theirs. And I’d still prefer progressive taxation of some sort where poorer people scraping by are allowed to freeride, but we need some sustainable equilibrium. We can’t have almost everyone freeriding or things break down.

How do I choose $10? I could do $20. What if very few others join? Hmm, I don’t really want to do $20 just on my own for very little results. Maybe I’ll test this system with $5. Oh, hey there’s another project that has a dozen supporters already… should I put in $5 there? Well, hmm. It’ll cost me $5, but the project will only get $3 of that (or whatever). Maybe by reducing the other patrons’ burden, they will be more likely to stick around long-term at least?

Sandra, how about presenting this as a formal mechanism? Even if nobody seriously advocates for it, I think this is a good thought experiment, a good comparison against which to test various hypotheses. It’s the sort of model I’d like to see empirical research done to see how people in fact pledge or not and how it plays out compared to others.

As a social signal, I see crowdmatching as a crowd saying “hey, we’re really happy to give more if it goes along with you helping, please join us and we’ll be even more enthusiastic, together we can change the world!” and your share-burden-from-beginning as the crowd saying, “hey, look we put our money where our mouth is, we’re doing this, but there’s more to do, and this is a real burden, would you please just come help out… many hands make light work”.

I could see a lot of interesting thoughts here. This simple proposal does have some form of mutual-assurance / interconnection. It’s not just unilateral giving. And it doesn’t have hard thresholds with all the things I complain about with those. What I want out of a good mechanism is flexibility, sustainability, motivation for growth, and explicit connection in how the crowd works together, building a sense of shared community responsibility where my participation is related to the agreement to work together and not just me choosing binary sides between self-sacrifice vs freeriding.

2 Appreciations

A strong sense but yeah.

“What the thinker thinks, the prover will prove” basically.

Right, since the goal is to create a platform that is appealing for patrons.

Expressing that is possible through pledging. The whole “I’ll give a few symbolic cents just to get the ball rolling and then increase that to my real donation later” is crowdmatching. Which is fine if it weren’t for the the quadratic curve going in the “wrong” direction.

I will.

That was my starting-point for many-hands. Pretending for a moment that I buy into the flexible-funding, any-nickel-helps mindset and, from there, being able to still express trepidation about crowdmatching’s curve-direction.

1 Appreciation

From IRC:

I feel worse giving $10 as part of $200 total than I do giving $20 as part of $15,000 — particularly if I understand that the $15,000 only exists because others like me made this mutual agreement. I’m not concerned just about my $10 or $20, I’m concerned about shouldering too much share of the burden.

With plain normal coffee donation (not that I advocate it, just that its hat-on-the-street basicness is such a baseline that any new proposal needs to beat), you might donate $15, which is a larger share of $200 than of $15000.

It’s 7.5% compared to 0.1%.

With crowdmatching, you instead donate $10 at $200 and $20 at $15000.

That’s 5% compared to 0.13%.

Your part of the burden increased under crowdmatching. Your share of the burden goes up (compared to the flat line baseline of plain normal coffee donating systems).

I want the results of the $15,000/mo well-funded project, and $20 for that seems very minor burden compared to $1

But the uselessness of the $1 donation to an impopular project is only exacerbated by crowdmatching, not alleviated by it (until and unless the project becomes more popular, in which case any witholding-based algorithm (such as crowdmatching) becomes moot anyway).

As I noted on IRC, many-hands is awful for the snowdrift scenario where a blocked road is useless to all, where people don’t wanna pay for vapor. Threshold systems are a good and tested solution to exactly the original snowdrift scenario.

Y’all should change name & image from that if that’s not what you want to apply to.

The baseline of plain boring donating is a flat line. You shovel your shovelfuls or you donate your $15. Crowdmatching has an ascending line, many-hands has a descending line. The descending line does make it hard to be an early-shoveler (which is exactly why I like threshold systems), but, it also means that small, under-thresholded project get a lot of help, which is exactly what you keep requesting whenever I bring up thresholds.

One might, but I usually don’t. I don’t give that sort of donation to anywhere near the number of projects I’d happily pledge to in crowdmatching. If we are only looking at the set of people who already give $15 donations readily to all sorts of projects, we don’t need crowdmatching.

This is the key point. We don’t need crowdmatching! We just don’t. Everyone could unilaterally donate and we could get the FLO reality we dream of. There’s no fundamental reason that can’t happen. People could just recognize that the Snowdrift dilemma exists, each independently decide not to play that game but to just do the “right” thing anyway. Our hypothesis is specifically that, in practice, we need crowdmatching (or similar, something motivating enough on a collective-action basis) in order to get much more popular participation that we just aren’t seeing with the status quo.

Our claim is that people mostly don’t do this and the snowdrifts mostly don’t get cleared and when they do it’s painfully slow and incomplete. That’s the baseline.

Your points about many-hands are correct in my view. That’s why I see it as an interesting counter proposal as a thought-experiment more than a snowdrift-dilemma solution. I’m still curious to think it through more.

1 Appreciation

A very common reply throughout this discussion, both three years ago and now, both with mray on IRC yesterday and here in this thread, is this Scotsman’s Goldilocks Goalposting of “the perfect crowdmatching project!” Not too supported, and simultaneously not too undeserving. It needs to have history, reputation, and no income.

In other words, there’s an expectation that the disincentivization to undershoot and overshoot (that Preshold’s mechanics attempt to do) instead be done manually&magically.

In such a magical worthiness filter, plain donation would make more sense than the “oh, you’ll only get a few cents for now and then if it does grow all the patrons will be on the hook for more”.

That was in comparison to being on the hook for a crowdmatched $1–$20 dollar pledge.

I get misconstrued sometimes as advocating for plain-boring “here is my paypal send money please” donation systems. No, we’re all aware that the Snowdrift dilemma is a huge factor (one of several) in causing hesitation there. I bring it up not to advocate for it but as a baseline to compare other proposals to.

Salt, you reacted with sceptical to the whole Snowdrift dilemma name thing. But I’m gonna double down on that:

This is what’s been one of the most frustrating and backwards parts of this thing.

The Snowdrift dilemma as classically presented is a perfect match for threshold pledge systems. The counter-argument on here whenever thresholsd pledge systems are brought up is that “but we are for projects that are unlike the Snowdrift / clear the road situation”. If so, the project could be well served by explicit clarification of what’s an appropriate project to fund, and how much funds it should receive. Applying scaling factors whether in the ascending (crowdmatchting) or descending (many-hands) directions is perhaps not needed or appropriate once you’ve figured out that weeding filter, or, the appropriate factor polarity (if any) would emerge with greater clarity after hashing out such a filter.

2 Appreciations

Let me try to clarify that reaction by drawing a scenario that I often dream of regarding crowdmatching a new project, rather than the “Goldilocks” partners we hope to launch with.

Let’s say a group of friends start a band. At first, they only play out of their garage, working full-time jobs to support the side hobby. When they discover, they tell all of their friends to back them there. A handful do, maybe 50-100, which leads to a meaningless contribution per month.

However, as they play local shows and gather a following, that number slowly climbs to maybe 1,000. It grows because people like the band, they like that their music is all released for fFree, and they like the idea of being part of a group that supports the dream of the band being able to play full-time. Now that the group hits this size, each patron is giving $1 per month, an amount that feels small, but is covering a lot of the band-specific expenses.

After a few more years, some patrons may have left, but many more joined, and now they have a decent following of 6,000 people. This is raising $36,000 per month, which allows most of the band members to work part-time or, for the more frugal ones, to devote all of their time to the music. Patrons are still only paying $6 per month, and they all feel great for having been part of the community that grew this band organically.

Now, to introduce a threshold, perhaps the band says that they need to reach 10,000 patrons to go on an international tour. One day the crowd gets there, and the band fulfills that promise. Since they reached the goal, new patrons are able to join at $10 per month, but it is just to join the community, not to enable further crowdmatching.

Long-term patrons are given a choice, depending on the direction we go with the mechanism. One proposal is to have a checkbox saying they’d like to continue matching as the crowd continues to grow. Another would have the project set the next goal/threshold, and these patrons could opt-in to crowdmatching towards that.

Going the other direction, once at the threshold a “share the burden” function could trigger. At this point, each additional patron (at that level, if we go with multiple thresholds of a certain variety) would mean that every patron gives less and less. This both makes the level of donations more sustainable and eases the ask per patron.

Of course these are only a few of the proposals out there, but this hopefully gives a decent picture. As always, thank you so much for engaging so deeply and passionately here!

And I will admit, after writing that all out, I am even more convinced that we should pursue the organizational split that has been discussed over the last month or two. We are currently trying to build and launch However, this discussion, and a number of others, are around resolving the snowdrift dilemma. And, as you surmise, crowdmatching is likely NOT the best mechanism for overcoming the initial snowdrift, but it can function to keep the roads clear in the long-term.

1 Appreciation

the perfect crowdmatching project

I wonder if some of the difference in perspective comes from our initial focus on FLO software.

I think many FLO software projects are such “Goldilocks” projects. In the Snowdrift metaphor, they’re situations where the software maintainers are already shoveling a narrow, uneven path, enough for some motivated folks to cross but that’s about it.

I can think of many examples of widely used/appreciated FLO software, but many fewer examples of FLO art or music which is similarly popular.

Normally this would be an issue, but with FLO projects I see it as a feature. Ultimately, we don’t want multiple projects that all do the exact same thing. There’s no need for 3 or 4 different photoshop like programs, it’s better to have one really good one. Also, the great thing about FLO and Open-Source, is that if someone wants to fork the project to make a better version, and then add the features they want, if it’s enough of an improvement that new project will become the new ‘standard’. For example: What happened with ublock, ublock Origin, and NanoBlock.

The thing about FLO projects is that working class people, if they are that hard come by, shouldn’t be paying ANY money into the project, much like Wikipedia. The people who are supporting these projects should ideally be middle class individuals who have disposable income. Which, at this point, should be trying to get the maximum utility for each dollar donated, economically speaking.
That’s why I’m so interested in, and FLO projects in general. By donating to a well run long lasting FLO project, each dollar donated has almost unlimited utility into the future as an infinite :infinity: number of people can benefit from the FLO program, whatever it is. But the best thing I love :heart_eyes: about FLO and open-source is that it is entirely consensual, unlike taxes, and socialism in general. The problem with “from each according to ability” is that the word ‘from’ is describing a non-consensual transaction.

To return to what I was saying about maximising the utility of my donation, crowd matching does this very well. Each dollar I pledge, has the chance of motivating even more people to pledge, which should lead to sustainable support of projects I enjoy. I mean, sure, I could donate $1000 a month or something to a single project, and that should be enough to fund a single programmer supporting a single project, but at that point the relationship is more like employer/employee, and if something happens to either one of us, the project fails. At that point, whatever the total of the donations happens to be is somewhat wasted, unless someone forks it and continues the work.

Finally, my biggest criticism of pre-shold is the uncertainty and instability it would create on the creator side of things. First, you’re asking them to budget their project, next they need to choose a threshold to set, next, they need to recruit/market enough people to get over the threshold, and finally, they’ll probably have to create a few fake sock-puppet accounts using friends and family if they are close enough to the threshold. Then they have to do that every month, or every 2 weeks, or whatever. Every single hour spent trying to achieve preshold is an hour that could be spent working on the project or recreationally.

If you’ve read my whole spiel, thank you. I think your support for preshold, and my support for snowdrift, come from our different ideologies around money and labour. Regardless, I’ve enjoyed reading your thoughts, its a different perspective and I always appreciate seeing things from a different view.

Also, just a friendly reminder that snowdrift is entirely FLO itself. There’s nothing stopping anyone, including yourself, from forking it and making preshold a thing.

2 Appreciations

Sure. I think a normal vanilla threshold system (a la Kickstarter / Patreon), or even your average normal “donate/coffee/put some money in my hat please” system, has some strong advantages compared to preshold. It’s less that I advocate for preshold and only preshold compared to those more basic vanilla systems. At every turn, preshold is optimized for the benefit and motivation of patrons (with the idea that that’s ultimately what’s going to bring in most money to projects). Preshold as a game-theoretical construct is also the best answer to the “Snowdrift Dilemma”, on paper. Now, the “Snowdrift Dilemma” isn’t the be-all, end-all of why people donate or not donate, which is why Preshold isn’t the be-all, end-all of protocols. Simplicity and straightforwardness, while preshold beats crowdmatching there, it doesn’t quite beat out the more vanilla and established crowdfunding methods.

Crowdmatching is instable and uncertain for everyone, both for donors and for projects.

If this ends up being correct then crowdmatching will win, and all my speculation of how bad crowdmatching is will end up having been wrong. Which I wouldn’t complain about since that would be a net win for FLO, which is our goal.

I disagree with this. I use all four of Krita, MyPaint, The GIMP (#changethename) and Inkscape every day, and sometimes Blender also. They are slowly picking up each other’s features but that road has a long way left to go.

3 Appreciations