[This post has now been subsumed by the following: blog.beeminder.com/akrasia. Also, the service described below, then known as Kibotzer, is now a real startup called Beeminder, announced here: http://lesswrong.com/lw/7z1/antiakrasia_tool_like_stickkcom_for_data_nerds/]

If you are a pathological procrastinator you're pretty screwed when it comes to weight loss.  You have this monumental goal like "lose 20 pounds" but there's no "last minute" that you can put it off until.

I and my partner have thought a lot about akrasia (ie, failure to do what we think we should be doing) and have a tool that tries to apply some anti-akrasia principles.  It's called Kibotzer (for "kibitzing robot") and is currently in private beta.

This is not necessarily the best way to use Kibotzer but if you're a pathological procrastinator and want to just embrace that flaw, Kibotzer can help:  (It's more general than weight-loss but that makes for a nice example.)

1. Pick your goal weight and goal date.

2. Kibotzer creates your "Yellow Brick Road."

kibotzer example graph

3. Place a bet with us that you'll stay on your Road.

   (if you go off your Road for even a single day, you lose.)

4. Procrastinate like hell until you're about to lose the bet.

The change in focus from "weigh 20 pounds less next year" to "be on the yellow brick road tomorrow morning" makes all the difference.  If you're in the wrong "lane" of your Road today then it's crunch time.  You have to be on your road tomorrow morning.  Pull an all-nighter on the treadmill if that's what it takes.

In one sense that mentality's crazy.  Whatever you do in any single 24 hour period makes essentially no difference to your weight next year. But that's the kind of thinking that let you drift away from your ideal weight in the first place.  The whole secret of Kibotzer is to automatically break down your long-term goal into day-to-day guidance.  And then, critically, add a wager to force you to stick to it.

Kibotzer's tagline is "Bring Long-Term Consequences Near!"  (Note that this differs from Stickk.com which adds consequences but can't bring them quite so near!)

We're interested to get the opinions of folks on LessWrong and perhaps some of you would like to be guinea pigs...

I'll put the rest of the details in the form of an FAQ.  Basically, we want to make sure we never cheat anyone out of money so we have safeguards we've worked out based on previous bets.


1. "What if I have a random up-day because I'm retaining water or something?"

The Yellow Brick Road adjusts its width so you shouldn't ever lose because of a random up-day.  We want to set unbending rules where each day matters, because that's what's motivating (no "I'll catch up later" where you dig yourself in a hole) but you should never lose on a technicality.


2. "What if I forget to reply to the bot or get too busy?"

If you stop replying to the bot you automatically get your money back. We only want your money if we're providing something so valuable that you want to interact with it continually.


3. "My goal is a year away; will you just hold my money that whole time?"

Whenever we're holding on to your money we pay a fair interest rate on it.


4. "It seems a little unforgiving; everyone makes mistakes..."

The Yellow Brick Road itself allows for a nice margin of error but to further ensure that you don't lose because of one or two mistakes, there's a "three-strikes" policy:  You can drive off your Road twice and the road will then be reset from where you currently are, targeting the same goal weight and goal date.  Only on the third time do you actually lose the bet.


5. "Do I still win if I go off the road once but end up reaching my goal in time?"

The short answer is that you lose if you go off at *any* time (modulo the three-strikes policy).  *But* the brilliance of Kibotzer is that it *knows* about random fluctuation, water retention, and hormonal cycles: the road is wide enough that you will never lose on a technicality.  What that roughly means is that you have to mostly stay in the right lane of your yellow brick road and reserve the left lane as your safety buffer for random (or monthly) up-days.

Recall Kibotzer's goal: "bring long-term consequences near".  In other words, the fact that you lose the game if you go off *tomorrow* is by design.  It's very hard to, for example, forgo that piece of pie merely because it will make it harder to weigh 20 pounds less 10 months from now.  Please!  One piece of pie won't make the difference and there's plenty of time to catch up!  Each individual piece of pie is *totally worth it*.  Same with each workout you really don't feel like doing right now.  Which of course is how you and everyone else in the country end up 20 pounds away from their ideal.  With Kibotzer that whole dynamic changes: when you're in the wrong lane of your road that one piece of pie could very well make the difference *tomorrow morning* and you're acutely aware of it.  The consequences are immediate.  And of course even better is the flipside of that coin: if you are well into your right lane then it's very very nice to be able to enjoy your hard-earned safety buffer and eat that piece of pie guilt-free!


6. "The graphs and numbers and betting seem a little gimmicky; is there another way to do this?"

Fundamentally it has to involve making a genuine commitment. Like, yes, I'm perfectly capable of staying below X pounds and I can commit to doing that. And then "commit" has to actually be made meaningful. Risking a painful chunk of money is the simplest way to do that.

It's sad but it often doesn't mean much when we verbally commit to something. (Some people are worse about this than others.) So the bet is like an acknowledgment that there's two "me"s: the me right now who definitely wants this to happen and then future-me who is going to thumb their nose and thwart it. You just have to force future-me's hand. Forget the charade that it's the same me -- it isn't. Verbalize the commitment all you want, history proves that future-me has a damn good chance of thumbing their nose at you (after all, how did you end up well over your ideal weight in the first place?). So if current-me is really serious then prove it by making it impossible for future-me to renege. Or, not impossible, just make future-me not *want* to renege. That's the best you can do and all that's needed.


7. "How do I actually place a bet?"

Email me with how much money you might be willing to risk (or indicate this on kibotzer.com/register).  I'll reply with the odds (how much you can win).  If that's acceptable, there's a "donate" button at kibotzer.com/money where you can put up the money.  The rest is honor system for now.


8. "I don't understand betting lingo.  What are 'odds'?"  (Probably don't need this one for LessWrong folks but interested to hear your ideas on how to explain this sort of thing to layfolk!)

First, "even odds" means that if you win you double your money.  If I'm betting on something where I'll probably lose then I'll want better odds to compensate, meaning that I'll more than double my money on the off chance that I win.  Higher risk, higher reward.  From your perspective, it's lower risk so you'll have a lower reward if you win.

For example, if you choose to risk $1000 then we'll figure you're highly likely to win so we might offer you odds where you risk the $1000 but only win $50 if you stay on your road (we'll also factor in how steep your Road is).  If you're sufficiently sure you can stay on your Road then that's a free $50 for you. (Of course the real advantage is the motivation it provides and the fact that you end up at the end of your Yellow Brick Road!)

New Comment
31 comments, sorted by Click to highlight new comments since: Today at 9:07 PM

Perhaps oddly, being threatened with a large negative consequence for inaction makes me even more paralyzed. I begin to act as though the negative consequence is inevitable, and just sit and sulk instead of taking action to avert the negative consequence. Instead of procrastinating and then making a heroic effort, I usually procrastinate until I decide that it's already too late, and give up completely.

"Carrot and stick" motivation simply doesn't work on me most of the time: I find myself not caring if I don't get the carrot and simply endure the the stick, suffering in silence.

Perhaps oddly, being threatened with a large negative consequence for inaction makes me even more paralyzed.

That's not odd at all; it's what I consider the normal definition of "pathological procrastinator".

People who find this sort of forfeiture arrangement motivating are probably not pathological procrastinators, the way poor people in wealthy nations aren't all that poor compared to relatively-well off people in poor countries. Doesn't mean they don't suffer from their problems, just that they're really not in the same class of hurt.

Your claim that poor people in rich countries suffer less from poverty is fallacious and insensitive. Statistical information shows that standard of living (and about every other imaginable method of judging well-being) is tied to relative wealth rather than absolute wealth.

How do you know that people won't lie about their weight in order to win bets? How do we know that you'll pay people who win their bets?

Great questions. It's just honor system for now. That's not necessarily crazy though. I mean, why do people on ebay actually send the goods after they get paid?

People on eBay who don't send the goods get negative feedback, which influences their ability to make future deals. Even that doesn't always stop thieves; a family friend once bought an iPod on eBay and was sent a padded envelope full of crumpled newspaper. (Then he went on court TV to get his money back from the scammer, but that is another story.)

Good point, though ebay still succeeds in the sense of providing a valuable service despite the fraud. Kibotzer should probably always remain community oriented -- perhaps an existing user has to recruit you and vouch for you (and for Kibotzer) before the bet starts.

Right now I'm just focused on the value of a service like this, assuming away the trust/fraud issues. I'm confident those are ultimately solvable. People are mostly pretty decent, in my experience.

By the way, thanks everyone for all the feedback! The comments would also be a great place for pointers to other anti-akrasia tools. I mentioned stickk.com already. Are there others? Anti-akrasia can be interpreted more broadly than "create new carrots and sticks".

I would upvote this because it's important that you answered the question and I don't want to discourage that, but I don't want to imply that I like your honor system solution.

Wouldn't it work just as well if instead of money, you used "reputation" in the form of a public profile?

I wouldn't be surprised if group pressure was more motivating than money, and it also lowers the barriers of entry so you get more people in. Plus it avoids all the legal and fraud issues.

Yes, to a degree, and that's why we have that (opt-in) public gallery on kibotzer.com. But if you (probably rightly) suspect that your friends have stopped paying attention to your graph, then it's easy to fall off the wagon. Speaking from experience. :) The wager idea also has an aspect that's hard to capture with shame (powerful as that is), namely the anti-procrastination aspect. You can dig yourself in quite a hole thinking you'll catch up soon and undo the (only gradually building) shame.

But if you (probably rightly) suspect that your friends have stopped paying attention to your graph, then it's easy to fall off the wagon.

Have a sidebar with "today's LOSERS"

Instead of money, have users give a list of email contacts. If the user fails, all of the contacts get spammed with a failure message written by the user when he starts (the contacts don't get spammed when the user starts or succeeds, so he can keep the list big).

Surely this is illegal in the US. First you're acting like a bank by taking deposits and paying interest, and probably you have to actually be a bank to do that; second you're setting yourself up as a bookmaker.

This may be mostly an issue of terminology. Treat the upfront payment as a fee for the service and then there's a contest with prizes. It's not gambling for the same reason it's not gambling when you pay an entry fee for a race with cash prizes.

Laws and regulations do not care what terminology you use, and many countries have extensive regulations on gambling and gambling web sites. If a technicality caused your site to be misclassified as gambling, then that would be the end of your site. And if you don't have a limited liability corporate shield around the site, it would also be the end of your personal finances. Unless you're really certain that you know what you're doing, you should consult a lawyer before proceeding.

The magic word here is "escrow". This is an escrow service, not betting or a bank. Alternatively, it might be considered a bonding service. Either way, though, it definitely needs competent legal advice and appropriate licensing for the jurisdictions you operate it in.

Seems to me that, after so much worrying over whether we can have any real-world impact, this should be upvoted more.

This is a great tool. One of my goals is to simplify my life by using checklists, schedules, software and other tools to avoid quotidian decision making.

The problem is that many important goals are not easily quantifiable; the partial solution is to reformulate goals in such a way as to admit quantification.

For example, if you want a quantifiable way to improve your math skills, download a formal proof checker and prove the 100 theorems listed here (or find an easier set). Another interesting possibility would be to create a wiki of canonical hard programming problems (e.g write a Lisp interpreter) along with test suites to determine if an implementation is correct. Then you can quantify your progress on the route to hacker nirvana by counting the number of hard problems you've solved. When you've solved all the problems in one language, start over with a new one.

Project Euler is a start on your last request.

I'm improving my maths by running through the exercises here

I like this idea. I would guess a big part of what is making it work for your users is simply the data entry.

This reminds me of something else that I thought is pretty good for combating exercise procrastination also. A little over a month ago I started the "100-day burpee challenge", where you do one burpee (it's an exercise; examples are on Youtube) on the first day, two on the second, until day 100. The important features of this for procrastinators are that it is very cheap to start—you just have to do one burpee—and that at any particular point it is not much more work than you've done before, and at the end you are doing a fair amount of exercise daily.

This will only work if you are the kind of person who can be motivated by the psychological cost of having failed, but still tend to procrastinate. I think that applies to a lot of people though; when I procrastinate, it is usually in part because of risk of failure.

I like this.

I wonder if it would work for people to have the software on their own computer, just as a way for them to keep score? Bringing the far near, but without the machinery of betting.

*cough* I came across a program some years ago for, one might say, a related purpose, although entirely different in its workings. A "virtual master" for people into BDSM. I will not post the link here, but if anyone seriously would like the URL and understands the sort of material they are likely to read there, let me know and I'll send it privately.

Why is it that some of the graphs on the Kibotzer home page have 30 or more points off of the yellow brick road? In fact, some of the graphs don't even have a yellow brick road that I can see. Are you still working out the kinks with the software?

Here are some examples of what I'm talking about http://kibotzer.com/jud-resistance/ http://kibotzer.com/dskate/ http://kibotzer.com/jill/

In any case, I think this is a nice idea. I'm a long-time user of Stickk for my anti-akrasia needs, but I'll definitely give this a shot if everything is working properly.

Yeah, those people (one of them is me) failed to stay on their yellow brick roads but kept reporting anyway, trying to catch up. The current system doesn't prevent that even though it misses the whole anti-akrasia point as I outlined it in this post.

Thanks so much, everyone, for this awesome feedback, as well as the huge influx of kibotzer applications (with some really creative goals). Sorry we're taking so long to work through the wait-list!

If you are a pathological procrastinator you're pretty screwed when it comes to weight loss. You have this monumental goal like "lose 20 pounds" but there's no "last minute" that you can put it off until.

No you're not. There are perfectly valid solutions like drugs (especially ECA), and for harder cases surgery. Or just talk to some bodybuilders for they know that stuff - and the very fact that you're talking about "weight loss" as opposed to "fat loss" and consider body weight a good indicator of progress shows that you're not thinking about it right.

Interesting that you mention surgery.
I believe that much of the benefit of stomach surgery can be achieved without the actual surgery. The surgery costs on the order of $20,000 and is typically not covered by insurance -- people are actually coughing up those amounts. What it accomplishes is to make you feel physically ill when you eat too much. That's all. There's no other magic going on. It just makes you eat less.
Or, more specifically, makes you want to eat less, since otherwise you'll be nauseated. So, what if you put the $20,000 in escrow and had a yellow brick road to follow such that if you ever went off the road, even for just a day, the $20,000 would be gone? You would first starve yourself for a while to make sure you were safely below the road. Then you would be cautious for a while. If you ever got close to the edge of the road then you'd really be nervous. That piece of pie would not look so appetizing. It could possibly be enough to put you over the edge. It might be a $20k piece of pie. You would really not want to eat it.

Finally, to respond to your point about weight loss vs fat loss, we don't disagree, just that if you're well outside your ideal weight range then scale weight is a perfectly good proxy for what you really care about and has the advantage of being much simpler to measure.

I never done surgery, I just know all papers show it's by a huge margin the most effective way of long term weight loss without least rebouncing of all. It beats all drug and diet based solutions ever tested by a huge margin. So diet would be no substitute for it. But I don't know anything more than pubmed and Wikipedia now, so I will shut up about it.

What I do know is that scale weight is a horrible proxy during dieting, because there are so many trivial ways to make very good progress by losing water (in so many different ways), and by losing muscles with any kind of shitty diet with serious calorie restriction. Actual fat lose only happens very slowly, and it's even possible to lose a lot of fat while on calorie deficit and building up weight - muscles are mostly water, so their scale weight per calorie is very high compared to fat.

I would even risk the suggestion that it's one of the main reasons why dieting fails - people still use this unbelievably wrong proxy to track their progress. Unfortunately body fat % as returned by cheap resistance based meters is extremely unreliable too. I found navel circumference to be the best free method - fat gets reduced equally over entire body, and navel is where there's a lot of fat, and very little muscle or anything else there, so changes in navel circumference are very good indicator of progress. But that's just a personal benchmark.

all papers show it's [surgery] by a huge margin the most effective way of long term weight loss

I don't have any evidence yet, but I claim that we can engineer a better way.

Thanks for the points about the problems with scale weight and the sensible alternative of navel circumference. I do think, though, that scale weight is not a crazy proxy, with some precautions (such as gradual loss). After all, if you're far from your bodyfat goal (and you're not a bodybuilder) then you know with certainty that the scale will say a smaller number when you're done. What I tell people is to first "kibotz" down to an upper bound on their ideal weight and then switch to a better metric.

In any case, Kibotzer is largely metric-agnostic.

You will have significant amount of muscles that you might loss even if you don't have anything to do with bodybuilding. Most seriously restricted diets put you on insufficient proteins (if you scale everything by the same %, you have insufficient proteins, and vitamins etc.), plus there's no reserve storage mechanism like with carbs and fats so you cannot just average them weekly, so your muscles are the only reserve source your body can use. So you will lose a lot of muscle, and all associated water, what will result in much faster weight loss than just losing fat on responsible diet.

The main difference is that normal people will rebounce back to normal amount of muscle, regaining all the water too etc. (that's part but not all of the whole after-diet rebounce) Bodybuilders cannot really do that, as they have much more muscles than they're supposed to.

Anyway if kibotzer works for you, go for it. I'm just guessing it won't work for most people, for the same reasons mainstream dieting doesn't.

My "unless you're a bodybuilder" caveat was just to point out that if you have, say, over a third bodyfat then you surely want your scale weight to go down as a side effect of losing the bodyfat, unless you also want to add massive amounts of muscle.

But we do mostly agree about the physiology and you definitely should add muscle as part of "weight loss". Muscle pretty much burns calories just sitting there, so it's a huge help. I definitely recommend some weight training as part of any fitness plan. Losing weight by just eating less can often be a death spiral where you have to keep eating less and less as your metabolism adapts to the fewer calories. The human body is impressively adept at maintaining its weight. Another thing I recommend is having occasional flat spots on your "yellow brick road". If it's hard to keep from gaining, let alone losing, then you're doing it wrong.

I guess we're getting far afield from the anti-akrasia aspects though. Again, kibotzer is agnostic about what metric you use to measure your progress, though the simpler it is to measure the better. (And we've put thought into the case of scale weight as a metric, figuring out how to adjust for random fluctuations and how to avoid losing weight the wrong way, like you're talking about.)

So I think we agree, "seriously restricted diets" are bad news and muscle gain should be part of any "weight loss" plan. But for most people that still means a gradually decreasing scale weight so I stand by it as a useful metric.