Probable question access at MIT Penn-ance

[reindeer]

As announced in the other thread, MIT is vacating its first place win at Penn-ance. We can’t provide proof of misconduct by Josh as definitive as NAQT can for SCT and ICT, but we wanted to explain the observations and events that led to this decision.

We were very under-staffed for Penn-ance, and Josh was extremely active in getting friends and acquaintances of his to volunteer to help out. At the time I was surprised by how prolific he was, although looking back at old emails it appears that he’d done similar things for previous tournaments, though on a smaller scale.
On 10/13, I received an email from Josh saying that his friend “Paul” (all names of Josh’s acquaintances have been changed) was available to staff, and could be reached at “[email protected]”. He also asked me to send Paul the packets ahead of time so that Paul could look them over and familiarize himself with words he didn’t know. Given that Paul was relatively new to quizbowl, this request didn’t seem all that unreasonable.

A few days later, I sent an email to all the Penn-ance staffers I’d accumulated so far, including Paul. Paul sent me a very confused response--he had no idea what I was talking about and had no recollection of expressing interest in staffing the tournament. I apologized to Paul and asked Josh for an explanation. Josh told me that he had asked two of his friends (Paul and “John”) to staff at the same time, and John had been the one who expressed interest, not Paul. John’s email address, according to Josh, was "[email protected]", and it was John who had wanted me to send him the packets.

At this point I should note that Josh also gave me the names and email addresses of several other people, both quizbowlers and friends of his, who were interested in staffing the tournament. Further, Josh told me that one of them, “Frank”, also wanted me to send him the set ahead of time for practice. Eric sent two versions of the set (with and without passwords) to Stephen and me at 6:27 on the morning of the tournament, and I forwarded them to Frank and John at 7:15. All other staffers read the tournament from laptops, receiving the passwords one round at a time on their scoresheets. (Anderson read in the morning and played in the afternoon; since we didn’t know when he’d be able to switch to playing, I sent him the rounds one at a time rather than all together).

Frank showed up to the tournament having received the set, but John did not come at all. This was inconvenient, but since there were several people who had committed to staffing and never arrived, it didn’t seem especially remarkable that he was among them. By the end of the day, though, enough people had expressed surprise at Josh’s performance to make me wonder if it was possible that he had been cheating. Neil told me that the John in question was a friend of Josh’s named John Peterson, but when I looked John Peterson up in the MIT directory, it said that his email address was “[email protected]”, not “[email protected]”. It’s my understanding that MIT students are only allowed to have one @mit.edu email account, but we can create “mailing lists” that forward to that account and make it appear that we have more than one. Thus, it was clear that “[email protected]” was not John Peterson’s official MIT email account, but it could have been a mailing list belonging to him. At the same time, it could also have belonged to anyone else.

In retrospect, this whole chain of events definitely looks very suspicious. However, Josh is often disorganized, especially when it comes to logistical matters, so the idea that he would have mixed up his friends like he claimed would not be unreasonable. In addition, his performance was in line with his SCT/ICT 2012 stats, which at that time we believed to be legitimate. Therefore, without any solid evidence to go on, we pushed it to the backs of our minds, and more or less forgot about it until NAQT contacted us this January.

At that point, it became clear that the major argument against Josh’s having cheated at Penn-ance (the continuity with his NAQT stats) had been negated. In meetings with an MIT staff member, we explained the chain of events concerning the "[email protected]" email address, and asked whether MIT could tell us who it belonged to. After a few weeks, we were told that MIT was not allowed to release that information. With no other obvious way of resolving the situation, we again mostly let the issue drop. We planned to vacate the Penn-ance win but were concerned about retribution from Josh; the recent re-opening of the cheating/question security discussion seemed to provide a good opportunity to do that, even in the absence of definitive proof.

We had considered emailing John Peterson and just asking if he is the owner of the "[email protected]" email address. Until now, we hadn’t done so out of concern about the possibility that John and Josh had been working together, and that being direct with John would alert Josh to our suspicions. Today, we decided that the potential benefit outweighed the risks, so I emailed John at his official MIT email address (jpeterson) with a short explanation and asked what he knows about johnp. He told me that the johnp email address “has nothing to do with me”.

While this is not incontrovertible proof that Josh acquired the Penn-ance set this way, I find it difficult to imagine other plausible explanations for the sequence of events. We would have vacated the Penn-ance win regardless of John’s response, but we wanted to wait to provide a lengthy explanation of our reasoning until we knew as much as possible.

With MIT’s victory removed, Brown B will move up to 3rd place; Yale A and Harvard A are free to agree to tie for first place or not as they see fit.

[Auks Ran Ova]

A good lesson to take away from this: never, ever, under any circumstances, send out an unplayed set to anyone for "practice reading". If I recall correctly a similar angle was used by Amit to fraudulently gain access to the NSC set in 2010. Keep in mind that any of the many hundreds of publicly-available sets will work just as well for practice reading material!

[reindeer]

Maybe I should have been more clear; the statement was that John wanted the set to practice reading those questions specifically, to practice pronunciation and so on. I get this request a lot when I run tournaments, mostly from inexperienced readers.

[Auroni]

Maybe I should have been more clear; the statement was that John wanted the set to practice reading those questions specifically, to practice pronunciation and so on. I get this request a lot when I run tournaments, mostly from inexperienced readers.

Even despite that, it's still not a good idea to send them the actual set that will be played. As a moderator, you'll read most of the same words from tournament to tournament. There might be a unique, difficult to pronounce word that's part of a clue that only shows up once at a given tournament, but that's hardly worth the security risk. The point is to practice reading quickly in general, not reading a specific set of questions quickly.

[Cheynem]

It's a reasonable request, although I think even if there's no malicious intent involved, still a bad idea, especially if you have never met the prospective readers (for instance, I, an experienced quizbowl person, sometimes have trouble talking about unclear sets before a tournament--I would imagine this to be harder for inexperienced quizbowl people).

[Auks Ran Ova]

Maybe I should have been more clear; the statement was that John wanted the set to practice reading those questions specifically, to practice pronunciation and so on. I get this request a lot when I run tournaments, mostly from inexperienced readers.

Even that seems like an unnecessary risk--any similar-in-difficulty already-released set will likely share many potential problem words with the set in question, and should provide an entirely sufficient simulacrum of the reading experience.

EDIT: To be sure, I'm not trying to accuse you of being neglectful or stupid or anything! When not made with malicious intent, it is a perfectly reasonable request. I'm just suggesting an alternate solution that will accomplish the same end with none of the potential security issues.

[Sen. Estes Kefauver (D-TN)]

Yeah tell prospective moderators "Sorry, but too many people have cheated that way for it to be an option. You can read the questions during the morning meeting if you need to."

[Ethnic history of the Vilnius region]

A good lesson to take away from this: never, ever, under any circumstances, send out an unplayed set to anyone for "practice reading". If I recall correctly a similar angle was used by Amit to fraudulently gain access to the NSC set in 2010. Keep in mind that any of the many hundreds of publicly-available sets will work just as well for practice reading material!

I think this needs to be communicated very clearly to mirror sites. Like people in the thread have mentioned, it is a reasonable request for an inexperienced reader to ask to see the questions in advance for practice reading purposes (assuming no malicious intent, of course). It's also reasonable to think that an inexperienced TD (heck, even some experienced TDs) would oblige such a request. Thus, editors should let their mirror hosts know to just point inexperienced readers to the archives and to not send unplayed sets out to anyone before t-time.

[mtimmons]

Do we know if "John" ever intended to staff or even knew about the tournament or if Josh was being dishonest when saying that John was interested in staffing?

[reindeer]

I don't know; he hasn't volunteered a lot of information in his emails to me. My guess is he was never intended to come (how would Josh explain John showing up without the set?), but again that's just a guess. I suppose I could email him again to ask directly, but I don't really want to interrogate him and I also don't think it would add much to the story.

[Masked Canadian History Bandit]

This is an edited transcript of what Sinan sent me over Facebook after the news broke out:

"Any chance he got his hands on the Penn-ance packets? He had the weirdest neg in our game. He negged Compton Effect with "triple alpha." A. triple alpha is too hard for this difficulty
but a couple of tossups later TRIPLE ALPHA COMES UP. Then he seemed to freeze up and completely underperform."

Seems to be clear cut evidence of cheating combined with all the other evidence.

[Adventure Temple Trail]

Olivia, thank you for doing all you've done to investigate and handle this situation well.

[reindeer]

"Any chance he got his hands on the Penn-ance packets? He had the weirdest neg in our game. He negged Compton Effect with "triple alpha." A. triple alpha is too hard for this difficulty
but a couple of tossups later TRIPLE ALPHA COMES UP. Then he seemed to freeze up and completely underperform."

Can I strongly suggest that people tell the TD or someone similar when things like this happen? I alluded to this in the post, but immediately after the tournament ended I was pretty worried about Josh having cheated. But there wasn't any real evidence, especially since I didn't see any games, and none of the staffers I talked to seemed to share those concerns. Because no one else seemed bothered, I attributed my concerns to residual paranoia from the Dimitri/TJIAT incident in high school and was willing to let it all drop. I'm virtually certain that if even one person had told me about an incident like this, that would have been enough to convince me to take some kind of action (like emailing John) right away.

[mtimmons]

I remember Josh negging triple alpha and then it coming up later as well but it didn't really stick out at me as obvious evidence that he cheated until we learned that he cheated at SCT/ICT last year. If you start with the premise that Josh's SCT/ICT stats were legitimate I don't think it's that surprising that he puts up something like 31/5/2.

[Cheynem]

I understand your point about telling the TD about these things, but remember: none of the players had the knowledge you did about the somewhat suspicious events leading up to it. There are plenty of instances when I've seen someone neg things that were going to come up later by chance (and I've done this as well). I would never tell a TD about it in and of itself.

The one thing that I hope doesn't emerge from this is cheating accusations thrown around willy nilly. I realize that's not what you were getting at, but I just wanted to point out it would be extremely hard for someone to go to a TD over just that incident.

[Windows ME]

I understand your point about telling the TD about these things, but remember: none of the players had the knowledge you did about the somewhat suspicious events leading up to it. There are plenty of instances when I've seen someone neg things that were going to come up later by chance (and I've done this as well). I would never tell a TD about it in and of itself.

The one thing that I hope doesn't emerge from this is cheating accusations thrown around willy nilly. I realize that's not what you were getting at, but I just wanted to point out it would be extremely hard for someone to go to a TD over just that incident.

Yes, this is why I didn't tell anyone - neither myself, my teammates nor his teammates for that matter thought nothing of it at the time other than an amusing coincidence.

[grapesmoker]

You can't put too much stock into one question. I recall playing tournaments where I'd neg with the same answer multiple times in the tournament and not pick it up when it did come up. It's not that unusual, really. On the other hand, I have never heard of a 31/5/2 statline obtained legitimately, and certainly not by someone who the year before had gone 0/2/0 in DII.

[Tees-Exe Line]

He had the weirdest neg in our game. He negged Compton Effect with "triple alpha." A. triple alpha is too hard for this difficulty
but a couple of tossups later TRIPLE ALPHA COMES UP. Then he seemed to freeze up and completely underperform."

Shantanu did this ALL THE TIME in practice. So it must just have been a coincidence.

[STPickrell]

You can't put too much stock into one question. I recall playing tournaments where I'd neg with the same answer multiple times in the tournament and not pick it up when it did come up. It's not that unusual, really. On the other hand, I have never heard of a 31/5/2 statline obtained legitimately, and certainly not by someone who the year before had gone 0/2/0 in DII.

Wait, 31/5/2 wasn't a typo??