Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The "overlearning the game" problem

andrewoneverything.com

Post reply on HN
Print view View original

131–140 of 144 posts

Re: The "overlearning the game" problem

#131

Post by learc83 »

drewcrawford wrote: post #12

If you pay a man by the hour, he'll work a lot of hours. If you pay him by the brick, he'll lay a lot of bricks. These "games" are basically the equivalent of counting lines of code or checkins. We're measuring poor proxies instead of the things we're actually interested in. The solution isn't an arms race to build bigger and better proxies, the solution is to measure real things instead of artificial ones. Here's ju…

I'm in favor of ending the cap on representatives in the House. The House was never meant to be capped at 435.

Re: The "overlearning the game" problem

#133

Post by r00fus »

utoku wrote: post #57
wisty wrote: post #10

It's related to Goodhart's Law: Any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes. This is often the result of attempting to overoptimize a system. You can optimize a race car to a huge degree, because you know exactly what you want it to do. You can't optimize a schooling system, because you don't know exactly what you want it to do. A little noise is a go…

I think the problem might be a deeper problem. The writer of the article is asking generally about how a system's rules can be exploited to the point that the original intention of the system is no longer achievable. The sortition is also a system with rules, therefore it can be exploited. As soon as I saw the rock-paper-scissors, I remembered a competition some time ago where different bots played rock-paper-scissor…

> So even sortition might be exploitable.

But the idea is that it's not easy to exploit. If it is more than trivially difficult to leverage, then that is wasted time/effort that dilutes from the greater message.

ie, why waste effort on the "tie-breaker" message when you can just spend that effort avoiding it altogether and push yourself beyond the margin-of-error?

The key to sortition is that effort expended on gaming sortition is wasted and non-complementary effort compared to the original contest/competition... right now, getting that 0.1% by dirty tricks is equivalent to getting-out-the-vote of 0.1% more of the populace.

Re: The "overlearning the game" problem

#134

Post by Greedy_Fools »

tsycho wrote: post #90
wisty wrote: post #10

It's related to Goodhart's Law: Any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes. This is often the result of attempting to overoptimize a system. You can optimize a race car to a huge degree, because you know exactly what you want it to do. You can't optimize a schooling system, because you don't know exactly what you want it to do. A little noise is a go…

Very interesting comment, hadn't heard of Sortition before. However, I see a few practical problems with this. First, it makes it harder to fix any errors/cheating. You can ask for a reevaluation of exam papers or a recount of the ballot if you felt that something shady happened, but then what about the randomization phase. In the admissions example for instance, if after a re-eval, the composition of students in the…

Do you guys know that shooting theives stops them from stealing ?

Re: The "overlearning the game" problem

#135

Post by r00fus »

jacques_chester wrote: post #27
drewcrawford wrote: post #24

Earlier quoted context omitted.

> The US founders specifically rejected that system because it leads to majoritarianism. It would be wonderful if you could cite a few primary sources discussing a direct vs indirect legislative branch as leading to majoritarianism, whether a longitudinal study or quotation of the appropriate historical person. You can still have separation of powers, judicial oversight, bills of rights, local governments, federalism…

> You can still have separation of powers, judicial oversight, bills of rights, local governments, federalism, and all the checks we have today against majoritarianism within a direct legislative system. This doesn't change the fact that legislation would still take on a majoritarian tone. You'd need even more circumscription of the powers of the legislature for it to work at all; history suggests that it's at best a…

> This doesn't change the fact that legislation would still take on a majoritarian tone.

It's not clear that this is substantially worse that legislation written for/by lobbyists.

The current process is thoroughly corrupt.

Re: The "overlearning the game" problem

#136

Post by Greedy_Fools »

drewcrawford wrote: post #12

If you pay a man by the hour, he'll work a lot of hours. If you pay him by the brick, he'll lay a lot of bricks. These "games" are basically the equivalent of counting lines of code or checkins. We're measuring poor proxies instead of the things we're actually interested in. The solution isn't an arms race to build bigger and better proxies, the solution is to measure real things instead of artificial ones. Here's ju…

That would be an actual representative government. A good way,logical and efficient and there are others. None of which will ever happen because government cannot be representative, no one will give anything up and inequality cannot be represented. You know the saying, if voting changed anything they'd outlaw it. We don't need a better way to communicate individually with our governors, we need better minds holding power.

Re: The "overlearning the game" problem

#138

Post by Greedy_Fools »

redthrowaway wrote: post #79
drewcrawford wrote: post #12

If you pay a man by the hour, he'll work a lot of hours. If you pay him by the brick, he'll lay a lot of bricks. These "games" are basically the equivalent of counting lines of code or checkins. We're measuring poor proxies instead of the things we're actually interested in. The solution isn't an arms race to build bigger and better proxies, the solution is to measure real things instead of artificial ones. Here's ju…

As you allude to, the problem with direct democracy is that people are stupid. Not all people about all things, but most people about most things. The realm of human knowledge is simply too vast for anyone to have a broad command of, so we specialize. This means that, for any given issue, the vast majority of people are ignorant about its specifics. Thus we cannot expect the masses to make wise decisions about most i…

It's not about people knowing a lot of things, it about people with a lot of things knowing. The fundamentals are what's lacking, the basic lies. "... ALL MEN are created equal" , lie from the very foundation, and why, why ,why can't they grasp "do until others as you would have them do unto you", the very heart of what they claim as their religion and proof of their legitimate claim to power ? It's only about holding key truths, not about vast sums of knowledge. People can do anything together, all the answers exist. It's the lack of power holding key truths that skews everything to the lack of progress,the lack of agreement and ability to find solutions. In short, we can fix it very quickly, the answers are actually simple and abundant, the people are deep and strong. BUT THEY WON'T LET US.

Re: The "overlearning the game" problem

#139

Post by Helianthus »

philwelch wrote: post #114
Helianthus wrote: post #103

Earlier quoted context omitted.

I'm late to the table and you might not read this, but something conspicuously absent from this conversation is a discussion of the representative nature of the US government by _state_. The higher legislative branch gives more weight to states as individual entities for very precisely balanced reasons. In a very real sense it is not representatives that act, but states themselves. This has been incredibly important,…

In other words, the popular vote is disregarded because different 'types' of people shouldn't have too much power: the density of cities should not grant that population dominion over the agrarian folk. In practice, the system grants sparse populations dominion over the majority. There's no well-thought-out philosophical reason for this, by the way, this was just a compromise states like Virginia had to make to get s…

>In practice, the system grants sparse populations dominion over the majority.

Not so! Populism _failed_, but it wouldn't have even had a chance to represent the interests of the agrarian states if statehood were not more important than sheer weight of numbers.

>There's no well-thought-out philosophical reason for this, by the way, this was just a compromise states like Virginia had to make to get states like Rhode Island to agree to the Constitution.

How is that not a well-thought-out philosophical reason? Balancing the needs of states is what enabled the Union. You act as if the compromise was not based in real, valid needs on behalf of the states. Rhode Island needed incentive to join. Virginia needed incentive to join. A two-senator Senate and a two-representative plus population-based-count representative HoR is the result: an _elaborate_ philosophical compromise, balancing power on several axes.

You're looking at it the wrong way: cities have much, much more by way of population. The most populous states would _steamroll_ everyone else if weighted solely by headcount, assuming they acted in unison--which has historically been the rule and not the exception.

Re: The "overlearning the game" problem

#140

Post by philwelch »

Helianthus wrote: post #139
philwelch wrote: post #114

Earlier quoted context omitted.

In other words, the popular vote is disregarded because different 'types' of people shouldn't have too much power: the density of cities should not grant that population dominion over the agrarian folk. In practice, the system grants sparse populations dominion over the majority. There's no well-thought-out philosophical reason for this, by the way, this was just a compromise states like Virginia had to make to get s…

>In practice, the system grants sparse populations dominion over the majority. Not so! Populism _failed_, but it wouldn't have even had a chance to represent the interests of the agrarian states if statehood were not more important than sheer weight of numbers. >There's no well-thought-out philosophical reason for this, by the way, this was just a compromise states like Virginia had to make to get states like Rhode I…

The mistake you're making is treating states as primary. States--and, more generally, federalism--are a means to an end. The purpose of government is to serve the interests of people. Rhode Island is a legal fiction delimited by imaginary lines; you and I are real people. The people who live within the imaginary lines of Rhode Island are just as important as anyone, but those million people are no more or less important than any other arbitrarily selected collection of a million people. Why should it matter where they choose to live, or where the imaginary lines are drawn?

In the United States, states are largely a historical accident stemming from the way North America was colonized. No one actually decided to establish a federal democracy in North America and divide it into states, different people from entirely different countries just happened to colonize different parts of the continent without ever intending that New Amsterdam and Jamestown would ever fall under the same American government. The Constitution was a political compromise based on the fact that thirteen of those colonies broke away from the mother country and needed a federal government. The Founding Fathers weren't gods, they were the same breed of hypocritical, self-serving politicians you see in any era, a scarce few wiser than the others for sure, but the same ugly sausage-making all the same.

Post reply on HN