Idea - Lottery Preference for #1 Ranked School

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:When did strategy become akin to Satan worship? Ooooh, no, can't do that, it'll lead to gaming. Shudder. Oh, how terrible! Wake up people. The whole charter system is gaming. We're gaming people who don't have the ability, education, or access to research or apply to charter schools.

A variation on the system proposed by the OP that assigns a weight to one's ranking of 12 schools (in addition to the random lottery number) would be quite workable, and would provide better, fairer outcomes, earlier in the process.

The problem with strategic ranking is that overall, it leads to sub-optimal outcomes. As soon as people start ranking in less-than-true-preference order, and assignments are done based on that ranking, there is a net loss of utility.

This pp must be an economist or other. They are correct. People should read up on this system and the reasoning behind it before they get so critical. No system will give everyone what they want, as we already know when supply is less than demand.

Other. Minored in Ec at Harvard.

[Anonymous]

Do people realize that most matching algorithms actually take into account the participants' preferences? Even the most basic form (the stable marriage problem) is based on creating a solution that most accurately reflects each person's preferences.

A more modern take is the algorithm that matches medical students to residency programs. The Nobel prize winners (who are condescendingly mentioned by every poster who thinks that preferences have no place in the DC charter school system) actually won their prize for the medical student-residency research. Yes, that's right. The research leading up to the creation of an algorithm that includes applicants' preferences is actually what won the Nobel.

The DC charter school allocation algorithm (which marginalizes preferences)? Hasn't even won a participation prize. And while we know who designed the DC charter system, we don't know what the specific directions or the limiting factors were. This wasn't an academic exercise. They were working for the DC government. Did DC express an opinion as to how the system should run? Was there be a difference in cost and/or time to the DC government if families' preferences were factored in--as opposed to basing the whole thing on a random lottery number?

Again, referring back to the actual Nobel-prize winning residency algorithm, the directions simply give student-applicants a disclaimer that in order to make the system function best for them, any rankings should reflect their true preference. Those who nevertheless choose to rank strategically, so so at their own peril. The result is a much fairer system than the one we have.

[Anonymous]

I don't understand how there's not already "strategy" involved in the current system. If you rank according to your true preferences, and just lottery for 12 super popular charters, you're likely to get shut out. If you want to get in somewhere, you have to be realistic about your chances, and including safeties = strategy.

[Anonymous]

Anonymous wrote:Do people realize that most matching algorithms actually take into account the participants' preferences? Even the most basic form (the stable marriage problem) is based on creating a solution that most accurately reflects each person's preferences.

A more modern take is the algorithm that matches medical students to residency programs. The Nobel prize winners (who are condescendingly mentioned by every poster who thinks that preferences have no place in the DC charter school system) actually won their prize for the medical student-residency research. Yes, that's right. The research leading up to the creation of an algorithm that includes applicants' preferences is actually what won the Nobel.

The DC charter school allocation algorithm (which marginalizes preferences)? Hasn't even won a participation prize. And while we know who designed the DC charter system, we don't know what the specific directions or the limiting factors were. This wasn't an academic exercise. They were working for the DC government. Did DC express an opinion as to how the system should run? Was there be a difference in cost and/or time to the DC government if families' preferences were factored in--as opposed to basing the whole thing on a random lottery number?

Again, referring back to the actual Nobel-prize winning residency algorithm, the directions simply give student-applicants a disclaimer that in order to make the system function best for them, any rankings should reflect their true preference. Those who nevertheless choose to rank strategically, so so at their own peril. The result is a much fairer system than the one we have.

Huh? I thought the systems (the Nobel prize winning one and ours) functioned in substantially the exact same way.

[Anonymous]

Anonymous wrote:Do people realize that most matching algorithms actually take into account the participants' preferences? Even the most basic form (the stable marriage problem) is based on creating a solution that most accurately reflects each person's preferences.

A more modern take is the algorithm that matches medical students to residency programs. The Nobel prize winners (who are condescendingly mentioned by every poster who thinks that preferences have no place in the DC charter school system) actually won their prize for the medical student-residency research. Yes, that's right. The research leading up to the creation of an algorithm that includes applicants' preferences is actually what won the Nobel.

The DC charter school allocation algorithm (which marginalizes preferences)? Hasn't even won a participation prize. And while we know who designed the DC charter system, we don't know what the specific directions or the limiting factors were. This wasn't an academic exercise. They were working for the DC government. Did DC express an opinion as to how the system should run? Was there be a difference in cost and/or time to the DC government if families' preferences were factored in--as opposed to basing the whole thing on a random lottery number?

Again, referring back to the actual Nobel-prize winning residency algorithm, the directions simply give student-applicants a disclaimer that in order to make the system function best for them, any rankings should reflect their true preference. Those who nevertheless choose to rank strategically, so so at their own peril. The result is a much fairer system than the one we have.

The current system takes into account the participants' preferences, it just doesn't weight the strength of those preferences.

There are some subtle differences with the medical residency matching (MRM). In MRM, schools also express preferences for candidates. In a way this is somewhat analogous to lottery numbers, but with a key difference: the graduates have some idea how strong or weak they are as candidates. It's like people knowing their lottery numbers before they make their rankings.

A more important difference is that in MRM everyone starts from essentially the same place. Nobody already has a spot at an OK school, there are no sibling preferences, and you can't repeat it year after year trying to improve your outcome.

If you believe a fairer system is possible -- fairer for all participants, not just you -- please explain with some detail how it would work.

[Anonymous]

Anonymous wrote:Do people realize that most matching algorithms actually take into account the participants' preferences? Even the most basic form (the stable marriage problem) is based on creating a solution that most accurately reflects each person's preferences.

A more modern take is the algorithm that matches medical students to residency programs. The Nobel prize winners (who are condescendingly mentioned by every poster who thinks that preferences have no place in the DC charter school system) actually won their prize for the medical student-residency research. Yes, that's right. The research leading up to the creation of an algorithm that includes applicants' preferences is actually what won the Nobel.

The DC charter school allocation algorithm (which marginalizes preferences)? Hasn't even won a participation prize. And while we know who designed the DC charter system, we don't know what the specific directions or the limiting factors were. This wasn't an academic exercise. They were working for the DC government. Did DC express an opinion as to how the system should run? Was there be a difference in cost and/or time to the DC government if families' preferences were factored in--as opposed to basing the whole thing on a random lottery number?

Again, referring back to the actual Nobel-prize winning residency algorithm, the directions simply give student-applicants a disclaimer that in order to make the system function best for them, any rankings should reflect their true preference. Those who nevertheless choose to rank strategically, so so at their own peril. The result is a much fairer system than the one we have.

Can you explain how the residency matching is different from the DC algorithm? How would you factor in student preferences in lieu of a lottery number?

[Anonymous]

Anonymous wrote:

Anonymous wrote:Do people realize that most matching algorithms actually take into account the participants' preferences? Even the most basic form (the stable marriage problem) is based on creating a solution that most accurately reflects each person's preferences.

A more modern take is the algorithm that matches medical students to residency programs. The Nobel prize winners (who are condescendingly mentioned by every poster who thinks that preferences have no place in the DC charter school system) actually won their prize for the medical student-residency research. Yes, that's right. The research leading up to the creation of an algorithm that includes applicants' preferences is actually what won the Nobel.

The DC charter school allocation algorithm (which marginalizes preferences)? Hasn't even won a participation prize. And while we know who designed the DC charter system, we don't know what the specific directions or the limiting factors were. This wasn't an academic exercise. They were working for the DC government. Did DC express an opinion as to how the system should run? Was there be a difference in cost and/or time to the DC government if families' preferences were factored in--as opposed to basing the whole thing on a random lottery number?

Again, referring back to the actual Nobel-prize winning residency algorithm, the directions simply give student-applicants a disclaimer that in order to make the system function best for them, any rankings should reflect their true preference. Those who nevertheless choose to rank strategically, so so at their own peril. The result is a much fairer system than the one we have.

Huh? I thought the systems (the Nobel prize winning one and ours) functioned in substantially the exact same way.

They do. pP bad no idea what she is talking about.

[Anonymous]

Anonymous wrote:

Anonymous wrote:Do people realize that most matching algorithms actually take into account the participants' preferences? Even the most basic form (the stable marriage problem) is based on creating a solution that most accurately reflects each person's preferences.

A more modern take is the algorithm that matches medical students to residency programs. The Nobel prize winners (who are condescendingly mentioned by every poster who thinks that preferences have no place in the DC charter school system) actually won their prize for the medical student-residency research. Yes, that's right. The research leading up to the creation of an algorithm that includes applicants' preferences is actually what won the Nobel.

The DC charter school allocation algorithm (which marginalizes preferences)? Hasn't even won a participation prize. And while we know who designed the DC charter system, we don't know what the specific directions or the limiting factors were. This wasn't an academic exercise. They were working for the DC government. Did DC express an opinion as to how the system should run? Was there be a difference in cost and/or time to the DC government if families' preferences were factored in--as opposed to basing the whole thing on a random lottery number?

Again, referring back to the actual Nobel-prize winning residency algorithm, the directions simply give student-applicants a disclaimer that in order to make the system function best for them, any rankings should reflect their true preference. Those who nevertheless choose to rank strategically, so so at their own peril. The result is a much fairer system than the one we have.

Can you explain how the residency matching is different from the DC algorithm? How would you factor in student preferences in lieu of a lottery number?

The biggest fundamental difference is that there are no preferences on the schools side in the DC lottery, unlike in the medical school ranking system. This is the fundamental difference between the DC lottery system and most "stable marriage" systems- the ranking only goes one way here (students rank schools), whereas most stable marriage systems (like medical schools) involve dual rankings- students would rank schools and schools would rank students.

Look people, no condescension with all the Nobel Prize talk- it's just that these are people who have been studying this stuff for decades and had others checking their work and had these systems go through multiple iterations to work out all the bugs. Unless you have pretty deep knowledge of all the parameters and in-depth knowledge of the actual algorithm, you just aren't going to be able to improve it.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:Do people realize that most matching algorithms actually take into account the participants' preferences? Even the most basic form (the stable marriage problem) is based on creating a solution that most accurately reflects each person's preferences.

A more modern take is the algorithm that matches medical students to residency programs. The Nobel prize winners (who are condescendingly mentioned by every poster who thinks that preferences have no place in the DC charter school system) actually won their prize for the medical student-residency research. Yes, that's right. The research leading up to the creation of an algorithm that includes applicants' preferences is actually what won the Nobel.

The DC charter school allocation algorithm (which marginalizes preferences)? Hasn't even won a participation prize. And while we know who designed the DC charter system, we don't know what the specific directions or the limiting factors were. This wasn't an academic exercise. They were working for the DC government. Did DC express an opinion as to how the system should run? Was there be a difference in cost and/or time to the DC government if families' preferences were factored in--as opposed to basing the whole thing on a random lottery number?

Again, referring back to the actual Nobel-prize winning residency algorithm, the directions simply give student-applicants a disclaimer that in order to make the system function best for them, any rankings should reflect their true preference. Those who nevertheless choose to rank strategically, so so at their own peril. The result is a much fairer system than the one we have.

Can you explain how the residency matching is different from the DC algorithm? How would you factor in student preferences in lieu of a lottery number?

The biggest fundamental difference is that there are no preferences on the schools side in the DC lottery, unlike in the medical school ranking system. This is the fundamental difference between the DC lottery system and most "stable marriage" systems- the ranking only goes one way here (students rank schools), whereas most stable marriage systems (like medical schools) involve dual rankings- students would rank schools and schools would rank students.

Look people, no condescension with all the Nobel Prize talk- it's just that these are people who have been studying this stuff for decades and had others checking their work and had these systems go through multiple iterations to work out all the bugs. Unless you have pretty deep knowledge of all the parameters and in-depth knowledge of the actual algorithm, you just aren't going to be able to improve it.

Students are ranked on the schools' side - first by preference and then by lottery # within each preference. It's just that other than the application high schools, there is no ranking based on performance in school. But it functions the same way as the residency model.

[Anonymous]

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:Do people realize that most matching algorithms actually take into account the participants' preferences? Even the most basic form (the stable marriage problem) is based on creating a solution that most accurately reflects each person's preferences.

A more modern take is the algorithm that matches medical students to residency programs. The Nobel prize winners (who are condescendingly mentioned by every poster who thinks that preferences have no place in the DC charter school system) actually won their prize for the medical student-residency research. Yes, that's right. The research leading up to the creation of an algorithm that includes applicants' preferences is actually what won the Nobel.

The DC charter school allocation algorithm (which marginalizes preferences)? Hasn't even won a participation prize. And while we know who designed the DC charter system, we don't know what the specific directions or the limiting factors were. This wasn't an academic exercise. They were working for the DC government. Did DC express an opinion as to how the system should run? Was there be a difference in cost and/or time to the DC government if families' preferences were factored in--as opposed to basing the whole thing on a random lottery number?

Again, referring back to the actual Nobel-prize winning residency algorithm, the directions simply give student-applicants a disclaimer that in order to make the system function best for them, any rankings should reflect their true preference. Those who nevertheless choose to rank strategically, so so at their own peril. The result is a much fairer system than the one we have.

Can you explain how the residency matching is different from the DC algorithm? How would you factor in student preferences in lieu of a lottery number?

The biggest fundamental difference is that there are no preferences on the schools side in the DC lottery, unlike in the medical school ranking system. This is the fundamental difference between the DC lottery system and most "stable marriage" systems- the ranking only goes one way here (students rank schools), whereas most stable marriage systems (like medical schools) involve dual rankings- students would rank schools and schools would rank students.

Look people, no condescension with all the Nobel Prize talk- it's just that these are people who have been studying this stuff for decades and had others checking their work and had these systems go through multiple iterations to work out all the bugs. Unless you have pretty deep knowledge of all the parameters and in-depth knowledge of the actual algorithm, you just aren't going to be able to improve it.

Students are ranked on the schools' side - first by preference and then by lottery # within each preference. It's just that other than the application high schools, there is no ranking based on performance in school. But it functions the same way as the residency model.

Good point. Exactly why I don't design these models and leave it up to experts in the field.

[Anonymous]

Anyone know where you can see the actual algorithm? Not that I would understand it, but I'd love to see it.

[Anonymous]

Anonymous wrote:Anyone know where you can see the actual algorithm? Not that I would understand it, but I'd love to see it.

Oh wait, I see it:

function stableMatching {
Initialize all m ? M and w ? W to free
while ? free man m who still has a woman w to propose to {
w = first woman on m’s list to whom m has not yet proposed
if w is free
(m, w) become engaged
else some pair (m', w) already exists
if w prefers m to m'
m' becomes free
(m, w) become engaged
else
(m', w) remain engaged
}
}

[Anonymous]

Anonymous wrote:Anyone know where you can see the actual algorithm? Not that I would understand it, but I'd love to see it.

You can see the algorithm as a set of logic questions (the sort you would use to do computer programming) here:

https://en.wikipedia.org/wiki/Stable_marriage_problem

[Anonymous]

Anonymous wrote:Anyone know where you can see the actual algorithm? Not that I would understand it, but I'd love to see it.

There's a video on the MSDC site.

[Anonymous]

Anonymous wrote:

Students are ranked on the schools' side - first by preference and then by lottery # within each preference. It's just that other than the application high schools, there is no ranking based on performance in school. But it functions the same way as the residency model.

This is a really important point. The "stable marriage problem" requires that both parties rank the other side. What OP is proposing is that DC use something other than the existing preferences plus lottery number. OP hasn't precisely explained how that ranking would work, other than it would give extra weight to having ranked a school #1.