Wednesday, March 24, 2010

Judging American Idol for Newsweek (seriously!)

Back in March, 2007, I posted to my blog about how upset I was with the results from American Idol the week before. Well, it appears that someone at Newsweek was looking for University faculty members who were into American Idol and who could judge the contestants, much the way the actual judges do on the show.

Yesterday afternoon, I received a call from Newsweek inviting me to be a judge in their "When Eggheads Attack" piece. I participated, along with two music professors who really sounded like they knew what they were talking about. I was their "Ellen", providing my opinions without much expertise. The piece is available here, on the Newsweek site.

Besides the judging comments that appear on Newsweek.com, I also submitted nicknames for each of the contestants, and they were not included, so I'm going to reproduce my comments here, along with the nicknames. Ann helped me out with these, although I get the blame for the corny ones.

Lee DeWyze:

Lee the wise. I liked the jazzy interpretation and the way the song brought out your soulful voice. You sounded a little like Danny Gokey from last season. However, I think your song choice prevented you from showing us what you're really capable of. There is no doubt you're coming back next week, but you should take a bit more of a risk next time and really show us how you can stand out.

Paige Miles:

Turn the Paige. I had high hopes for you this week. You have incredible vocals and a wide range. Against all odds, the song started out weak and went downhill from there. I did not like the arrangement of this song, and I don't think it showcased your voice as well as your choices in previous weeks. Based on previous performances, I really hope you'll be back, and that you'll turn things around next week, but this performance puts you at risk.

Tim Urban:

Urban Warfare. On the plus side, you were definitely "safe" when you slid home. But that's the problem. The song was safe. You took on Queen, one of the hardest bands to simulate, and now, I think you're facing a crazy little thing called elimination.

Aaron Kelly:

Kid Kelly. It was fun watching you drool over Miley Cirus in the pre-performance video clip. You guys make a cute couple. If I close my eyes, your performance makes sense. When I open them, there's a mismatch between your body and that big voice. You are to singing what Doogie Houser was to medicine.

Crystal Bowersox:

Take a Bow-ersock: I expected it to be good. It was better than the original. Janis would be proud. I got goose bumps. You're definitely in the top two - see you at the finale.

Mike Lynche

Big Daddy: I would probably say you did a great job even if you didn't out of pure fear for my life. But luckily, I don't have to lie. You are incredible. You make every song interesting, and I can't wait to hear what you're going to do each week. Loved it!

Andrew Garcia

Straight Up Garcia: I don't understand how someone with such a good voice could make me dislike one of my favorite songs so much. I have to go listen to the original now to get that out of my head.

Katie Stevens

Cool KT: What you did was try to imitate the original, and you fell a bit short. I think you deserve to be in the top ten, but probably not the top nine. Definitely not the top eight. You have a good voice, and you're talented, but you're overmatched in this competition.

Casey James

Casey at the Bat: Best vocal I've heard from you all year. Sounded better than Huey Lewis. There is joy in Mudville. Mighty Casey has hit a home run

Didi Benami

Didi Banimi: You're good, you're good, you're good. Baby, you're good. But not great.

Siobhan Magnus

Driving on the Auto (Sio)bhan: You have been my favorite all year. They saved the best for last. I can't wait to see what you're going to do each week. This is how I felt about David Cook and Adam Lambert. The anticipation for your song is what makes idol exciting this season.

Follow up - teaching crypto to children

This is a follow up to my blog post yesterday about teaching cryptography to 5th grades.

It turned out that my class consisted of four students each from grades 5, 6, 7 and 8, not just fifth graders. The makeup of the class, which included older kids than I expected, did not really change anything, and the class size of 16 was perfect for the exercise I had planned. As many people pointed out to me, I had way too much material for one hour. In fact, the lesson I had planned out was better suited to a four hour session.

My typical students at Hopkins are about 15-20 years older than the Schechter students, and I found the class refreshing and entertaining. My initial observation was that some of the older kids did not appear too happy to be there. I saw kids staring at the ceiling or off into space. The younger children on the other hand, the 5th and 6th graders, seemed eager for me to start and made much better eye contact with me than their older classmates. However, once I started the class, I had everyone's attention.

Paul Revere's ride and the issue of signaling a code in the absence of cellphones and radio provided for a lively discussion, and I realized right away that this was a particularly bright and engaging group of students. What a pleasure.

Next, I described Caesar's cipher and substitution ciphers in general, and I asked the class to come up with ideas for how to break a simple substitution cipher. I wrote the list on the board as they made suggestions.


  • Pairs of letters
  • Common combinations, such as th, ch, sh
  • Using letter frequency
  • Small words, A, I, the, you


    Next, I had the students break into teams and come up with a cipher and a plaintext sentence and encrypt the sentence. Unfortunately, this took quite a bit longer than I would have liked, and by the time they finished, there were only about 20 minutes left. So, I gave them 10 minutes to spend cryptanalyzing the messages, and I had to break things off before any group had made any real progress. The students had a lot of fun reading their messages to the class. One interesting thing that happened is during the cryptanalysis phase of the project was that two of the teams decided to trade revealing one letter of their choosing with the other group. I had not specified that such bartering was allowed, but I let them do it, hoping that they would be able to finish in time.

    With 10 minutes left, I went back to lecture mode and showed them how to build increasingly complex ciphers from the cipher wheels. I got some oohs and ahhs, and I think that the exercise of trying to break a simple substitution gave them an appreciation for how hard it would be to break a three wheel cipher with keyed rotations. One particularly bright student, one of the younger ones, who asked some great questions (wonder if there's room for him at JHU some day in our Ph.D. program) said that he thought all of the strategies that were on the board were useless once the wheels started rotating, and he did not accept that anyone could actually break such a cipher, even using a computer. I did my best to assure him that computers could try many combinations of rotations of the wheel and look for recognizable plaintext, but I don't think he bought it. Once I stated that in WWII, the real progress in breaking the Enigma happened when an encrypting machine was captured, he seemed satisfied.

    So, overall, I had a blast. The students were great, and of course, the material made it easy. I just wish I had had a longer session. Okay, now back to my "adult" students.
  • Tuesday, March 23, 2010

    Teaching Cryptography to 5th Graders

    My kids attend the Krieger Schechter Day School. Today, they are having a learning festival where professionals from many different fields are teaching classes in their areas of expertise. The topics include, Drama, Animal Communication, Art, Conflict Resolution, Israeli Dance, Sports Signals, Hieroglyphics, Sign Language, Media, Electric Circuits and Morse Code, Woodshop and many others. I'm teaching a unit called Codes and Ciphers.

    Designing the class proved to be more challenging than I initially realized. These are very smart kids, but they are only 11 years old, and so one time pads, modular exponentiation, and Diffie Helman key exchange are off limits. I decided that the best approach is to give them something hands on to do. I teach in a few hours, and I'll post something afterwards about how it went. Here is the lesson plan.

    First, I'm going to talk a bit about basic codes. I'll describe "One if by Land, Two if by Sea". Although that code was not designed for confidentiality, it was nonetheless a code. Next, I'll move into Caesar ciphers and general substitution ciphers. Then, I'll divide the class into 4 groups of four students, as I was told there will be 16 students in the class.

    Elana (my 5th grade daughter) and I prepared cipher wheels out of paper plates. Each wheel consists of an inner and an outer paper plate. The inner plate was cut so that it is a few inches smaller in diameter than the outer one, and a tack was placed in the middle of the two plates so that they can spin independently of each other. We filled in the inner plates with the letters of the alphabet around the perimeter. The groups in the class will each fill in the outer plates with letters, lined up with the letters on the inner plate, to produce a substitution cipher. At the same time, they will generate a message of several sentences and encode it with the cipher.

    Each group will receive an encoded message from the another group, and they will use letter frequency and other clues to try to decipher it and to reveal as much as they can about the cipher from the other group.

    In the remaining time, I will show them how they can build an increasingly sophisticated cipher by turning the wheel after each letter is encrypted by a fixed amount and then by a variable amount. I'll show them how decryption would work by running an example with one of the cipher wheels and encrypting/decrypting a simple sentence on the blackboard.

    Next, I will show them how to combine three cipher wheels and rotations via a key to obtain a cipher similar to the Enigma Machine, and I will talk about the role of the Enigma in WWII, and how it was ultimately broken, as well as the importance of capturing a device.

    Finally, I will explain the intuition behind modern ciphers that use a key, and I hope that I can get them to appreciate Kerckhoff's Principle that the algorithms can be made public, and that all of the security of a cipher system needs to lie in the key.

    So, it's an ambitious undertaking, but I hope I can get through all of this and give the students an appreciation for the beauty and complexity of Cryptography while showing them a good time and not overwhelming them.

    Tuesday, March 16, 2010

    24 is GOOD again

    I've posted before about the Fox TV show, 24. My DVR records it on Monday nights, and I watch it early in the morning on Tuesday when I work out on my exercise bike. This morning, I watched last night's episode, and for the first time all year, I was completely gripped. I won't spoil it for those who have DVRed it and haven't watched or those who will watch it on Hulu. I will just say that the plot twist at the end was ingenious, and I felt stupid for not picking up on it. Well done. Probably the best episode in 3 or 4 years. The writing was clever - almost perfect. The only negative was the silly sub-plot about Dana's ex-boyfriend. Other than that, this may be a top 3 or 4 all time episode. (Either that, or the recent ones have been so dull and predictable that this one just stands out by comparison.) Anyway, 24 is good again. I'm very excited to see next week's episode on Tuesday morning.

    Saturday, March 06, 2010

    14 hours in Vegas

    After the RSA conference in San Francisco this week, I had 24 hours before I had to be in San Antonio for my nephew's bar mitzvah. So, I rearranged my flights, got a comped room at Harrah's, and ... free trip to Las Vegas!

    After checking in and dropping off my luggage in my room, I walked over to the poker room at Harrah's. It was dark, dingy and depressing looking, and there were only 2 tables with players. So, I crossed the street over to the Mirage, and found a beautiful, large poker room that was hopping. I bought into the 1-2 no limit Hold 'em game for $200 and started playing at 6:30 pm. The table was mostly weak with a few aggressive players pushing the action and several tourists who were in way over their head.

    I watched a couple of rounds without playing any hands until I picked up K-Q suited in middle position. An active player with a stack about like mine in second position (player A) raises to $6 (the bet amounts in this trip report are the best approximation I can make based on what I remember). I call, and two other players behind me (players B & C) call. The flop came K-T-3 rainbow, giving me top pair with a decent kicker. Player A, the original raiser, bets $12, which was half the pot, and I call. Player B, behind me folds and player C raised to $30. Player A folds, and after giving it some thought (although clearly not enough), I call. The turn card was another ten. That was a scary card given the betting so far, so I check. Player C pushes all in. I didn't see how I could call for all my chips with top pair decent kicker on my very first hand on that board. Furthermore, I had pegged player C as tight and this was her first big move since I sat down. So, after contemplating it a bit, I folded. Player C shows a pair of tens for quad tens, and in addition to the pot, she collects a special jackpot that the casino paid for anyone showing quads or better.

    So, I had a bit of a dent in my stack after one hand, but I had averted early disaster. About 4 hands later things weren't going well, I was down to $15, and in less than 15 minutes from when I played my first hand, I bought in for another $100. I had some unlucky breaks and in short order, I was down to $55. At this rate, it was going to be a very short night of poker. I got up and walked around and tried to relax and decided to tighten up and focus, so that my $55 would last a little longer.

    About an hour later, I was doing better and up to about $90 playing very tight and occasionally stealing some pots due to the table rep I had established. I wasn't getting any cards. Then I hit a turning point. I was in the big blind wit Q-5 offsuit. It limps around to me with 8 players and I check. The flop comes 5-5-3. After a couple of checks, one player (A) bets $10 into a $16 pot. Another player (B) calls, and I call, so three players see the turn. The turn card is a Q, completing my well-disguised boat. Player A bets $25; player B who had us both covered raises to $50. I push all in. Player A folds, and player B calls. The river card is irrelevant, and player B shows a pair of 3s for a lower full house than mine. I'm back in business, doubling up to $180. I got lucky because player B flopped his boat and I didn't get mine until the turn. Furthermore, had the Q not come, I would have lost a lot of money with my trip fives.

    I decided to shift gears and loosen up a bit. The more aggressive players at the table had left and were replaced by tight, passive players. I was able to chase out limpers pretty often. There was one guy at the table, Don Quixote (pronounced 'Donkey Chote'), in particular who like to limp and then raise on the flop and then fold on the turn to a big bet. I noticed this pattern with regularity, so I started exploiting it very successfully. His tactic, however, had been working for him surprisingly well (due in large part to luck), and he had a decent stack, well over $400.

    I had built myself up to about $300 (breaking even for the night at that point) when I was dealt suited connectors. I don't remember which cards exactly, but I ended up in an all in showdown with Don Q., and my flush beat his two pair, and so I doubled up again, to twice my total buy-in.

    My worst beat of the night also came against Don, who at that point, not surprisingly, was down to about $100. Under the gun, I bet $8 with pocket aces, and only Don calls. Heads up. The flop comes 8-J-Q with two spades. I bet $25, hoping to shut out any draws he might have and take the pot right there. Don raises me all in. I had to think a little bit. The hands I can think of where I'm not a favorite are 9-T, two pair, and the three sets. Don was always raising preflop with pairs, even low ones, so I had to figure he had flopped two pair or a straight. He was also very capable of a bluff, and I also figured he might make that play with hands like A-Q or (more likely since I had 2 aces) K-Q. The way Don played, I also thought he could push with any two spades as some kind of misguided semi-bluff. I had no idea which of these I was up against, although I suspected it was not a set, and I didn't see much choice with a third of my stack already in the pot and a decent hand, so I called, and he turns over 9-T of spades and ends up getting his flush on the river, although he already had me beat with the straight.

    The characters at the table changed over time, and I stayed until 3:00 a.m. - my longest poker session ever. At one point my stack got as high as around $700, and I ended cashing out $518. When I got back to my room, I got ready for bed and called Ann to say hi. She was just getting up at 6:20 a.m. EST in Baltimore. I set a wake-up call for 4 hours later and got a little bit of sleep. I'm on the plane now heading to San Antonio. Tired as hell, but I had a total blast. Can't wait to play poker again, and incredulous that I came out ahead on my first trip to play poker in Las Vegas. I'm sure I'll be redistributing my winnings to the folks at my regular house game before long.