Accuracy of the Match Algorithm

[hopefulmatchapplicant]

Has anyone heard of any system errors or mistakes in the matching process? Do they actually check every single result manually? Any insight would be appreciated.

Essentially a program who told me that I had an impressive interview yet did not rank me, contacted me asking if I matched. I can't help but feel that there is some type of error with my result.

---

Members don't see this ad.

 

[disgonbgud]

Did they tell you they didn't rank you? That's a strange thing for a site to even do IMO.

 

[precisiongraphic]

Did they tell you they didn't rank you? That's a strange thing for a site to even do IMO.

Yes, do you mean that they implied that they did rank you and thought that you ranked another place higher, resulting in you not ending up there?

 

[hopefulmatchapplicant]

So I know that they did not rank me as they have at least 1 position open in the Phase II list and the program that I matched to was ranked lower on my list. However, they told me they were very impressed by me (after interviews by multiple people) and someone from the hospital asked me if I ended up matching (why even ask if you did not like me?) I just find all of this very strange and am wondering if there could have been some kind of mistake in my match result

 

[precisiongraphic]

So I know that they did not rank me as they have at least 1 position open in the Phase II list and the program that I matched to was ranked lower on my list. However, they told me they were very impressed by me (after interviews by multiple people) and someone from the hospital asked me if I ended up matching (why even ask if you did not like me?) I just find all of this very strange and am wondering if there could have been some kind of mistake in my match result

No mistake. They didn't rank you, therefore you didn't match there.

 

[hopefulmatchapplicant]

Yes I realize that but why would this program that did not even like me enough to rank me, contact me and ask me how my match went? Seems rather bizarre

No mistake. They didn't rank you, therefore you didn't match there.

 

[precisiongraphic]

Yes I realize that but why would this program that did not even like me enough to rank me, contact me and ask me how my match went? Seems rather bizarre

Maybe they had second thoughts and would have given you a chance had you not matched. Did you match?

 

[hopefulmatchapplicant]

I did match although to a less desired program. So no one has experienced any so called mistakes in the ranking system?

 

[iPharm12]

I did match although to a less desired program. So no one has experienced any so called mistakes in the ranking system?

Same thing happened to me when I was applying for PGY1s. Matched to my number 2 and my number 1 program emailed me to see if I matched afterwards because they were in the scramble. I think they probably didn't like me as a first choice candidate but when they ended up in the scramble they might have been more comfortable going with a candidate they had experience with rather than trying to find a candidate within the week of the scramble. So maybe that program liked you but just not as much as they implied.

At our program now we send every candidate basically the same email saying we really enjoyed getting to know them, enjoyed their presentation, we wish them luck and hope to work with them next year. So I wouldn't really read too much into what you hear from programs.

Edit: It probably wasn't a mistake in the algorithm that you didn't match there in phase I.

 

[ethyl]

They're just putting out feelers to help them with Phase II. There's no error.

 

[hopefulmatchapplicant]

Anyone actually tried calling NMP before for an explanation? I can't help but think something got messed up. It wasn't just that program but other programs who literally told me that I would be placed high on their rank list and then no match from ANY of those programs? There was at least 4 programs like that

 

[mfdoom]

Anyone actually tried calling NMP before for an explanation? I can't help but think something got messed up. It wasn't just that program but other programs who literally told me that I would be placed high on their rank list and then no match from ANY of those programs? There was at least 4 programs like that

Don't ASHP match rules say neither the program nor the candidate should indicate whether or not they will rank each other, let alone how high? In any case, to echo the other replies, an error seems highly unlikely. An interview panel may have said they liked you the best during a certain portion, but the ranking process involves considering all interview sections and interviewees together. There may have been only tiny differences between you and the people ranked above you. Who knows. The point is, things can change a lot between your interview and the ranking deadline. I knew a handful of people confused with their match results, but no one called NMP for an explanation.

 

[gkrgus]

Anyone actually tried calling NMP before for an explanation? I can't help but think something got messed up. It wasn't just that program but other programs who literally told me that I would be placed high on their rank list and then no match from ANY of those programs? There was at least 4 programs like that

Yeah but there were other applicants who they ranked higher than you, which means you probably did okay across all the interviews and nobody ranked you as a top candidate. That can explain why you fell all the way down to your 7th choice. (If you are still unmatched by the time it gets to one of your lower choices
and you are ranked higher than one of those people on the hospitals matched list, then you kick out the bottom person. So you might have gotten kicked out by a higher ranked candidate)

You can't blame it on the system just because you did not get the optimal result. Just suck it up and move on.

 

[Pharmd2017tx]

Anyone actually tried calling NMP before for an explanation? I can't help but think something got messed up. It wasn't just that program but other programs who literally told me that I would be placed high on their rank list and then no match from ANY of those programs? There was at least 4 programs like that

I'm sorry, but you sound like a spoiled entitled child. Be grateful for the opportunity you were presented, even if it isn't your top choice (or 2 or 6). I would hazard a guess that the majority of the people that are interviewed are super amazingly awesome just like you-everyone can't match to their top choice! It'll be a great year if you go into it with a good attitude. Choose to be happy!!


Sent from my iPhone using SDN mobile

 

[Rockinacoustic]

Think of the match like the date that says they "had a great time" only to never hear from them again. Lots can happen once you walk out of that interview.

 

[Abby Atwood]

They probably messed up and narrowed down the list more than they should have. It is weird that they didn't rank you because it seems like they liked you well enough.

I can say that, unless they weren't thinking about how the match works, they didn't rank you. If they ranked you there would be no need to ask if you had matched. Any candidate they rank is going to wind up with them or with a more preferred program. If they had ranked you they would have assumed that you wound up at a more preferred program and they wouldn't have reached out.

Edit: unless they assumed that you didn't rank them I guess (still a loophole)

 

[Pharmibippy]

Anyone actually tried calling NMP before for an explanation? I can't help but think something got messed up. It wasn't just that program but other programs who literally told me that I would be placed high on their rank list and then no match from ANY of those programs? There was at least 4 programs like that

Of the applicants we interviewed, 6 - 8 of them were deemed as potentially great residents. 2 - 4 were seen as potentially good residents.

But here's the kicker - we only offer 2 spots. Entirely possible that you were good, just not good enough compared to others.

 

[hopefulmatchapplicant]

Thanks for the replies/support! I'm definitely not an entitled person- furthest from it. I'm a student who worked very hard through school and fell in love with my top choice, only to not match to it or any of my top choices. I am very happy to have matched at all, it was just disheartening as nearly all of my top choices went out of their way to tell me that I would be a shoe in, that they ranked me #1 in the preliminary round when scheduling interviews, etc. On top of that many of them had 3-4 resident positions. So it's definitely tough to understand why they would tell me all of that info (which they didn't even have to say) and yet I still did not manage to match.

According to match stats, more than half of matched applicants matched to their top choice so regardless of how competitive it is, not everyone ranks the same program #1 and the programs that I applied to were not "name brand" programs. That's the only reason why I questioned the possibility of an error but perhaps things changed when they interviewed more applicants, idk.

Anyway, I suppose I will just have to move on, try to make the most of my program, and try to gain the experiences that my program lacks

EDIT: was not a 4.0 GPA student (3.3) so perhaps it came down to GPA and other factors apart from the interview at the end

 

[spacecowgirl]

Grieve a little and move on. There's no sense trying to dissect what happened.

 

[Pharmres420]

Don't take it personal. I didn't match somewhere I've worked for two years and I have a well rounded cv with high gpa. I was pissed at first but time to move on

 

[KARM12]

Who called you from the program? Was it the program director or someone else? If it was someone else, they might have no knowledge of how the actual rank list went down and might not know where you were ranked.

 

[anonymouspharmd]

Don't ASHP match rules say neither the program nor the candidate should indicate whether or not they will rank each other, let alone how high?

Yes, I would also like to know how people know how different programs feel about them. Is it just that people have much better intuition than I am giving them credit for, or do people from programs reveal that information more freely than they should?

 

[Abby Atwood]

Yes, I would also like to know how people know how different programs feel about them. Is it just that people have much better intuition than I am giving them credit for, or do people from programs reveal that information more freely than they should?

I have no intuition at all. Some of the programs hinted pretty heavily. It didn't mean anything in terms of their ranking of me. For all I know, I didn't even appear on their rank list. The programs just wanted me to rank them higher.

No harm done

 

[hopefulmatchapplicant]

Who called you from the program? Was it the program director or someone else? If it was someone else, they might have no knowledge of how the actual rank list went down and might not know where you were ranked.

It was a member of the RAC. I suppose that is possible however the member seemed to know that overall I was not ranked although she claims that she did rank me high on her list. Not sure what that means but I guess the other preceptors preferred other candidates. My interview was also very early in process so they potentially could have forgotten how well I interviewed or just wanted other candidates over me

 

[hopefulmatchapplicant]

Yes, I would also like to know how people know how different programs feel about them. Is it just that people have much better intuition than I am giving them credit for, or do people from programs reveal that information more freely than they should?

My programs literally told me I would be ranked high, that they really liked me, that I had a lot to offer to the program. In hindsight that was all bs, but I ranked those programs high as well but did not match to any of them. I suppose they did rank me potentially but not as highly as they implied. I've come to discover that they chose applicants with little work experience, no pharmacy extracurriculars, and a decent GPA over myself, with a modest GPA, ample hospital work experience, extracurriculars, etc. and after implicating that I'd be matched so who knows

Also there were many strange things about my interview process like being asked exactly what other programs I applied to and I wasn't in a position to lie

 

[Pharmibippy]

It was a member of the RAC. I suppose that is possible however the member seemed to know that overall I was not ranked although she claims that she did rank me high on her list. Not sure what that means but I guess the other preceptors preferred other candidates. My interview was also very early in process so they potentially could have forgotten how well I interviewed or just wanted other candidates over me

This was probably only 1 opinion of the residency panel/committee. So they could've liked you, but perhaps you weren't viewed so favorably by others. Since you interviewed early, others may have left a better impression.

My programs literally told me I would be ranked high, that they really liked me, that I had a lot to offer to the program. In hindsight that was all bs, but I ranked those programs high as well but did not match to any of them. I suppose they did rank me potentially but not as highly as they implied. I've come to discover that they chose applicants with little work experience, no pharmacy extracurriculars, and a decent GPA over myself, with a modest GPA, ample hospital work experience, extracurriculars, etc. and after implicating that I'd be matched so who knows

Also there were many strange things about my interview process like being asked exactly what other programs I applied to and I wasn't in a position to lie

At least at our facility, it seems like candidates were ranked based on the interview. I don't necessarily agree with this but candidates here were viewed on the same playing field once they got an interview.

 

[Rockinacoustic]

My programs literally told me I would be ranked high, that they really liked me, that I had a lot to offer to the program. In hindsight that was all bs, but I ranked those programs high as well but did not match to any of them. I suppose they did rank me potentially but not as highly as they implied. I've come to discover that they chose applicants with little work experience, no pharmacy extracurriculars, and a decent GPA over myself, with a modest GPA, ample hospital work experience, extracurriculars, etc. and after implicating that I'd be matched so who knows

Also there were many strange things about my interview process like being asked exactly what other programs I applied to and I wasn't in a position to lie

Again, you're getting duped by your own emotions. For all you know these applicants with little work experience, extracurricular, and high GPA's were a way better fit personally and professionally when all interviews were said and done.

If you're used to an world where everyone's "word is their bond" then maybe this is a reality check. Move on, and make the best of this upcoming year.

 

[pharmdpharmd]

Maybe you were polarizing, some people loved you and some people couldn't stand you.. the folks who loved you told you how great you are, and those who didn't made sure the program didn't rank you.

 

[lord999]

They probably messed up and narrowed down the list more than they should have. It is weird that they didn't rank you because it seems like they liked you well enough.

I can say that, unless they weren't thinking about how the match works, they didn't rank you. If they ranked you there would be no need to ask if you had matched. Any candidate they rank is going to wind up with them or with a more preferred program. If they had ranked you they would have assumed that you wound up at a more preferred program and they wouldn't have reached out.

Edit: unless they assumed that you didn't rank them I guess (still a loophole)

https://www.uibk.ac.at/economics/bbl/lit_se/papierews06_07/roth_pearson(1999)_.pdf

Not necessarily. It is quite possible that depending on the match statistics, a candidate and a site can rank each other, but depending on "best fit", the "worse" site may match before the "better" site if the "worse" site ranks the applicant highly enough while the "better" ranks lower. Again, the rule that you should not match anywhere that you don't want to go applies here. You could have convinced the "worse" site that you are a highly desirable candidate. So ironic.

We've done the math for these cases, and it's a fun observation that everyone forgets that the potential resident and potential site could pick in a way that everyone loses (this is a voting paradox and specifically the Ellsberg paradox problem). I kind of wish every potential resident was informed of how that actually plays out in practice as there are clear ways to game the system in your favor if you are the one choosing the interviews.

If you bother to read the paper, you will also note that there is a Buridan's ass problem where there are perfect ties between resident and site preferences, and those are broken by numbers of residents, and then arbitrarily broken otherwise.

 

[Abby Atwood]

https://www.uibk.ac.at/economics/bbl/lit_se/papierews06_07/roth_pearson(1999)_.pdf

Not necessarily. It is quite possible that depending on the match statistics, a candidate and a site can rank each other, but depending on "best fit", the "worse" site may match before the "better" site if the "worse" site ranks the applicant highly enough while the "better" ranks lower. Again, the rule that you should not match anywhere that you don't want to go applies here. You could have convinced the "worse" site that you are a highly desirable candidate. So ironic.

We've done the math for these cases, and it's a fun observation that everyone forgets that the potential resident and potential site could pick in a way that everyone loses (this is a voting paradox and specifically the Ellsberg paradox problem). I kind of wish every potential resident was informed of how that actually plays out in practice as there are clear ways to game the system in your favor if you are the one choosing the interviews.

If you bother to read the paper, you will also note that there is a Buridan's ass problem where there are perfect ties between resident and site preferences, and those are broken by numbers of residents, and then arbitrarily broken otherwise.

Hmm. I'll try to find some time to read the paper. It's not really in my specialty. If the paper describes something different than an applicant proposing system in which each applicant matches to a program in order of preference provided that a program has not already matched with a more preferred candidate, then it might be out of date (or the ASHP website is full of lies )

 

[lord999]

Hmm. I'll try to find some time to read the paper. It's not really in my specialty. If the paper describes something different than an applicant proposing system in which each applicant matches to a program in order of preference provided that a program has not already matched with a more preferred candidate, then it might be out of date (or the ASHP website is full of lies )

That is the overall goal, but the implementation is something where candidates (and programs) do not really understand that it's a global, not a local fit. This is the trick about the linear programming.

(Explicit) Rule 1: Candidates: Do not rank a program that you wouldn't want to go to. Programs, do not rank candidates that you would not take.
Everyone knows the consequences when a party violates this.

(Implicit) Rule 2: ONLY Candidates (Selectivity bias rule): If you know that interviews are very few (or that the candidates they rank are few), ONLY interview if you absolutely know when you want to go there. If questionable, DO NOT interview.

Example:
Jane Doe ranks:
1: Harborview
2: VA Beacon Hill

Harborview Ranks (3 slots):
1: Not Jane Doe
2: Not Jane Doe
3: Jane Doe

VA Beacon Hill Ranks (3 slots):
1. Jane Doe
2 and below don't matter

Where does Jane Doe match? VA Beacon Hill as the rank preference is lower for Harborview.

Here is why I give that warning about interviewing.

Crap Example:
Jane Doe ranks:
1: Harborview
2: VA Beacon Hill

Harborview Ranks (3 slots):
1: Not Jane Doe
2: Jane Doe
3: Not Jane Doe
4: Not Jane Doe
5: Not Jane Doe

VA Beacon Hill Ranks (3 slots):
1. Not Jane Doe
2. Not Jane Doe
3. Jane Doe

Where does Jane Doe match? VA Beacon Hill again EVEN THOUGH the rank preference is lower for VA Beacon Hill and the candidate locally. This is why:

In the case of Harborview, even if Jane Doe is not picked up, Harborview has a deeper roster. VA Beacon Hill if Jane Doe is NOT taken, loses the spot completely.
The National Match Program considers this a fault, so even though both candidate and program rank lower, the match happens with that not best because it is a scenario where more slots are filled while the more permissive program "loses" a better candidate, they are filled with someone that they are supposedly ok with too.

This happens! It's a result of programs being very selective of who they end up ranking.

(Implicit) Rule 3: (ONLY Programs: Same example as above) When there are more great candidates than they are slots, programs should be more selective about ranking anyone. Paradoxically, for weak programs, they should consider rolling the dice and picking selectively as if they are more selective than a popular, promiscuous in terms of selection program, their likelihood is greater of matching. I also think that weak programs have more to lose from selecting weak candidates, while strong programs can weather those crisis residents.

By the way, a candidate being selective does not work in the same way as the program. The program's "advantage" depends on them having to declare the number of slots open and having to accept for those slots.

Now, that is the static analysis, there is a more complicated dynamic analysis, but because there is normally a geographically correlated market, this is not usually a problem for pharmacy residency. For medical residency, it's reasonable to interview geographically and rank somewhere in the 10s or so. Pharmacy is now competitive, but it still has nothing on medicine with respect to the outcomes. Do you think ASHP really knows how this algorithm works? Not really. Neither does ACGME for that matter. They left all the linear programming to the economists, which gives you the solution that it does. The Health Economists and the Informaticists collectively go when they are shown the actual algorithm implementation was:

"The *$(#! This *()%ing algorithm decided my fate and I didn't even know the real rules of the match."

 

[hopefulmatchapplicant]

That's insane. It's entirely possible that the site where I matched ranked me high. I ranked the site (low but still ranked) because I would have rather ended up there than have not matched. It's not a terrible site in the least, just does not have some of the specialities that I'm interested in which is why I ranked them low.

So you mean to say if you match a candidate to a less desirable program just so it doesn't go unmatched then what is the point of Phase II and scramble? Why do those other programs end up there and rather not be matched to candidates who ranked them but did not prefer them?

That is the overall goal, but the implementation is something where candidates (and programs) do not really understand that it's a global, not a local fit. This is the trick about the linear programming.

(Explicit) Rule 1: Candidates: Do not rank a program that you wouldn't want to go to. Programs, do not rank candidates that you would not take.
Everyone knows the consequences when a party violates this.

(Implicit) Rule 2: ONLY Candidates (Selectivity bias rule): If you know that interviews are very few (or that the candidates they rank are few), ONLY interview if you absolutely know when you want to go there. If questionable, DO NOT interview.

Example:
Jane Doe ranks:
1: Harborview
2: VA Beacon Hill

Harborview Ranks (3 slots):
1: Not Jane Doe
2: Not Jane Doe
3: Jane Doe

VA Beacon Hill Ranks (3 slots):
1. Jane Doe
2 and below don't matter

Where does Jane Doe match? VA Beacon Hill as the rank preference is lower for Harborview.

Here is why I give that warning about interviewing.

Crap Example:
Jane Doe ranks:
1: Harborview
2: VA Beacon Hill

Harborview Ranks (3 slots):
1: Not Jane Doe
2: Jane Doe
3: Not Jane Doe
4: Not Jane Doe
5: Not Jane Doe

VA Beacon Hill Ranks (3 slots):
1. Not Jane Doe
2. Not Jane Doe
3. Jane Doe

Where does Jane Doe match? VA Beacon Hill again EVEN THOUGH the rank preference is lower for VA Beacon Hill and the candidate locally. This is why:

In the case of Harborview, even if Jane Doe is not picked up, Harborview has a deeper roster. VA Beacon Hill if Jane Doe is NOT taken, loses the spot completely.
The National Match Program considers this a fault, so even though both candidate and program rank lower, the match happens with that not best because it is a scenario where more slots are filled while the more permissive program "loses" a better candidate, they are filled with someone that they are supposedly ok with too.

This happens! It's a result of programs being very selective of who they end up ranking.

(Implicit) Rule 3: (ONLY Programs: Same example as above) When there are more great candidates than they are slots, programs should be more selective about ranking anyone. Paradoxically, for weak programs, they should consider rolling the dice and picking selectively as if they are more selective than a popular, promiscuous in terms of selection program, their likelihood is greater of matching. I also think that weak programs have more to lose from selecting weak candidates, while strong programs can weather those crisis residents.

By the way, a candidate being selective does not work in the same way as the program. The program's "advantage" depends on them having to declare the number of slots open and having to accept for those slots.

Now, that is the static analysis, there is a more complicated dynamic analysis, but because there is normally a geographically correlated market, this is not usually a problem for pharmacy residency. For medical residency, it's reasonable to interview geographically and rank somewhere in the 10s or so. Pharmacy is now competitive, but it still has nothing on medicine with respect to the outcomes. Do you think ASHP really knows how this algorithm works? Not really. Neither does ACGME for that matter. They left all the linear programming to the economists, which gives you the solution that it does. The Health Economists and the Informaticists collectively go when they are shown the actual algorithm implementation was:

"The *$(#! This *()%ing algorithm decided my fate and I didn't even know the real rules of the match."

 

[hopefulmatchapplicant]

Also does this mean for next year I should not apply to as many sites? I feel like that is a lose lose situation

That's insane. It's entirely possible that the site where I matched ranked me high. I ranked the site (low but still ranked) because I would have rather ended up there than have not matched. It's not a terrible site in the least, just does not have some of the specialities that I'm interested in which is why I ranked them low.

So you mean to say if you match a candidate to a less desirable program just so it doesn't go unmatched then what is the point of Phase II and scramble? Why do those other programs end up there and rather not be matched to candidates who ranked them but did not prefer them?

 

[MaximusD]

Your first example is incorrect. I'd encourage you to re-examine the nrmp explanation of how the match algorithm works. If you rank a program #1 and the program ranks you, you receive a tentative match unless another applicant who ranked the program #1 is more preferred by the program. Period.

That is the overall goal, but the implementation is something where candidates (and programs) do not really understand that it's a global, not a local fit. This is the trick about the linear programming.

(Explicit) Rule 1: Candidates: Do not rank a program that you wouldn't want to go to. Programs, do not rank candidates that you would not take.
Everyone knows the consequences when a party violates this.

(Implicit) Rule 2: ONLY Candidates (Selectivity bias rule): If you know that interviews are very few (or that the candidates they rank are few), ONLY interview if you absolutely know when you want to go there. If questionable, DO NOT interview.

Example:
Jane Doe ranks:
1: Harborview
2: VA Beacon Hill

Harborview Ranks (3 slots):
1: Not Jane Doe
2: Not Jane Doe
3: Jane Doe

VA Beacon Hill Ranks (3 slots):
1. Jane Doe
2 and below don't matter

Where does Jane Doe match? VA Beacon Hill as the rank preference is lower for Harborview.

Here is why I give that warning about interviewing.

Crap Example:
Jane Doe ranks:
1: Harborview
2: VA Beacon Hill

Harborview Ranks (3 slots):
1: Not Jane Doe
2: Jane Doe
3: Not Jane Doe
4: Not Jane Doe
5: Not Jane Doe

VA Beacon Hill Ranks (3 slots):
1. Not Jane Doe
2. Not Jane Doe
3. Jane Doe

Where does Jane Doe match? VA Beacon Hill again EVEN THOUGH the rank preference is lower for VA Beacon Hill and the candidate locally. This is why:

In the case of Harborview, even if Jane Doe is not picked up, Harborview has a deeper roster. VA Beacon Hill if Jane Doe is NOT taken, loses the spot completely.
The National Match Program considers this a fault, so even though both candidate and program rank lower, the match happens with that not best because it is a scenario where more slots are filled while the more permissive program "loses" a better candidate, they are filled with someone that they are supposedly ok with too.

This happens! It's a result of programs being very selective of who they end up ranking.

(Implicit) Rule 3: (ONLY Programs: Same example as above) When there are more great candidates than they are slots, programs should be more selective about ranking anyone. Paradoxically, for weak programs, they should consider rolling the dice and picking selectively as if they are more selective than a popular, promiscuous in terms of selection program, their likelihood is greater of matching. I also think that weak programs have more to lose from selecting weak candidates, while strong programs can weather those crisis residents.

By the way, a candidate being selective does not work in the same way as the program. The program's "advantage" depends on them having to declare the number of slots open and having to accept for those slots.

Now, that is the static analysis, there is a more complicated dynamic analysis, but because there is normally a geographically correlated market, this is not usually a problem for pharmacy residency. For medical residency, it's reasonable to interview geographically and rank somewhere in the 10s or so. Pharmacy is now competitive, but it still has nothing on medicine with respect to the outcomes. Do you think ASHP really knows how this algorithm works? Not really. Neither does ACGME for that matter. They left all the linear programming to the economists, which gives you the solution that it does. The Health Economists and the Informaticists collectively go when they are shown the actual algorithm implementation was:

"The *$(#! This *()%ing algorithm decided my fate and I didn't even know the real rules of the match."

 

[MaximusD]

I've gotta say, I've re-examined and I'm wrong. What ****ery!

 

[lord999]

That's the difference between local versus global linear programming. The match advice is broadly trustworthy but for those odd scenarios where a global solution outweighs a local one. It's intentionally written to have the global effect that most residents are matched to the most sites.

The reason why this comes up at times is due to to having a really odd market for medical residents where I'm from. The arguably premier medical center in the state that has the international reputation is not in a major metropolitan (and quite frankly is a huge historical anomaly to be located in an otherwise rural area), and they have been victimized by Scenario 1 from having to balance between national and local candidates. This medical center has no problem matching their national candidates, but for the local candidate market, they always get shafted with worse candidates even though those candidates rank that center more highly due to the small area dynamics of the state. It does cause some issues when both that center and mine would have rather traded candidates at times. However, the rules of the match expressively forbid that sort of collusion for obvious reasons.

That said, it's been brought up and protested multiple times that it should be a first rank system should be the way it goes (the way you thought it worked), except that if you do it that way, less residents match overall causing more scrambles. Such is the fate of us having to balance national policy against honest intentions of both the candidates and the sites to have the best from each other.

By the way, if you're still Air Force, AAAS works on similar random to the Match, except there's a whole subsystem devoted to dealing with Tier 3 and Tier 4 placement. PROFIS in the Army works differently and is more like the first pass system and keeps out Tier 4 assignments as those are only human assigned.