MARCATo General Thoughts

[ThisIsMyUsername]

This thread is for general thoughts about the tournament, such as about the format, mechanics, philosophy, etc. Basically, for anything that doesn't pertain to just Wonyoung's questions or my questions.

[tpmorrison]

I really enjoyed this tournament - thanks so much to John, Wonyoung, and Noah for all their hard work! The questions were lots of fun to play, and even the ones I knew nothing about were interesting to listen to.

The daily format was a fantastic idea and kept me a lot more excited about the tournament throughout than an all-day event would have. I’d love to see some more online side events use this format in the future, when possible.

I was also really impressed by the Qblitz platform and thought the “Review Questions” and “View Matches” sections were great features. One minor comment is that, a few times, I accidentally clicked on the play button instead of the buzz button when trying to buzz on a question. It’s quite possible I was the only moron who made this mistake, but I think graying out the play button when the question is being read might remove this issue.

[Asterias Wrathbunny]

I really enjoyed playing this tournament. Clearly a lot of effort went into making this happen, and I thank Noah, John, and Wonyoung for all their hard work. Thanks especially for being responsive and attempting to work out a solution when I scheduled travel in the middle of the tournament.

[gimmedatguudsuccrose]

First off, I wanted to say that I really enjoyed this tournament! The classical music content was excellent and exceeded my expectations (which were already extremely high for John Lawrence music tournament) in that they managed to keep the difficulty below previous ILs while still being both exciting and fresh. I'm less qualified to discuss the popular music, but I really enjoyed that portion as well! My only complaints are along the lines of what Tim pointed out in the other thread. I will say that on questions that I have relatively deep knowledge on (Californication and Def Leppard to name a few off the top of my head), Wonyoung did an extremely good job of clue selection in terms of both a difficulty gradient and picking interesting moments. Thanks again to you two for writing such a spectacular tournament!

This was the first tournament I played on the Qblitz platform, and I was relatively impressed by how easy the interface was to use! The tournament proceeded very smoothly despite the occasional hiccup. I particularly enjoyed the timing of the tournament - not only was it a nice change of pace from the typical one tournament over the course of a day/weekend, it meant that a time commitment of at most half an hour was required every day, which is nice for people with busy schedules.

I have some thoughts about possible improvements to MARCATo's format. I've been informed that the format that was used with 2 parallel playoff brackets was the only format that guaranteed that all 12 rounds were played by every team. Personally, I would've personally preferred a format where there was a single championship bracket at the end of the day with round robin play, with perhaps an extra round used for seeding. I'm not just saying this because it would've been advantageous to my team, but because I think it's fair in general if all the championship contenders have a chance to play one another.

One format that I believe should be considered for future tournaments like this on Qblitz is every team playing every other team on all of the packets! This seems like the logical next step in asynchronous online league-style tournaments, and gives results closer to the true strength of the teams relative to one another due to the large increase in number of games. The way that Qblitz works, we were able to visualize our results against every team in every round, so I don't think this would be that challenging to implement (although I could be wrong). Alternately, the tournament could use a system similar to Buzzword with celerity scores. Even if people wanted to continue the 1v1 feel of an IRL quizbowl tournament, the rounds against every other team could be employed either in lieu of the standard 'prelim' format or as a tool for breaking ties (instead of PPG).

[ErikC]

I agree with previous posts about the difficulty and bias towards recent music in this thread, but I'd like to add some more thoughts.

Isaac pointed out the dearth of 90s content, and I'd also say there wasn't too much 00s music either, particularly in the more radio-centric genres. I'd need to see the actual numbers to be sure but I think the 21st century basically meant 2010s most of the time. Pointing to a few tossups for a set with 120 tossups as squarely 90s when its a decade with a lot of iconic music isn't very fair to people who love those artists.

I think lumping R&B with hip-hop is incredibly unfair to the genre. Hip-hop deserves at least 1 tossup every pack as the most popular genre (1) (2) currently with plenty of well known artists and albums from several decades (was there any 90s hip-hop besides Biggy? It's a pretty well-liked era). R&B is a completely different genre with a different origin and history, and I think the majority of the artists asked about in the genre were not representative of all the genre has. I'd understand if some of this as to keep it "easier" but considering how hard some tossups were like Lauv (do that really need an entire tossup?), I think some questions on artists like Usher or Mariah Carey would have been fairly within the tournament's scope.

Another quirk of this tournament was how few questions there were on electronic music. Perhaps a few of the pop questions had some EDM songs in them, but I don't think there were any answerlines that were about electronic artists besides Zedd. This might not have been a conscious decision, but instead a product of the distribution being half a mix of questions. This mix ended up largely being common links that covered the genres already in every pack, besides the country questions which seemed fairly hard for a casual listener. I'd have preferred more straight-forward questions on artists from a variety of genres. I think MIXTAPE's distro would have been a good model.

There was a bit of bias towards indie-famous rock and pop artists like Dacus, Fiona Apple, and Lauv. These questions were often quite hard. I think its demonstrative that London Calling and Dark Side of the Moon were the choices for those bands and Tidal was chosen for Apple. The difficulty felt pretty bimodal because of this. There was a fair amount of questions with early clues (the Roses piano opening, the opening of Nonstop) that were easier than the giveaway for these indie-famous artists that seem to be popular with a sub-set of the qb community.

I think most of this comes from having a single writer for such a wide range of music and from the distribution being set up the way it was. On the music I particularly like in this tournament I really liked how it was clued. Buzzing on albums was a bit of a challenge, and most of my negs were from them. From a purely gameplay standpoint I think working a way to reduce them would have at least improved my enjoyment.

[magin]

I was also really impressed by the Qblitz platform and thought the “Review Questions” and “View Matches” sections were great features.

Agreed. I was not expecting to enjoy seeing the buzz distribution for each question while re-listening to it, but that became one of my favorite things to do each day. One question--how much additional work would it be to determine a how a team did against the rest of its bracket or field per round? That could be an interesting way to run this instead of head-to-head matches in order to reduce pure packet variance.

I was also really impressed by the Qblitz platform and thought the “Review Questions” and “View Matches” sections were great features. One minor comment is that, a few times, I accidentally clicked on the play button instead of the buzz button when trying to buzz on a question. It’s quite possible I was the only moron who made this mistake, but I think graying out the play button when the question is being read might remove this issue.

I also did this at least 5 times. It's a minor fix, but I agree that graying it out would improve the experience overall.

[warum]

I tried to simulate a full round robin on all rounds of the first MARCATo mirror, based on buzzpoint data that Noah kindly provided. The resulting team records and PPG, and individual player PPG, are available at https://github.com/natan-holtzman/MARCATo_stats, as well as the Python code used to run the simulation. The team stats spreadsheet is sorted first by number of wins, and then by number of ties if two teams have the same number of wins. The player scores spreadsheet is sorted by points per game.

To simulate games, I made a few simplifications to the rules:
(1) If the first buzz for a given tossup in a game is incorrect, it is counted as -5 points if before 60 seconds, and as 0 points if after 60 seconds. Not all tossups necessarily lasted exactly 60 seconds, but the buzzpoint data didn't specify where each tossup ended, so I used 60 seconds as a cutoff between negs and zeros.
(2) I accounted for ties where two players have the same correct first buzz at the same time, by splitting the 10 points evenly among the two players/teams. If three or four buzzes occur simultaneously, the simulation only counts the first two listed in the buzzpoint data. This has the potential to affect team scores. In the situation that one team buzzes incorrectly and two players on the second team later buzz in with the correct answer simultaneously, I did not try to split the buzz among both players. Such a buzz would arbitrarily be credited to one player. This would only affect player scores, not team scores or records.

[Asterias Wrathbunny]

This is late, but I was talking with Isaac about different MARCATO buzzes as we prepare for PAveMEnT, and decided to import the csv of buzz data into Google Sheets, filtered it for just the correct answers, cleaned it up, and sorted it so that it's readable. Here's that sheet if anyone's interested.

https://docs.google.com/spreadsheets/d/ ... sp=sharing

[ErikC]

This is late, but I was talking with Isaac about different MARCATO buzzes as we prepare for PAveMEnT, and decided to import the csv of buzz data into Google Sheets, filtered it for just the correct answers, cleaned it up, and sorted it so that it's readable. Here's that sheet if anyone's interested.

This is really cool, is there a chance we could see the second mirror's stats too?

[Asterias Wrathbunny]

Unfortunately, I don't have the buzz data for the second mirror, and I don't know if Noah's sent it out to anyone. If someone does have it, I could put it in this format pretty quickly.