Bidding: Symplicity Algorithm Issues

(On Campus Interviews, Summer Associate positions, Firm Reviews, Tips, ...)
Forum rules
Anonymous Posting

Anonymous posting is only appropriate when you are revealing sensitive employment related information about a firm, job, etc. You may anonymously respond on topic to these threads. Unacceptable uses include: harassing another user, joking around, testing the feature, or other things that are more appropriate in the lounge.

Failure to follow these rules will get you outed, warned, or banned.
morningjoe
Posts: 8
Joined: Fri Mar 04, 2011 9:40 pm

Bidding: Symplicity Algorithm Issues

Postby morningjoe » Tue Jun 26, 2012 9:14 am

Took this from another post, but placing it in the proper forum as there doesn't seem to be an answer yet on TLS.

Does anyone know if the symplicity algorithm operates based on the ranking number of the bids? Or is there some type of priority based on how many interviews the person has previously gotten in the lottery?

For example:
My friend and I both bid firm A at #1 and firm B at #2. Tons of people bid on both these firms with very high bids. My friend gets firm A from the lottery and I don't. Does the algorithm correct for this and give me a better shot than my friend at firm B, either by shifting all of my bids up when I don't get an interview, or by prioritizing everyone's bids based on how many interviews they already have?

This would have a big impact on whether or not to bid on very popular firms/firms with few interview slots 4th or 5th (because there's no risk if it fills up, as your bids will just shift up) or whether or not to strategically bid such a firm at the bottom of your bid list (and work on getting the interview outside OCI) if the shifting mechanism is not the case (so as not to waste a #4 bid on a firm you anticipate enough students to put in their top 3).

User avatar
Arbiter213
Posts: 2249
Joined: Mon Oct 11, 2010 10:16 pm

Re: Bidding: Symplicity Algorithm Issues

Postby Arbiter213 » Tue Jun 26, 2012 11:16 am

morningjoe wrote:Took this from another post, but placing it in the proper forum as there doesn't seem to be an answer yet on TLS.

Does anyone know if the symplicity algorithm operates based on the ranking number of the bids? Or is there some type of priority based on how many interviews the person has previously gotten in the lottery?

For example:
My friend and I both bid firm A at #1 and firm B at #2. Tons of people bid on both these firms with very high bids. My friend gets firm A from the lottery and I don't. Does the algorithm correct for this and give me a better shot than my friend at firm B, either by shifting all of my bids up when I don't get an interview, or by prioritizing everyone's bids based on how many interviews they already have?

This would have a big impact on whether or not to bid on very popular firms/firms with few interview slots 4th or 5th (because there's no risk if it fills up, as your bids will just shift up) or whether or not to strategically bid such a firm at the bottom of your bid list (and work on getting the interview outside OCI) if the shifting mechanism is not the case (so as not to waste a #4 bid on a firm you anticipate enough students to put in their top 3).


From my understanding, at least for Cornell, there is no adjustment of the lottery.

morningjoe
Posts: 8
Joined: Fri Mar 04, 2011 9:40 pm

Re: Bidding: Symplicity Algorithm Issues

Postby morningjoe » Tue Jun 26, 2012 4:58 pm

Is that credited through a discussion with your career office? There doesn't seem to be any information on the Internet answering this question. Has anyone else had this question?

User avatar
Arbiter213
Posts: 2249
Joined: Mon Oct 11, 2010 10:16 pm

Re: Bidding: Symplicity Algorithm Issues

Postby Arbiter213 » Tue Jun 26, 2012 5:13 pm

morningjoe wrote:Is that credited through a discussion with your career office? There doesn't seem to be any information on the Internet answering this question. Has anyone else had this question?


Yes, from the career office.

WearyCartographer
Posts: 42
Joined: Mon Jun 25, 2012 12:41 am

Re: Bidding: Symplicity Algorithm Issues

Postby WearyCartographer » Tue Jun 26, 2012 7:59 pm

I discussed it in passing with career services at Duke. She said it was pretty straight forward. The computer gives everyone their first choice in a random order, and then moves on to second choice in entirely different randomized order, etc. If all the slots for one firm fill up and you have it ranked lower, that firm won't drop of your list, you just won't get an interview for that round of randomized matching.

My career counselor also mentioned that its good to be weary of your own personal schedule too. You could get to a point where you have firm ranked that still has spots left, but all of those times are already taken by other firms on your schedule, so you will not get a firm that round of randomized matching either.

We have a straight lottery system where anyone can bid any firm and there are no preselects.

User avatar
Arbiter213
Posts: 2249
Joined: Mon Oct 11, 2010 10:16 pm

Re: Bidding: Symplicity Algorithm Issues

Postby Arbiter213 » Tue Jun 26, 2012 8:11 pm

WearyCartographer wrote:I discussed it in passing with career services at Duke. She said it was pretty straight forward. The computer gives everyone their first choice in a random order, and then moves on to second choice in entirely different randomized order, etc. If all the slots for one firm fill up and you have it ranked lower, that firm won't drop of your list, you just won't get an interview for that round of randomized matching.

My career counselor also mentioned that its good to be weary of your own personal schedule too. You could get to a point where you have firm ranked that still has spots left, but all of those times are already taken by other firms on your schedule, so you will not get a firm that round of randomized matching either.

We have a straight lottery system where anyone can bid any firm and there are no preselects.


This seems odd. The way Cornell runs it is the converse: Firm X is first with 20 slots. All people who put firm X at 1 fill it. Let's say 15. Then all people who put it at 2, let's say 3. Then all people who put it at 3, let's say 3. Uh oh, that'd put us to 21. So of those three, 2 will be randomly selected. That last person and everyone who put the firm after it miss the firm.

EDIT: Turns out I was completely wrong- here they randomly select someone, give them the current top of their list, randomly select again, etc. If the firm is full, you get your next choice. So fucking stupid.
Last edited by Arbiter213 on Thu Jun 28, 2012 2:24 pm, edited 2 times in total.

morningjoe
Posts: 8
Joined: Fri Mar 04, 2011 9:40 pm

Re: Bidding: Symplicity Algorithm Issues

Postby morningjoe » Wed Jun 27, 2012 7:27 am

Regardless, it looks like the situation at both Duke and Cornell is that symplicity in no way shifts your bids up (and instead skips you) if the firm you bid on at slot X is full by the time the algorithm gets to that point, and this issue gets to the heart of my question.

dixiecupdrinking
Posts: 3139
Joined: Sun Oct 26, 2008 2:39 pm

Re: Bidding: Symplicity Algorithm Issues

Postby dixiecupdrinking » Wed Jun 27, 2012 7:57 am

Arbiter213 wrote:
WearyCartographer wrote:I discussed it in passing with career services at Duke. She said it was pretty straight forward. The computer gives everyone their first choice in a random order, and then moves on to second choice in entirely different randomized order, etc. If all the slots for one firm fill up and you have it ranked lower, that firm won't drop of your list, you just won't get an interview for that round of randomized matching.

My career counselor also mentioned that its good to be weary of your own personal schedule too. You could get to a point where you have firm ranked that still has spots left, but all of those times are already taken by other firms on your schedule, so you will not get a firm that round of randomized matching either.

We have a straight lottery system where anyone can bid any firm and there are no preselects.


This seems odd. The way Cornell runs it is the converse: Firm X is first with 20 slots. All people who put firm X at 1 fill it. Let's say 15. Then all people who put it at 2, let's say 3. Then all people who put it at 3, let's say 3. Uh oh, that'd put us to 21. So of those three, 2 will be randomly selected. That last person and everyone who put the firm after it miss the firm.

My understanding is the same as yours (NYU). Basically, the matching is run by firm, not by applicant. For Firm X, they run all the people who bid it #1. If there are still slots, they run everyone who bid it #2. Etc, until it's full.

In any event, I also agree that there isn't some sort of "consolation prize" thing built in. You can bid a firm at 4, and just miss it, and then bid a firm at 5 and just miss it and one at 6 and just miss it, etc. Each firm is treated independently.

User avatar
Arbiter213
Posts: 2249
Joined: Mon Oct 11, 2010 10:16 pm

Re: Bidding: Symplicity Algorithm Issues

Postby Arbiter213 » Wed Jun 27, 2012 8:11 am

dixiecupdrinking wrote:
Arbiter213 wrote:
WearyCartographer wrote:I discussed it in passing with career services at Duke. She said it was pretty straight forward. The computer gives everyone their first choice in a random order, and then moves on to second choice in entirely different randomized order, etc. If all the slots for one firm fill up and you have it ranked lower, that firm won't drop of your list, you just won't get an interview for that round of randomized matching.

My career counselor also mentioned that its good to be weary of your own personal schedule too. You could get to a point where you have firm ranked that still has spots left, but all of those times are already taken by other firms on your schedule, so you will not get a firm that round of randomized matching either.

We have a straight lottery system where anyone can bid any firm and there are no preselects.


This seems odd. The way Cornell runs it is the converse: Firm X is first with 20 slots. All people who put firm X at 1 fill it. Let's say 15. Then all people who put it at 2, let's say 3. Then all people who put it at 3, let's say 3. Uh oh, that'd put us to 21. So of those three, 2 will be randomly selected. That last person and everyone who put the firm after it miss the firm.

My understanding is the same as yours (NYU). Basically, the matching is run by firm, not by applicant. For Firm X, they run all the people who bid it #1. If there are still slots, they run everyone who bid it #2. Etc, until it's full.

In any event, I also agree that there isn't some sort of "consolation prize" thing built in. You can bid a firm at 4, and just miss it, and then bid a firm at 5 and just miss it and one at 6 and just miss it, etc. Each firm is treated independently.


I will say that they re-run the lottery if someone misses every firm.

dixiecupdrinking
Posts: 3139
Joined: Sun Oct 26, 2008 2:39 pm

Re: Bidding: Symplicity Algorithm Issues

Postby dixiecupdrinking » Wed Jun 27, 2012 8:16 am

Arbiter213 wrote:
dixiecupdrinking wrote:
Arbiter213 wrote:
WearyCartographer wrote:I discussed it in passing with career services at Duke. She said it was pretty straight forward. The computer gives everyone their first choice in a random order, and then moves on to second choice in entirely different randomized order, etc. If all the slots for one firm fill up and you have it ranked lower, that firm won't drop of your list, you just won't get an interview for that round of randomized matching.

My career counselor also mentioned that its good to be weary of your own personal schedule too. You could get to a point where you have firm ranked that still has spots left, but all of those times are already taken by other firms on your schedule, so you will not get a firm that round of randomized matching either.

We have a straight lottery system where anyone can bid any firm and there are no preselects.


This seems odd. The way Cornell runs it is the converse: Firm X is first with 20 slots. All people who put firm X at 1 fill it. Let's say 15. Then all people who put it at 2, let's say 3. Then all people who put it at 3, let's say 3. Uh oh, that'd put us to 21. So of those three, 2 will be randomly selected. That last person and everyone who put the firm after it miss the firm.

My understanding is the same as yours (NYU). Basically, the matching is run by firm, not by applicant. For Firm X, they run all the people who bid it #1. If there are still slots, they run everyone who bid it #2. Etc, until it's full.

In any event, I also agree that there isn't some sort of "consolation prize" thing built in. You can bid a firm at 4, and just miss it, and then bid a firm at 5 and just miss it and one at 6 and just miss it, etc. Each firm is treated independently.


I will say that they re-run the lottery if someone misses every firm.

I mean, you would have to be really unlucky to have bid more than a few firms in the exact slot where some people got the interview and others didn't. You'd have to be even unluckier to lose out multiple times on the random lottery aspect of that situation. I'd guess most people have this happen maybe once or twice.

User avatar
Arbiter213
Posts: 2249
Joined: Mon Oct 11, 2010 10:16 pm

Re: Bidding: Symplicity Algorithm Issues

Postby Arbiter213 » Wed Jun 27, 2012 8:23 am

dixiecupdrinking wrote:
Arbiter213 wrote:I will say that they re-run the lottery if someone misses every firm.

I mean, you would have to be really unlucky to have bid more than a few firms in the exact slot where some people got the interview and others didn't. You'd have to be even unluckier to lose out multiple times on the random lottery aspect of that situation. I'd guess most people have this happen maybe once or twice.


Missing on EVERY firm would take incredible luck and skill in planning the worst order possible. But remember, if you miss the first 5, the rest becomes much more likely to miss- as you get further down the list, the more likely it is a firm is filled. Poor strategy can easily contribute to a poor showing of firms.

dixiecupdrinking
Posts: 3139
Joined: Sun Oct 26, 2008 2:39 pm

Re: Bidding: Symplicity Algorithm Issues

Postby dixiecupdrinking » Wed Jun 27, 2012 8:29 am

Arbiter213 wrote:
dixiecupdrinking wrote:
Arbiter213 wrote:I will say that they re-run the lottery if someone misses every firm.

I mean, you would have to be really unlucky to have bid more than a few firms in the exact slot where some people got the interview and others didn't. You'd have to be even unluckier to lose out multiple times on the random lottery aspect of that situation. I'd guess most people have this happen maybe once or twice.


Missing on EVERY firm would take incredible luck and skill in planning the worst order possible. But remember, if you miss the first 5, the rest becomes much more likely to miss- as you get further down the list, the more likely it is a firm is filled. Poor strategy can easily contribute to a poor showing of firms.

I suppose. I'm talking only about "missing" in the sense that you bid a firm at number x, and some people who bid them at x get it and some don't. If a firm is simply filled by the time your bid rolls around, then that's simple. Even so... the vast majority of people tend to get very similar numbers of interviews in the end.

morningjoe
Posts: 8
Joined: Fri Mar 04, 2011 9:40 pm

Re: Bidding: Symplicity Algorithm Issues

Postby morningjoe » Thu Jun 28, 2012 5:19 pm

Arbiter213 wrote:
WearyCartographer wrote:I discussed it in passing with career services at Duke. She said it was pretty straight forward. The computer gives everyone their first choice in a random order, and then moves on to second choice in entirely different randomized order, etc. If all the slots for one firm fill up and you have it ranked lower, that firm won't drop of your list, you just won't get an interview for that round of randomized matching.

My career counselor also mentioned that its good to be weary of your own personal schedule too. You could get to a point where you have firm ranked that still has spots left, but all of those times are already taken by other firms on your schedule, so you will not get a firm that round of randomized matching either.

We have a straight lottery system where anyone can bid any firm and there are no preselects.


This seems odd. The way Cornell runs it is the converse: Firm X is first with 20 slots. All people who put firm X at 1 fill it. Let's say 15. Then all people who put it at 2, let's say 3. Then all people who put it at 3, let's say 3. Uh oh, that'd put us to 21. So of those three, 2 will be randomly selected. That last person and everyone who put the firm after it miss the firm.

EDIT: Turns out I was completely wrong- here they randomly select someone, give them the current top of their list, randomly select again, etc. If the firm is full, you get your next choice. So fucking stupid.



Did you speak to the Dean or one of the other counselors? I'm also at Cornell and I've heard conflicting answers. DeRosa explained it to me your original way, and then the other counselors explained it to me your edited way. Either way, I'd like to get that straight before bids are due.

AllTheLawz
Posts: 369
Joined: Mon Jun 25, 2012 11:20 pm

Re: Bidding: Symplicity Algorithm Issues

Postby AllTheLawz » Thu Jun 28, 2012 5:49 pm

morningjoe wrote:
Arbiter213 wrote:
WearyCartographer wrote:I discussed it in passing with career services at Duke. She said it was pretty straight forward. The computer gives everyone their first choice in a random order, and then moves on to second choice in entirely different randomized order, etc. If all the slots for one firm fill up and you have it ranked lower, that firm won't drop of your list, you just won't get an interview for that round of randomized matching.

My career counselor also mentioned that its good to be weary of your own personal schedule too. You could get to a point where you have firm ranked that still has spots left, but all of those times are already taken by other firms on your schedule, so you will not get a firm that round of randomized matching either.

We have a straight lottery system where anyone can bid any firm and there are no preselects.


This seems odd. The way Cornell runs it is the converse: Firm X is first with 20 slots. All people who put firm X at 1 fill it. Let's say 15. Then all people who put it at 2, let's say 3. Then all people who put it at 3, let's say 3. Uh oh, that'd put us to 21. So of those three, 2 will be randomly selected. That last person and everyone who put the firm after it miss the firm.

EDIT: Turns out I was completely wrong- here they randomly select someone, give them the current top of their list, randomly select again, etc. If the firm is full, you get your next choice. So fucking stupid.



Did you speak to the Dean or one of the other counselors? I'm also at Cornell and I've heard conflicting answers. DeRosa explained it to me your original way, and then the other counselors explained it to me your edited way. Either way, I'd like to get that straight before bids are due.


I don't see why these two ways necessarily have to be inconsistent.. could be that they are just discussing it from different perspectives. On the firm side, the people who selected it number 1 still fill first, then the ones who chose it second, and so forth under either scenario. On the student side, the randomizer just changes the order in which you get your number one and then moves to your next top choice if the firm is filled. I guess there could be a difference in what it does if your number one is filled.. lf so, the second way (mentioned in the edit) sounds better because it gives you your next choice rather than you risking losing out on number 2 as well.

User avatar
Arbiter213
Posts: 2249
Joined: Mon Oct 11, 2010 10:16 pm

Re: Bidding: Symplicity Algorithm Issues

Postby Arbiter213 » Thu Jun 28, 2012 5:59 pm

AllTheLawz wrote:
morningjoe wrote:
Arbiter213 wrote:
WearyCartographer wrote:I discussed it in passing with career services at Duke. She said it was pretty straight forward. The computer gives everyone their first choice in a random order, and then moves on to second choice in entirely different randomized order, etc. If all the slots for one firm fill up and you have it ranked lower, that firm won't drop of your list, you just won't get an interview for that round of randomized matching.

My career counselor also mentioned that its good to be weary of your own personal schedule too. You could get to a point where you have firm ranked that still has spots left, but all of those times are already taken by other firms on your schedule, so you will not get a firm that round of randomized matching either.

We have a straight lottery system where anyone can bid any firm and there are no preselects.


This seems odd. The way Cornell runs it is the converse: Firm X is first with 20 slots. All people who put firm X at 1 fill it. Let's say 15. Then all people who put it at 2, let's say 3. Then all people who put it at 3, let's say 3. Uh oh, that'd put us to 21. So of those three, 2 will be randomly selected. That last person and everyone who put the firm after it miss the firm.

EDIT: Turns out I was completely wrong- here they randomly select someone, give them the current top of their list, randomly select again, etc. If the firm is full, you get your next choice. So fucking stupid.



Did you speak to the Dean or one of the other counselors? I'm also at Cornell and I've heard conflicting answers. DeRosa explained it to me your original way, and then the other counselors explained it to me your edited way. Either way, I'd like to get that straight before bids are due.


I don't see why these two ways necessarily have to be inconsistent.. could be that they are just discussing it from different perspectives. On the firm side, the people who selected it number 1 still fill first, then the ones who chose it second, and so forth under either scenario. On the student side, the randomizer just changes the order in which you get your number one and then moves to your next top choice if the firm is filled. I guess there could be a difference in what it does if your number one is filled.. lf so, the second way (mentioned in the edit) sounds better because it gives you your next choice rather than you risking losing out on number 2 as well.


There's a huge difference in the variance- my original way (as DeRossa explained it to me as well), there will always be a normal distribution of probabilities and possible variance to the extremes is minimized. The way I'm hearing now (secondhand through the other counselors) allows for extremely lucky and unlucky outcomes. There's absolutely 0 reason to go with the pure random lottery instead of the by firm random.

morningjoe
Posts: 8
Joined: Fri Mar 04, 2011 9:40 pm

Re: Bidding: Symplicity Algorithm Issues

Postby morningjoe » Thu Jun 28, 2012 6:42 pm

I've also been told that the way the other counselors explained it, a student is selected at random "out of the hat", then given their top remaining ranked firm, they are then immediately placed back into the hat with just as good a chance to be selected next. So theoretically, a student can be selected two or three (or more) times before another student is assigned their first firm. Which is ridiculous.




Return to “Legal Employment”

Who is online

The online users are hidden on this forum.