Jeopardy Progress Update

It's been a long time since I've written here, and I think the primary culprit has been trivia taking over most of my recreation time. I don't watch as much premium TV and movies as I used to, much less find motivation to write about them. So I'll somewhat explain my absence and get to where I am with how prepping for Jeopardy! is going, and some things I've been doing to track my progress.

First off, I'll clarify that "progress update" is solely about how well I'm performing at playing the game and not about an imminence of me appearing on the show. I'm not too far along in the testing / interview / selection process (and since we haven't seen a new contestant on Jeopardy! all season long, seems like no one else is either), so I don't want to make it seem like I'll be appearing on someone's television anytime soon.

I posted my trivia origin story years ago on this site, but the short version is that in 2017, I was one of the worst trivia players around and knew very little about anything. I started "playing" Jeopardy in 2018 -- sitting in front of my TV with a buzzer and recording, on each question, whether I'd failed to buzz in, buzzed in and got it correctly, or buzzed it and got it incorrectly. I would then add up the dollar values and compute my Coryat Score, a thing described briefly here and more extensively here (great URL, Karl). Basically it's the sum of the dollar amount of the questions you get right, less the sum of the dollar amount of the questions you buzz in on and get wrong.

It was clear that I was pretty bad, and I expected to be. A "good" score for me at the beginning was cracking 5-digits, which would happen if I stumbled into a couple higher value clues and didn't make too many wrong guesses. But I have many, many sheets of paper showing that I didn't crack $10,000. I think I even finished in the negatives a couple times (which I know people do on the show sometimes, but it's a LOT worse to finish in the negatives at home when there's no buzzer competition and no wagering). Work needed to be done.

After playing Jeopardy! for awhile, it was clear that the best strategy was to know a little bit about a lot of different topics. This was not ideal for me, a person who knew a lot about the White Sox, poker, and Starcraft and not all that much about anything else. I didn't need to be able to write an essay on the French Revolution, but I did need to know key dates and people. Initially I started with memorizing countries and capitals (Sporcle is a great resource) and paintings/artists. I read Alan Sepinwall's excellent book on television, and started memorizing the histories of the big awards at the Oscars and Tonys (retention is not great at these...)

It became clear after playing both Jeopardy! and LearnedLeague for awhile that history was probably the single most important subject to study, given its prevalence in both formats and significant overlap with other categories of knowledge. I was appalled at my performance in U.S. history, so I read through an online U.S. history text. I then moved on to read a world history textbook that some high school posted online that I can no longer seem to find. These were important to do, but it became clear that much like in school, I wasn't going to permanently retain this information. Flash cards are apparently an invaluable resource in the trivia community, so I started to make those from key figures and events in the text. 

History seemed like a great place to start because it seemed finite. There's so many different types of science and a ridiculous amount of books that have been written over the years, but I figured I was unlikely to encounter many history questions on things that weren't in one of these two textbooks (side note: WRONG!). For the other subjects, I just relied on expanding the scope of trivia I'd play. I joined BPTrivia. I signed up to play in Jeopardy! great Alex Jacob's School of Trivia. More recently, I joined PAR, a fantastic competitive speed trivia competition where I can go up against some of the greatest trivia players around (which has been humbling, but also an excellent experience). Whenever I encounter a question in any of these formats that I don't know but seems in the realm of knowledge, I make a flash card and add it to the bank.

Slowly, all of this prep has borne fruit, but it has not been easy. I have a wife, two children, and a job, all of which are wonderful but none of which are conducive to improving trivia performance. It has meant that in addition to spending my spare time watching new Jeopardy! shows and having reruns on in the background while cooking or doing chores, I needed to carve out time to study flashcards. Doing this for 15 minutes constitues a "good" day, but it's not enough time to see quick results. It has meant that at many points, the simple variance of day-to-day questions on Jeopardy! has made me question whether I'm even getting better at all. The plateaus in Coryat score are frustrating, but the steps back? They've made me want to incinerate my self-created Jeopardy! game cards and chuck my buzzer at the wall. 

But it's clear that despite short-term glitches, the overall trend has been good. Here's what my 30-day average scores have done over the last 676 games. I don't have dates on here, but it's roughly my progress over the most recent three years and change (so, starting in fall 2020).

I'm hoping I can push this to be consistently over 30,000 in the next year or so, but it will keep getting harder to improve. I don't know what level would make me satisfied that I'd have a good chance to win going in. I can only imagine that superstars like Matt Amodio and Amy Schneider average something closer to 40,000.

J-Archive computes a "combined Coryat" that combines the individual performances of the players on the show. I wanted to see how I stacked up, so I started taking my averages and comparing them to this combined performace of the contestants. It's a little hard to interpret, but I think a rating of 100% on this metric would mean that I played as well as all three contestants would have working together. Clearly I'm not there yet, but I think it's a reasonable goal to average that if I want to have strong hopes of success.

Note that the dip on the right, while seemingly discouraging, is actually expected. This season, due to the writers' strike, Jeopardy! has only had past champions and those who they've determined to be the very best runners-up of the last few seasons on the show ("Second Chance Tournament"), so the contestants have been much better on average in Season 40 than the second half of Season 39. The player quality in Season 39 was also very much front-loaded; the back portion with that nice bump around the 280-300 range was well after greats like Cris Pannullo, Ben Chan, Hannah Wilson, Troy Meyer, and Ray Lalonde made their mark, so it was a little artificial.

So, what does this all mean? What would my chances be of actually winning on the show? I've started to assign each game I play a "W", "L", or "M" (maybe) to indicate whether I think I would have come out ahead of the contestants on a particular day. In my earlier days of playing, there'd be games that I'd assign a "W", but almost all of them would be due to the flukiest method possible -- weaker contestants allowing me to be in striking distance and then a Final Jeopardy that I happened to know that the leading contestant(s) missed. Despite improving enough to win by other means, Final Jeopardy does seem to be the single biggest determinant of whether I feel like I'd be successful on a given day.

Unfortunately, Final Jeopardy seems to be one of the weakest aspects of my play at the moment. I'm running hot right now, but over the long-term, I'm tracking at just a little over 50%, which isn't gonna cut the mustard. 

                         

A big reason for this is the way I've prepped, which is built largely on associating one thing with another (year/President, book/author, atomic number / element, etc.) and Final Jeopardy doesn't lend itself as much to simple memorization. It favors people who have had a passion for knowledge/experience all their lives who have deeper understandings of different topics over dolts like me who started grinding trivia in their 30s.

Daily Doubles are the other big determinant of success, and with a larger sample, it's a little less noisy than Final Jeopardy. Unfortunately, not much progress in this department over the course of the last year-plus (I didn't start tracking these until more recently). 

So all in all, if I had to compete on the show tomorrow, I probably would do well enough to look competent and might have a one-in-three chance or so of winning, as it seems from my tracking that I could steal a game or two per week. But it could be a blessing that I haven't been selected yet, because there's much more improvement to go. The counterpoint is that I'm not going to be (relatively) young forever, and with each passing year I get a little more anxious that it's just never gonna happen. In a normal year, close to 400 people from the general population will get their first cut at the show (this has decreased with the recent emphasis on tournaments and past contestants), and compared to how many people are trying, that's just not much.

So anyway, this passes as some semblance of an explanation of why I haven't been writing about TV and movies, and probably why I've become less on top of things baseball-wise than I was a half-decade ago (though the White Sox have done their part!). I'll keep you all posted on whether all of this nonsense looks like it's gonna pay off some day. Even if it doesn't, everyone needs something to shoot for, and this has given me a new competitive outlet, which I always need. So I guess....I REGRET NOTHING!!!!!