I made a dumb mistake in final-12. For some reason I did not consider the other path for B :) It is the only mistake. Otherwise I would have had 100% in the course.

And just with one mistake I am in top 10% and not in top 5 or top 1. I would take a positive lesson out of this, it takes A LOT to be among the best. One point out and BANG, see you next time.

Just wanted to share this lesson ;)

asked 22 Dec '11, 05:44

calculemus1988's gravatar image


edited 22 Dec '11, 05:45

Lyle Lovett has a song about that, with this refrain: "It's such a shame, 'cause you've been so good up to now."
(22 Dec '11, 06:05) EllenL EllenL's gravatar image
thats amazing - i feel for you. Im pretty sure that with so many people getting in the 90% range for this course - they should really switch over to the ML class model (allowing retries) - the educational objective wouldn't be compromised, just the silly ranking thing where you can get so severely punished for just one part of one question - when that same question in a homework would hardly hurt.
(23 Dec '11, 04:24) tom tom's gravatar image
Indeed, the ML-class model was much better IMO, no more losing points for a silly oversight. I'm used to the kind of grading where getting a 70% means pretty much being among the best, so making a few silly mistakes isn't really a rank killer because they get diluted within all the real errors or missing answers. Here we have pretty much everyone above 90%. Considering that a normal human makes about 1% error on trivial repetitive tasks, well, you get the picture...
(23 Dec '11, 04:31) patheticcock... patheticcockroach's gravatar image
Great post! It was late at night 3 AM in the morning and I was staying up to see the result! I was doubtful about one question and at last minute I changed my answer and went to sleep! and I ended up in 10% ! if I had gone to bed early that night things were different!!! but what really cheered me up right now is your post!!!! YES,damn it takes a lot to be among the best!
(23 Dec '11, 12:10) Sunny Sunny's gravatar image

Actually, the lesson for me would rather be it's not worth taking advanced tracks if the final ranking will be noise anyway, particularly since the coefficients would allow people who did 6 mistakes in HW (plus fully screwed up 2 more HWs) to be ranked higher than people who did only one error on the final. You learn just as much from basic tracks without the hassle of sticking to a schedule which isn't properly enforced anyway. Sticking to basic tracks also sounds like a good way to be able to take more classes.


answered 22 Dec '11, 07:10

patheticcockroach's gravatar image


edited 22 Dec '11, 07:13

Or you could do the Advanced track and just care a bit less about the final score. I think the extra hassle and scheduling was rather useful for learning.
(22 Dec '11, 09:40) jojo jojo's gravatar image
I guess it all depends on your personal schedule and ability to motivate yourself. There was some weeks where I had extra free time but couldn't work in advance because the materials wasn't released yet, but some other weeks where I had difficulties to squish the smallest second to work on the videos during the small time window between the time when videos were released and the time when the HW was (supposed to be) due. I'd definitely have been more comfortable browsing the whole course at my pace.
(22 Dec '11, 11:16) patheticcock... patheticcockroach's gravatar image
And now imagine what it was for us needing the subtitles to clearly follow the lessons. Sometimes we were simply squeezed in time and could not view and learn when it was the most appropriate for us. I must admit, sometimes it was really pain in the ass. You had to shuffle what can be done from all the courses you are following. I hope they put the subtitles in a top priority list to improve quality for non-domestic students and hearing impaired students. Don't quote me wrong, I am not whining about the course, this is really a problem and an area to improved. The Intro to Databases was about 100% reliable on this side, only once we had a problem. Machine Learning was almost reliable until the third quarter of the lessons, after that, we ran into problems with timely subtitles publishing. AI was at problem well before half completition of the course. And I realize everyone did their best, that's not my point.
(22 Dec '11, 18:37) AchilleTalon AchilleTalon's gravatar image
I do imagine, I'm not a native English speaker myself (I take it from your nickname that we might be from the same country ;))... Although I watched enough IT crowds and stuff to be able to do okay without subtitles.
(23 Dec '11, 04:24) patheticcock... patheticcockroach's gravatar image

That reminds me of a midterm in pre-organic chemistry at MIT many long years ago. I was scared of chemistry and the first lecture started with Schroedinger's equation which was incomprehensible, so I studied very hard and was thrilled to find the midterm easy. I got 110 points -- all 100 points correct, plus one of two extra-credit questions.

Then I found out I was below average. Oh man, did that rankle! Getting an extra credit question wrong shouldn't hurt your score, but everything was graded "on the curve" so what mattered was the class average, not the point count.


answered 22 Dec '11, 05:55

EllenL's gravatar image


edited 22 Dec '11, 06:01

Yep, but hey you did awesome. I really messed up the final REALLY bad because I was tired and needed to get it out of the way, so I'm in the 50% bracket ... but I finished the course!!!

I have a certificate and for that I'm proud. I've gained more through this course than just a score and my position in the class is irrelevant when I'm up against PhD's and CS graduates that have already covered the material.

All in all, despite my grumblings in the past I've made it through the other side into a world where I have a foundation upon which to take this subject in any number directions. I will watch the videos again to make sure that I understood everything. I've started reading Sebastian's book and I'm planning out my first robot.

In summary, we're all winners, if some won more than others ... kudos to them! I'm not gonna let that put me off the next challenge.


answered 22 Dec '11, 06:08

Merlin's gravatar image


edited 22 Dec '11, 06:09

Don't worry, if you do quizzes, it's really easy to do a mistake. Also, some questions were really strange, if you think about it, somewhat about definitions, even of philosophical type, more than on data.

Then, really, I don't think that the 100% people are going around pretending to be "experts" in the AI field! This would be ridiculous, because this course was really at a basical level.

I am a computer programmer with 32 years of experience and I am in my 50's, so I already knew how to play with flipped coins :->

Bu there have been many new things to learn and this directly from some of the major experts in these fields. This was great! I will follow future courses, but without enrolling, so that I can concentrate on learning and not on votes.


answered 22 Dec '11, 12:18

Finix's gravatar image


edited 22 Dec '11, 12:19

I got 84% for the class, which is in the bottom 50%. But I'm not too disappointed... I'm reminded of a joke:

Q: What do you call the guy who graduates last in his class at medical school? A: Doctor.


answered 22 Dec '11, 10:27

malcolmtech's gravatar image


Yeah, that's why to be world famous it's so hard. Think about it! 200k students joined aiclass! Only 20k remained for the final! So even if you get 90% overall it's among the top 5k, so it's not like a typical University - which has just 200 students in all and 90% is really high!

This answer is marked "community wiki".
Users with 100 karma points or above may edit this wiki.

answered 22 Dec '11, 05:52

AI_Occam's gravatar image


At least you didn't make a mistake in question #2! It got me down to the 25th percentile!! I wouldn't mind if it counted once, but it counted 4 times!


answered 22 Dec '11, 07:15

achkar's gravatar image


Oh man, question #2. I was on track for a 99.7% for the course until that baby swung into view -- it only takes one iceberg.
(22 Dec '11, 09:53) Tom Chappell Tom%20Chappell's gravatar image
Yes! I had 6 100s, a 100 on the midterm. I had #2 and k-means... I disagree with the k-means answer but 1 real mistake getting us down that much.. bummer
(22 Dec '11, 18:41) achkar achkar's gravatar image

I had 98% on midterm and got the "top 1000" mail. Then it took just two answers marked as wrong on Q3 (one of which is probably the most argued on this Community) to drop me down to 96.9% score and 25% percentile.

So what's next? Should I take the invitation to send them my resume as withdrawn? That sounds like waving a cookie under a dog's nose and then throwing it away... I wonder why they didn't wait until the final exam was closed.

I'm obviously aware that just being given the opportunity to send them my CV would give no guarantee to get a job in the Bay, but still it was an opportunity that is now gone and... well, that hurts.

The Higher You Are, The Harder You Fall


answered 22 Dec '11, 11:31

Fred%20G's gravatar image

Fred G

edited 23 Dec '11, 04:07

That is interesting. I thought that those who got 98% on midterm are already out of the top 1000. I for example did some silly mistake there, like swapped the order of answers or something like that. And haven't seen the top 1000 letter. Could it be that You did solve all Your other homework to 100%? You are correct about the waving the cookie effect. It is studied phenomenon that losing only by a small margin is much more painful than losing by large margin.
(22 Dec '11, 18:14) levitation levitation's gravatar image
I had 93% on the first one and 100% on all other HWs.
(23 Dec '11, 04:11) Fred G Fred%20G's gravatar image
Don't be silly. Just send them your resume, and good luck to you.
(23 Dec '11, 04:45) EllenL EllenL's gravatar image

I'm in top 25% instead of top 5% because I wasn't able to read correctly and in final-8 I checked nodes that will NOT be expanded instead of nodes that will be expanded.

So my lesson is learn to read :-)

(And to write as well, because I'm not good at that too)


answered 22 Dec '11, 12:16

zdenda's gravatar image


Funny, exactly the same for me. But top 25% is not that bad after all. And you know what? I understand particle filters!
(23 Dec '11, 04:39) ThM ThM's gravatar image
It's funny you mention that, because I did get all HWs and exams about particle filters right, but I'm still not sure I understand precisely how they work...
(23 Dec '11, 04:44) patheticcock... patheticcockroach's gravatar image

Well, I'll share my story as well.

I answered all the questions of the final on the day when the initial deadline was. I was satisfied with my answers. I was sure they are saved, even though there were server issues. Even though deadline was extended for 24 hours, I did not review the questions the next day. And that bit me in the ....

When I logged in and saw my score for the final exam, I was shocked at how could I have so few points. When I looked at my answers, well, I did not see many because many were not saved. I'd have had ~94% if all of my answers were saved, but I ended up with 77%. And it's completely my fault, I know.

I felt bad, but only for a while. The only problem is, now I don't feel like mentioning the score to anyone. I got 100% in ML and 96% in DB. 77% in AI, which I consider the easiest among the three does not feel that great. And I even got the digitally signed statement of accomplishment. Oh well :)

Still, I learned a lot. And will take another Stanford online class or two next semester.


answered 22 Dec '11, 11:49

awerti's gravatar image


Your answer
toggle preview

Follow this Question via Email

Once you sign in you will be able to subscribe for any updates here

Q&A Editor Basics

  • to upload an image into your question or answer hit
  • to create bulleted or numbered lists hit or
  • to add a title or header hit
  • to section your text hit
  • to make a link clickable, surround it with <a> and </a> (for example, <a>www.google.com</a>)
  • basic HTML tags are also supported (for those who know a bit of HTML)
  • To insert an EQUATION you can use LaTeX. (backslash \ has to be escaped, so in your LaTeX code you have to replace \ with \\). You can see more examples and info here