Power theory (was: Am I just stupid/ignorant?)

[Ondes Martenot]

So I'm looking at the ACF article on the quizbowl and I noticed that it says:

"However, NAQT represented a regression to certain College Bowl-like features...gimmicks such as the power tossup"

So I have to ask, what is gimmicky about powers? It rewards someone for having a deeper level of knowledge. It seems to make perfect sense. Am I just stupid/ignorant?

[Matt Weiner]

So I have to ask, what is gimmicky about powers? It rewards someone for having a deeper level of knowledge.

Getting the tossup before the other team and the bonus rewards someone for having a deeper level of knowledge. Focusing on the extra five points instead of on the 80 point swing involved there is something that's done for artificial "excitement." As many people have said before, it's among the more harmless gimmicks, since 5 points here or there does not impact the outcome of the game as much as the quality of any given question does. But, it's a gimmick nonetheless.

[Ondes Martenot]

True. Games are won on bonuses, not powers (usually). But there's a difference if you get a tossup right through a buzzer race at the very end of the question as opposed to getting right unopposed early on in the question

[Mechanical Beasts]

But there's a difference if you get a tossup right through a buzzer race at the very end of the question as opposed to getting right unopposed early on in the question

Yes, that difference being that against good competition the former won't happen. Against bad competition, when your team is getting 90% of the tossups, it's true that there's no other metric to distinguish between a buzz on the first and last word. But then again, you'll beat bad competition any way, so it doesn't much matter. Against good competition, you only get the tossup if you buzz early: deep knowledge is rewarded by points, whereas shallow knowledge is rewarded by nothing. The extra five points don't really say nearly as much about how valuable the player is, though it does say a lot more from an individual stats perspective.

I'm holding out hope until we build our perfect quiz machine so that stats can note the word players buzz on.

[cvdwightw]

The one thing that can be said that is good for powers outside of this artificial "excitement" is that it provides a second, complementary measure of a team's depth. Bonus conversion measures both breadth and depth of a team's knowledge; powers arguably measure how much of that is depth. If team A and team B each have 100 tossups and a bonus conversion of 15 over the course of the tournament, but 20 of team A's tossups are for power whereas 5 of team B's tossups are for power, I think it would be safe to say that team A is likely to have better depth than team B. Without power marks, team A and team B look roughly equivalent from a statistical point of view.

Also, that line of reasoning about "against good competition the former won't happen" is ridiculous, especially coming from someone who played at ACF Nationals, a tournament where THE FORMER HAPPENED ALL THE TIME in matches between two of the best teams in the nation. If your team grabs 5 powers and 2 regular tossups and my team has only 2 powers but beats you to 8 buzzer races well after the power mark, then my team is probably going to win. Or maybe your team is so good that you never let the question get to the end before getting it right, and if so you have a team that could blow TJ '05 out of the water. Against good competition, you get the tossup by either knowing more than your opponent, or by knowing the same as your opponent but being a fraction of a second faster than them. And the latter happens a heck of a lot more than you're giving it credit for.

[Mechanical Beasts]

The one thing that can be said that is good for powers outside of this artificial "excitement" is that it provides a second, complementary measure of a team's depth. Bonus conversion measures both breadth and depth of a team's knowledge; powers arguably measure how much of that is depth. If team A and team B each have 100 tossups and a bonus conversion of 15 over the course of the tournament, but 20 of team A's tossups are for power whereas 5 of team B's tossups are for power, I think it would be safe to say that team A is likely to have better depth than team B. Without power marks, team A and team B look roughly equivalent from a statistical point of view.

Yes, which is a good argument for tracking early buzzes, but which doesn't support them being worth more. Arguments in favor of powers suggest that an early buzz should earn you more points, on top of the fact that the earlier your buzzes, the more tossups you're getting.

Also, that line of reasoning about "against good competition the former won't happen" is ridiculous, especially coming from someone who played at ACF Nationals, a tournament where THE FORMER HAPPENED ALL THE TIME in matches between two of the best teams in the nation.

This is true, but this is the newbie board dedicated to high school quizbowl, the beneficiaries of which don't play ACF Nationals. So, yes, if you're playing questions astronomically harder than NAQT A sets, then it's valuable to be able to get tossups after the last word.

But against good high-school competition--say a team that will go 6-4 at HSNCT--early buzzes are rewarded plenty because playing against empty chairs, that 6-4 team might get 75% of tossups on the second-to-last line, so your before-power-mark buzzes are extremely valuable compared to your buzzes on the last word, since only 25% of the time do you get a shot at displaying that skill.

Let's say you're playing an opponent through a computer. You both buzz at some point in the question and are right; whoever first gets the bonus. Ceteris paribus, the earlier buzz is more valuable because it's more likely to have come before your opponent's.

If your team grabs 5 powers and 2 regular tossups and my team has only 2 powers but beats you to 8 buzzer races well after the power mark, then my team is probably going to win.

I don't like your argument here. It's true--powers in themselves aren't something that magically make you win; if you answer seven tossups and your opponent gets ten, you will probably lose. I don't think I was ever contesting otherwise. I was saying that an early buzz shouldn't be motivated by a cookie; it should be motivated by the fact that the opponent may buzz at any moment and thus deprive you of the bonus. That doesn't have anything to do, so far as I can tell, with a hypothetical situation in which a team buzzes early but doesn't answer many questions. A better example would be: my team knows 17 answers, your team knows 18. Let's say that my knowledge is a subset of yours. We buzz earlier than you on 13 of those and miss two due to aggression, but we're buzzing earlier than you so we get eleven bonuses (and two negs), while you get seven bonuses. That's where the value comes from.

[Matt Weiner]

Since powers don't really matter, it's good to have them if they really do bring in a bunch of new teams. Again--a gimmick that you shouldn't be overly concerned about, but gimmicks that increase participation in real formats without having much impact on game outcomes should be supported as a net benefit.

[Frater Taciturnus]

How much energy is put into making and evaluating power marks before sets are played though? If the answer is a lot, then think of how that energy could be refocused into improving (or even expanding) production.

If the answer is not much, then obviously not much is cut by cutting them, except criticism along the lines of "That was in power? Seriously

[yoda4554]

I'm curious what people would think about a more concrete example. At this year's Penn Bowl, my team was playing Maryland B on the Princeton packet. Mostly, that game was a shootout between me and Andrew Lim; we answered 8 and 9 tossups respectively, with me taking most of the lit. and arts, him taking hard science and history, and us fighting over the rest toward the giveaway. In the end, Maryland B won 200-195.

Let's assume for the sake of argument that my memory's a lot better than it actually is, because I'm interested in this as an intellectual experiment rather than quibbling about the outcome of a game in the middle bracket of a regular tournament. I recall that I took my questions in this particular packet, on average, a fair bit earlier than Andrew took his. Looking packet at the packet (Princeton's), my guess is that if powers had been assigned these questions, I'd have had three. I don't think Andrew (or Tricia, on the tossup she got), would've had more than one. Thus, if this game had powers, Temple probably would've won 210-205. Even, in fact, if we only outpowered Maryland by one, the first tiebreakers is lit., and considering that I got all four previous lit. tossups, I think it's a good bet I'd have gotten that too, and we probably would've won.

So this is my question (again, intended out of curiosity): in this type of situation, would people rather see the team that gets more tossups, but gets them relatively later in the question, win the game, or a team that gets fewer, but earlier on?

[Mechanical Beasts]

So this is my question (again, intended out of curiosity): in this type of situation, would people rather see the team that gets more tossups, but gets them relatively later in the question, win the game, or a team that gets fewer, but earlier on?

So in this situation, Maryland, by answering the questions later, gave you more of a chance to win the game, buzzing in on later clues in those questions. So had you had superior knowledge in those situations, you had a chance to demonstrate it, and you didn't, so they got more tossups. You dominated your categories with depth, sure, but I don't think the game is meant to be about hearing four words total from four lit tossups, and getting the win because each of them was a super-duper-trooper power: that's obviously not good QB.

Though your depth is more valuable in terms of building a team, they could cover two of the big three (science and history) and you covered one, so you lost. I think that's fair. At a single-subject side-tournament, I see powers as much more reasonable, or whatever sorts of constructs celebrate extraordinarily deep knowledge--hell, give a team that answers after the first clue an extra bonus part. But breadth is too important for a team that gets more points from tossups and bonuses to lose not because they didn't buzz before you did on science and history--because they did--but rather because they didn't buzz before some arbitrary standard.

[First Chairman]

On the impact of powers, this would be easy to assess impact on a rough level:

Look at the various statistics for a standard tournament (toss/bonus no rebounds) on a game-by-game basis and determine whether the outcome of a match would change if you did not reward powers. What is the percentage of games that would have a different outcome if you did this? I would be interested in a number, and hard-pressed to think the answer is above 10% depending on the field.

PACE's record on the highest number of powers converted in the stretch round part of the game still remains around 6 though I think we will discontinue this as a "record" in identifying games where it happens: I may record how many games that the number 6 or greater has been reached in a regulation game. But whether powers made a difference in the outcome of PACE games is a difference of no greater than 100 points, and I think on average is an advantage of around 30 points. We also have no negs so I hope we provide more incentive for teams to buzz in early if they sniff a significant hint.

[yoda4554]

But breadth is too important for a team that gets more points from tossups and bonuses to lose not because they didn't buzz before you did on science and history--because they did--but rather because they didn't buzz before some arbitrary standard.

The key word here is "arbitrary." There's an theory, often more true than it should be, that power placement is more or less arbitrary. But power placement should not be arbitrary, at least not by its very nature. I think that in most questions, there's a pretty definite place in the question--or at least an area of no more than a few words--where the material moves from "stuff that you need to have actually studied the material to know" to "stuff you might know if you've read a bunch of packets where this topic has come up." I also think that an editor competent enough to consistently make questions pyramidal can usually spot that point. That transition should be where the power mark is placed. I think most people agree that quiz bowl should try primarily to test something resembling "real" knowledge first, and qb knowledge second, and this seems like a fine, small way to do it, particularly when dealing with questions that often go down to the final sentence, which tend to elevate (by necessity) the importance of qb knowledge.

I know I never feel terribly good about winning games like the one we played against the Swarthmore frosh at the same Penn Bowl, on Minnesota A's packet. We were tied in that game through 20 questions before Temple won in the tiebreaker--but many of their tossups came from actually knowing something about the material being asked, where most of mine came from having developed long mental lists of things like the works of William Dean Howells (whom I've never read) by virtue of having played at the college level many years more than them. In games like that, I feel like they've demonstrated knowledge considerably superior to mine and that I've been outplayed; it seems strange that my team should end up with the win.

[Matt Weiner]

The key word here is "arbitrary." There's an theory, often more true than it should be, that power placement is more or less arbitrary. But power placement should not be arbitrary, at least not by its very nature. I think that in most questions, there's a pretty definite place in the question--or at least an area of no more than a few words--where the material moves from "stuff that you need to have actually studied the material to know" to "stuff you might know if you've read a bunch of packets where this topic has come up."

As people such as Jeff Hoppes have noted, this sort of grading becomes a more exact science as the difficulty level moves up. Who should know early clues for "Great Expectations" at a high school national championship? Probably about four fifths of the teams. Who will know early clues for the Mysore Wars at a collegiate national championship? Maybe 2 or 3 teams. The latter is more impressive and easier to spot as worthy of extra reward, but for various historical reasons, it's the former that we use powers in. At the A-level, or even the IS-level when elite teams are playing, the difference in difficulty between the power clues and all the other non-final clues can be quite "subtle" indeed.

As an aside, the attempt to meld powers with calculation is pretty much the definition of "arbitrary," though that's the price you pay for trying to shoehorn calculation tossups in at all.

[Mechanical Beasts]

As people such as Jeff Hoppes have noted, this sort of grading becomes a more exact science as the difficulty level moves up. Who should know early clues for "Great Expectations" at a high school national championship? Probably about 4/5 of the teams. Who will know early clues for the Mysore Wars at a collegiate national championship? Maybe 2 or 3 teams.

Right, and more distinguishes a "high-difficulty" clue about whatever that second thing is from "easy" clues about it than a "high-difficulty" clue about Great Expectations can be distinguished from a low difficulty one. In high school it's especially hard to judge what's power-worthy among early clues. Should the first of two early clues be for power? Should power last until a middle clue?

From NAQT's sample packet of IS 36A:

7. After settling in Hollywood in 1947, his interest in the Ramakrishna Mission produced mystical works
like The Doors of Perception. Sheryl Crow fans know he died in November 1963, which may have been 31
in the year of our (*) Ford. For 10 points—name this author who numbered years A.F. in his Brave New
World.
answer: Aldous (Leonard) Huxley

The clue structure goes: biofact, key term, famous work, pop culture reference for power. The clue that bridges power is the vague suggestion of an alternative calendar, (*), then Ford--indicating that unambiguously this is the alternative calendar from Brave New World, if anyone playing has read it or read about it. Then FTP comes--and then we get a recapitulation of the last clue and the book it came from.

I argue that putting power anywhere in that tossup is completely arbitrary.

Another example below:

15. Austria recently netted around one billion dollars due to voiding the value of (*) schillings that were not
converted to this new unit. For 10 points—name this new currency used by Austria and eleven other
European nations.
answer: the euro [YOO-roh]

Here, power can't come after schillings, because giving away the fact that the question is about currency and current events is too much to be worth fifteen points. This means that teams that wait for the first clue to be completed get ten points, and the teams that reflex buzz on Austria having done something recently that made some money get fifteen. The tossup is too easy to have a power mark.

But as Matt said, in college, if you get a tossup on Billiards at Half-Past Nine on something that Jochen the bellhop says (something of which I certainly know I'm incapable) without hearing the name Jochen... I'll FIFTEEN that. Perhaps even TWENTY it.

EDIT: other example.

[Deviant Insider]

I don't like either one of those examples above, and they may provide some evidence for Weiner's theory that power marks are difficult for lower level tournaments. Probably the most appropriate power mark for that round is:

Mama Lena and her children Beneatha and Walter Lee, Jr., are part of the (*) Younger family, who need to decide on what to do with the money from a life insurance policy. For 10 points—this describes what 1959 play by Lorraine Hansberry about deferred dreams?

answer: A Raisin in the Sun

In this case, you need to recognize the names of the characters to get 15, while having vaguely heard of the work gets you 10.

For the Huxley example, you get 15 if you read about LSD or listen to bad music (or possibly if you anticipate well), while you get 10 if you read literature.

For the Euro example, you only get 15 if you take an unreasonable risk, since Schilling is a good answer before the power mark.

By the way, we're referring to NAQT's sample packets here.

I think it would be possible to write a question with power for Great Expectations by referring to a minor character or incident before the power mark. My guess is that you could write an entire set of good questions at that difficulty with power marks, though I wouldn't say that the sample packet posted achieves that goal. (I generally like NAQT, but that particular round has some clues out of order and some transparency issues.)

The Goldfish Tournament would be a place where you could have powers systematically determined by making a rule that the first 5% (or whatever percentage) of teams that answer a given tossup get power in the matchups where they answered before their opponent. (That barn door is probably best left closed, however.)

[yoda4554]

From NAQT's sample packet of IS 36A:

7. After settling in Hollywood in 1947, his interest in the Ramakrishna Mission produced mystical works
like The Doors of Perception. Sheryl Crow fans know he died in November 1963, which may have been 31
in the year of our (*) Ford. For 10 points—name this author who numbered years A.F. in his Brave New
World.
answer: Aldous (Leonard) Huxley


15. Austria recently netted around one billion dollars due to voiding the value of (*) schillings that were not
converted to this new unit. For 10 points—name this new currency used by Austria and eleven other
European nations.
answer: the euro [YOO-roh]

I would argue that the problem here is not that powers must be arbitrarily placed in easier questions--the problem is that these are not terribly well-written questions (note that my comment above is contingent on the questions being good, pyramidal stuff, regardless of length). The first substantial clue in the first question is something that, if it's known by JV HSers, is known via Jim Morrison, the second is a clue about Sheryl Crow, and after that we have a pretty stock BNW clue; in the second, the question is entirely obvious by the time you know what is being asked for.

I think it's possible to write much better, easy pyramidal tossups of A-level length, so long as you have the discipline to select a quick hard clue, another quick middle one, and then (*) go toward a couple giveaway-type things. Here's a question I wrote for a power-less NAQT TV set in '06-'07, which I think is at least moderately decent for what it's supposed to do and is about the length/difficulty of an A-set:

Iraq's King Faisal I had previously ruled this other country for a few months in 1920. Taken over by the Baath Party in 1963, its Hama Massacre was ordered by Hafez al-Assad [AH-sahd], who sent its army in Lebanon in 1976. Name this Arab nation with capital Damascus.

answer: _Syria_ or _Syrian Arab Republic_

Now, I don't know the abilities of JV teams well enough to say whether power should be before Assad or Lebanon, but I'd argue that knowing Faisal ruled Syria, that Syria is the other country Ba'ath took over, or where the Hama Massacre took place constitutes serious Syria knowledge for that level, something that would only be acquired via real studying, while the other clues fall into a category of things that can be associated with Syria from list or general knowledge. Obviously, at higher levels some of those early clues become more stock, and you'd want to either move them back or add something harder.

Similarly, I've not read Great Expectations in a while, but as with any answer that's "easy" for the level of tournament at which it's used, I would think a good writer should fill the question's first half with slightly harder clues than one would for a lower-level tossup on that subject; for example, while you could write a pyramidal tossup that included Mr. Jaggers or Miss Havisham in the second sentence, it would make the question too easy for HSNCT and more appropriate for an IS set, even though listing supporting characters before power would be fine for a harder book. Certainly, there exist harder clues for Great Expectations, ones that could distinguish between people who've read the book multiple times (which for a book that major is not wholly unreasonable to imagine, even in HS) or could distinguish people who really got into the book from people who just slugged through.

Put another way: the fact that there are more stock clues from Great Expectations at the highest HS levels than for more obscure works doesn't mean that there are fewer hard clues to put before the power mark; I think that it's still possible to make that division between stock and non-stock, at least if you're aware of what current high school players are capable of converting.

[Mechanical Beasts]

Iraq's King Faisal I had previously ruled this other country for a few months in 1920. Taken over by the Baath Party in 1963, its Hama Massacre was ordered by Hafez al-Assad [AH-sahd], who sent its army in Lebanon in 1976. Name this Arab nation with capital Damascus.

answer: _Syria_ or _Syrian Arab Republic_

Now, I don't know the abilities of JV teams well enough to say whether power should be before Assad or Lebanon, but I'd argue that knowing Faisal ruled Syria, that Syria is the other country Ba'ath took over, or where the Hama Massacre took place constitutes serious Syria knowledge for that level, something that would only be acquired via real studying, while the other clues fall into a category of things that can be associated with Syria from list or general knowledge. Obviously, at higher levels some of those early clues become more stock, and you'd want to either move them back or add something harder.

I don't think it's good, though, to have your model for where you place your power mark "before this mark, we have serious knowledge; after it, we have list knowledge." Or "before it, we have serious Syria knowledge; after it, we have less serious knowledge." Either you have a tossup with a big jump in difficulty / non-specialist answerability, or you have a tossup (one I'd call good) with an even decline in difficulty, where placing a power mark is an arbitrary exercise--do you place the power mark where 0% of teams will know the answer (before the first word), 5%, 10%, etc.?

In a perfect tossup, where each clue opens up the field gradually, by another few percent, there's no reason to put the power mark before the clue that gives 25% of teams the answer than to put it before the clue that gives 20% of teams the answer besides your convention. Most power-marked tossups try to do what you were talking about with Syria in order to accommodate their existence: make a jump in difficulty. At SCT, I was involved in a number of buzzer races on the word or two after power. I think that suggests that there's a difficulty jump, and that's not good QB.

And if powers must be accommodated in a way that breeds bad QB, I don't like them.

[cvdwightw]

The fact that you were involved in a number of buzzer races a word or two after power at SCT implies that power placement was good. The "difficulty jump" stems from the length limit that makes it difficult to cram in all the hard, medium, and easy clues that your heart desires. In general, I am of the opinion that power marks should be placed in front of the first clue that the writer/editor considers "medium". Obviously power placement then depends on where the perceived division between "hard" and "medium" clues are, and not on where people will actually get the question. If you and your opponent are both buzzing off the first "medium" clue, then the buzzer race winner does not get power, plain and simple. If you've got 4 lines of hard clues, and then some medium clues and a giveaway, then I don't see a problem with putting a power mark somewhere in line 4 or 5; on the other hand, if you get to your medium clues quickly, then a later power mark is bad, regardless of how many people actually buzz off your first couple of medium clues.

I honestly think that this "perfect tossup", where each clue gradually opens up a larger percentage of the field, is an ideal that will maybe be achieved a couple of times a tournament. No question writer knows the exact field that will hear the question he's writing, and even if he did, no one knows the sum total of what that field knows. So we can't say "oh, 35% of the field knows this clue, 50% knows this clue, 75% knows this clue", because we don't know what the field knows. We make educated guesses based off where we get answers, where we see other people get answers, and where clues have shown up in previous packets. Putting clues that we assume a lot of people will know (names of main characters, famous commanders in battles, artist of certain paintings) in the beginning of tossups is bad, because we assume that the vast majority of the field will recognize those clues. That doesn't necessarily mean that the field will recognize them, just as putting in a supposedly hard clue at the beginning doesn't mean that people won't get it. We have limited empirical evidence to work with, and we try the best we can, but this doesn't mean that a question that is answerable by 75% of the field in the middle of the question is somehow promoting bad quizbowl. If it happens on a lot of questions, then we could have written a bunch of perfectly good questions that just happened to miss the target difficulty. That doesn't mean that they're all inherently bad questions.

[Mechanical Beasts]

The fact that you were involved in a number of buzzer races a word or two after power at SCT implies that power placement was good. The "difficulty jump" stems from the length limit that makes it difficult to cram in all the hard, medium, and easy clues that your heart desires.

Yes, as I said, good power placement necessitates that. However, if good power placement necessitates something I think is bad, and I don't think powers are a big benefit, I say no powers.

If you've got 4 lines of hard clues, and then some medium clues and a giveaway, then I don't see a problem with putting a power mark somewhere in line 4 or 5; on the other hand, if you get to your medium clues quickly, then a later power mark is bad, regardless of how many people actually buzz off your first couple of medium clues.

Right, there's no problem with that placement. But what about putting the power mark two-thirds of the way through line three? Maybe getting the tossup after a few hard clues, but before the last hard clue is what ought to be worth five extra points? Who can say? What's your calculus? Or do you believe that power should be getting it off a hard clue, not a medium or easy one? In that case, what defines a hard clue--and what justifies making that clue harder enough than the next that there will be the difficulty jump that you say indicates good placement?

I honestly think that this "perfect tossup", where each clue gradually opens up a larger percentage of the field, is an ideal that will maybe be achieved a couple of times a tournament.

Very true, but I think that powers jeopardize the potential for a perfect tossup, so why have them?

No question writer knows the exact field that will hear the question he's writing, and even if he did, no one knows the sum total of what that field knows. So we can't say "oh, 35% of the field knows this clue, 50% knows this clue, 75% knows this clue", because we don't know what the field knows. We make educated guesses based off where we get answers, where we see other people get answers, and where clues have shown up in previous packets.

I certainly don't pretend that when I talked about that sort of a progression I actually believed that for the most part question writers can do that. I meant to describe the shape and feel of the perfect tossup, in contrast to one that goes "5%, 10%, okay I'm not worth 15 points anymore, 35%, 45%,..."

]Putting clues that we assume a lot of people will know (names of main characters, famous commanders in battles, artist of certain paintings) in the beginning of tossups is bad, because we assume that the vast majority of the field will recognize those clues. That doesn't necessarily mean that the field will recognize them, just as putting in a supposedly hard clue at the beginning doesn't mean that people won't get it.

This is true; I don't see how this indicates that powers are good or contradicts what I've said.

We have limited empirical evidence to work with, and we try the best we can, but this doesn't mean that a question that is answerable by 75% of the field in the middle of the question is somehow promoting bad quizbowl.

No, but a question that, the clue before, was answerable by 35% of the field just so that a moderator can exclaim "FIFTEEN" and becomes far far easier ALL OF A SUDDEN is promoting bad quizbowl. It's closer to

Question: "Timber!"
Player: I do not know what they are going for here. Not at all.
Question: "Paul Bunyan felled the tree."
Player: This is giving me nothing, man.
Question: "FTP, spell 'timbre' "
Player: Which?
Question: "as it applies to music."
Player[s]: Buzz!

That went from 0% to a hose to a buzzer race very, very quickly, and is bad quiz bowl. Without the hose, and assuming the player is well-prepared for transparency, maybe 25% of a field could spell timbre after "Clarinets have different timbre from a dog whistle." I bet 90% could after "spell timbre." Obviously, this is extreme, but it's the difficulty jump that's bad quiz bowl, not being low-difficulty.

EDIT: for clarity.

[First Chairman]

The last few posts reminded me of the difference in what I consider is a "gimmick" in quiz bowl.

Variable-value tossups (like has) are a gimmick. It renders moot any standard reward for skill at a game and gives victory more or less to "chance" in close matches.

As for powers, I will note that PACE powers are transparent powers. Everyone knows that after the reader says "for ten points," the value changes to ten. It is an attempt to reduce any arbitrary factors in powermark placement (because I prefer there be some more objectivity to the meaning of getting the power bonus points) while also helping teams focus on the fact the tossup is about to end. For my opinion, a powermark should be placed whenever one is confident that a critical number of teams (or percentile) should be able to buzz in before that clue and get the right answer. However, any geographic sample may skew the probability a team could get that answer. So could unscrupulous negbeasts. I have wondered whether I wanted negs with powers (which makes some sense to evaluate "tossup aggressiveness", and there are times I'm still not sure. But for sure, no one has figured out the "percentile" yet. Plus for a national tournament, adjusting for the field for the power points is difficult given the range in ability within the field.

[Mechanical Beasts]

The last few posts reminded me of the difference in what I consider is a "gimmick" in quiz bowl.

Variable-value tossups (like has) are a gimmick. It renders moot any standard reward for skill at a game and gives victory more or less to "chance" in close matches.

As for powers, I will note that PACE powers are transparent powers. Everyone knows that after the reader says "for ten points," the value changes to ten. It is an attempt to reduce any arbitrary factors in powermark placement (because I prefer there be some more objectivity to the meaning of getting the power bonus points) while also helping teams focus on the fact the tossup is about to end. For my opinion, a powermark should be placed whenever one is confident that a critical number of teams (or percentile) should be able to buzz in before that clue and get the right answer. However, any geographic sample may skew the probability a team could get that answer. So could unscrupulous negbeasts. I have wondered whether I wanted negs with powers (which makes some sense to evaluate "tossup aggressiveness", and there are times I'm still not sure. But for sure, no one has figured out the "percentile" yet. Plus for a national tournament, adjusting for the field for the power points is difficult given the range in ability within the field.

I think your powers are good in terms of transparency; inaudible power marks have always bothered me, and I think what you have is a step in the right direction.

But is it appropriate to award ten more points for a buzz before the first word of the FTP clue? Either it is, and then there's a difficulty jump, or it isn't, because there's no difficulty jump. Or is your philosophy based on the idea that a certain proportion of teams should get an extra ten points on the tossup, in some sense as though they were given a thirty-five (or forty) point bonus, with a gimme part? I think that's valid except for the fact that the points assigned changes suddenly. If the number of points you get for a tossup declines from twenty to five constantly as it's read, maybe--but that's just impractical, as should be obvious.

[First Chairman]

But is it appropriate to award ten more points for a buzz before the first word of the FTP clue? Either it is, and then there's a difficulty jump, or it isn't, because there's no difficulty jump. Or is your philosophy based on the idea that a certain proportion of teams should get an extra ten points on the tossup, in some sense as though they were given a thirty-five (or forty) point bonus, with a gimme part? I think that's valid except for the fact that the points assigned changes suddenly. If the number of points you get for a tossup declines from twenty to five constantly as it's read, maybe--but that's just impractical, as should be obvious.

Allow me to sit on my rocking chair for a bit...

I do recollect on the reasons why we went with 20 points for power. For one thing, it is a remnant of "It's Academic" grab bag points where if you are familiar with the show the tossups at the end are worth 20 points or 30 points for a visual question. Also, Chip had questions in his Stump the Experts round that were worth 15 points for tossups at times. So the issue of selecting 20 points as the point value for our powers did not seem to me and Samer as anything so out of proportion in terms of changing the dynamic of the game. The other reason is that we also fielded a lot of "issues" from coaches at that time who complained (and still do) that tossup/bonus is "boring." If there is a blowout in the first-half, there's really no reason to continue playing. Now, that's worth a discussion, and why I recognize the format PACE has adopted may be construed by many as a "gimmick." However, there was a lot of discussion on why the stretch round was formatted the way it was over 10 years ago, and it mainly was to address two things: my issue with blind powers being a potential standard of quiz bowl play and valuing the more "random" nature of standard tossup/bonus that many more areas of the country were not accustomed to at the time. (Again, we are talking 12 years ago before NAQT and the only respected national out there was 's and Jeopardy.)

The 20-point transparent power allows each tossup/bonus cycle to be 50 points possible. 10 of these means you have 500 points possible in the last 10 questions of the game in a more rigorous format that did not exist in the "previous half/halves" for the game format... specifically related tossup/bonus and category quiz (where one chooses a one-part bonus). The symmetry of the points possible means that if one does not do so well on the even more gimmicky sections that your true depth of knowledge would be rewarded. Finally it also means that everyone has a chance to play the last half of the game (as witnessed by the consensus-pick for greatest quiz bowl game ever). Because bonuses can be stolen, it also means that the dynamic of the game requires teams to be on guard for a lot longer.

Finally, the transparent power point in the question I think offsets the fact we don't have negs. In a way it becomes an effective equivalent to a neg followed by a power by the opposing team. The intrigue though is that by having known powers, you give the opposing team a chance to know when to buzz in and take advantage, while blind powers I think can inappropriately tempt the opposing team to try to buzz in to beat the mystery power mark. So you actually work as a team "better" in this format (as I have seen people play it now). This means that actual game strategy and preparation are very important to close the deal and win the game. It also helps teams to remind the reader to check for power without disrupting the game as much as the NAQT version. (Believe me, there have been times when I realize too late to say FIFTEEEEEEN because I had to move on to a bonus, and no one reminded me until much later after I forgot where I stopped.)

I think it should force question-writers to reserve the more "listy" Level 1-reflex clues for after the FTP prompt. It reduces the horrible questions that I have played on (and admittedly written) that have no significant clues whatsoever prior to the FTP comment. But it also should remove the responsibility from writers and editors from having to make arbitrary judgment calls on what "they think" is hard enough and placing the power mark there. To directly answer your question, the structure is NOT based necessarily on what proportion of teams the writers should feel teams should buzz prior to that point. It is more a means to structure questions to enhance the value of strategy in the game. The responsibility to determine how "hard" the clues ought to be before the FTP prompt is the responsibility of the editors and writers (as it should be) but not at the expense of rewarding a team for appropriate knowledge or appropriate game play.

I think part of what you're thinking was also an influence for us: NTN counts down point values based on time to read the stem and answer choices, then the goofy clues. Based on when you lock your answer in, you get the points out of 1000. That would be a bit less practical to do that for a game format like this because that distracts the reader way too much. Simplify the reader's job, but make things more transparent regarding points... that certainly was another philosophical point for crafting the NSC format.

[yoda4554]

I don't think it's good, though, to have your model for where you place your power mark "before this mark, we have serious knowledge; after it, we have list knowledge." Or "before it, we have serious Syria knowledge; after it, we have less serious knowledge." Either you have a tossup with a big jump in difficulty / non-specialist answerability, or you have a tossup (one I'd call good) with an even decline in difficulty, where placing a power mark is an arbitrary exercise--do you place the power mark where 0% of teams will know the answer (before the first word), 5%, 10%, etc.?

But the thing is, knowledge is rarely a continuous spectrum over the domain of players. In fact, it can't be, because the domain of players is discrete; by necessity, there must be a sort of punctuated dropping off. For example, let's say there's an HS question on Daisy Miller. If you've read it, you know lots of things about it, including the basic structure of the scenes, various quotes, names of minor characters, etc. etc. But if you've not read it, you probably only know that it's the Henry James novella where the girl goes to Rome and gets sick at the Colosseum; if you've studied a bit more, maybe you know the name of the guy we follow around, Winterbourne.

Now let's say two teams are playing, both of whom have a player who's read the book: it's easy enough to write a question that rewards the player who remembers it better. It's even easier with one, and you can probably arrange it so that that player gets power. But what if no one's read it? Then we have to distinguish between who's studied more, and that's not very easy. Why? Because there are only a couple stock facts that a non-reader will know. It's pretty likely that both teams in that situation are going to know exactly the same amount of stuff about the book, and race to pick up the question there. It is thus necessary that there will be a drop-off at Winterbourne, power or not. And since this is the case, placing a power there seems a good way to acknowledge that.

[evilmonkey]

Mr. Letzler - this makes excellent sense, but I think that only works for literature.

Of course, you could make an analogous rule for history and such. Science... I'm not sure about, but then again, I don't know.

If you did power by this definition, it would almost necessitate the elimination of common link tossups.

[Mechanical Beasts]

But the thing is, knowledge is rarely a continuous spectrum over the domain of players. In fact, it can't be, because the domain of players is discrete; by necessity, there must be a sort of punctuated dropping off. For example, let's say there's an HS question on Daisy Miller. If you've read it, you know lots of things about it, including the basic structure of the scenes, various quotes, names of minor characters, etc. etc. But if you've not read it, you probably only know that it's the Henry James novella where the girl goes to Rome and gets sick at the Colosseum; if you've studied a bit more, maybe you know the name of the guy we follow around, Winterbourne.

Now let's say two teams are playing, both of whom have a player who's read the book: it's easy enough to write a question that rewards the player who remembers it better. It's even easier with one, and you can probably arrange it so that that player gets power. But what if no one's read it? Then we have to distinguish between who's studied more, and that's not very easy. Why? Because there are only a couple stock facts that a non-reader will know. It's pretty likely that both teams in that situation are going to know exactly the same amount of stuff about the book, and race to pick up the question there. It is thus necessary that there will be a drop-off at Winterbourne, power or not. And since this is the case, placing a power there seems a good way to acknowledge that.

I think it'd be very difficult to write a Daisy Miller tossup that I couldn't fraud for your idea of power by reading Masterplots and memorizing every character name and some memorable quotes. I surely believe that it's possible, however--but we can't make policy based on every Daisy Miller tossup being That One Perfect Tossup.

And really, I again don't see why power divides "serious knowledge" from "I just read a list" considering that I think as small a part of the tossup as possible should reward anything but serious knowledge; having a power there is a sort of concession.

[First Chairman]

Can I differentiate something here? Reading Masterplots or Cliffs notes or Sparknotes may not be the same as reading the book, but clearly it is better as a study guide than just memorizing lists. I have seldom been surprised at mid-level clues and some upper-level clues that are mentioned in those guides, but they go above and beyond the sentence-long synopses that most students "should know" as a basis of a particular work in the general canon.

The other thing is the audience. Remember that not all students get access to or appreciate Masterplots. Some of us readers don't get the opportunity to read all the works of literature. (Besides, poetry.... poetry!!!! that's another topic!)

Granted the other source of "fake knowledge" is going through old questions. But let's not be so willing to raise what would have earned power last week for a question on a work of lit to a mid-level clue next week. Not everything in academia gets to be accepted knowledge that quickly.

[yoda4554]

But the thing is, knowledge is rarely a continuous spectrum over the domain of players. In fact, it can't be, because the domain of players is discrete; by necessity, there must be a sort of punctuated dropping off. For example, let's say there's an HS question on Daisy Miller. If you've read it, you know lots of things about it, including the basic structure of the scenes, various quotes, names of minor characters, etc. etc. But if you've not read it, you probably only know that it's the Henry James novella where the girl goes to Rome and gets sick at the Colosseum; if you've studied a bit more, maybe you know the name of the guy we follow around, Winterbourne.

Now let's say two teams are playing, both of whom have a player who's read the book: it's easy enough to write a question that rewards the player who remembers it better. It's even easier with one, and you can probably arrange it so that that player gets power. But what if no one's read it? Then we have to distinguish between who's studied more, and that's not very easy. Why? Because there are only a couple stock facts that a non-reader will know. It's pretty likely that both teams in that situation are going to know exactly the same amount of stuff about the book, and race to pick up the question there. It is thus necessary that there will be a drop-off at Winterbourne, power or not. And since this is the case, placing a power there seems a good way to acknowledge that.

I think it'd be very difficult to write a Daisy Miller tossup that I couldn't fraud for your idea of power by reading Masterplots and memorizing every character name and some memorable quotes. I surely believe that it's possible, however--but we can't make policy based on every Daisy Miller tossup being That One Perfect Tossup.

And really, I again don't see why power divides "serious knowledge" from "I just read a list" considering that I think as small a part of the tossup as possible should reward anything but serious knowledge; having a power there is a sort of concession.

Yeah, that's because you're a pretty good college player--one who's willing to put in a bunch of time for the game--and this tossup, as I said at the start, is designed for a typical IS set. Look, this is obviously a simplification for the sake of argument; there's no real way to distinguish between someone who's really intensely studied secondary material and someone who's read the book, and I'm basically resigned to that; furthermore, there's no real way to distinguish a high-schooler who's read the book and a college player who's been immersed in lots and lots of clues. I'm talking about the standard stock clues at the level for which the question was written--those are, incidentally, all that's left of your Masterplot memorization a few months after the fact anyway. And I see no reason why the theory for this can't be extended to at least all literature questions, as well as many other arts questions.

And look, I'd love it if all questions could only test "real" knowledge, but you know as well as I that (particularly at the college level, but in HS too) that the first thing most good players do is learn how to fraud stuff. At a given level, in a appropriately-difficult game between two average teams, there will be many tossups on things that neither team has really legitimately studied; seriously, look at any packet you've played this year of normal college difficulty and ask yourself how many people in the game you played had read the book, listened to the piece, performed the reaction, etc.. Thus, to preserve conversion levels we tend to want, most questions transition from harder stuff to easier stuff, which means from stuff a specialist knows to stuff a person who's read a few packets and has little more than a general canon awareness knows. The question is, by necessity, more often picked up in the latter parts than the former. As real and dense as we want questions to be, those early clues tend, mostly, to make very fine distinctions between several players, and then have to roughly divide up a huge number more.

Anyway, I think we're just asserting ourselves back and forth at this point, so this'll be my last post on this. One more thing:

If you did power by this definition, it would almost necessitate the elimination of common link tossups

Why? Consider this common-link tossup on "Daisy" for IO Lit:

In Barnaby Rudge, one of Old John’s three cronies, Solomon, has this surname, and in Little Men it is the first name of Demi’s twin sister. A woman of this name is left behind when Eliza leaves her husband Eugene Gant in Look Homeward Angel, while another one with the surname Fraser brags that she contributes “To the School Fund in (*) Spoon River,” even though she is the village prostitute. It is the first of the four names of Mr. Zero’s secret love in Elmer Rice’s The Adding Machine and the first name of Moses Herzog’s first wife. It’s the name of Berenger’s girlfriend in Rhinoceros, and another character with this name goes with Giovanelli to the Colosseum and dies of malaria in a novel narrated by Winterbourne. FTP, identify this first name of Jay Gatsby’s love.
ANSWER: Daisy

Now, this was intended as a rather challenging tournament; it's a given that the depth of the stock clue well is much deeper. I'm not sure this is the best Daisy question ever written, but I think minus a quibble or two on some placement it's pretty good. I think Andrew Hart's logic in placing power was this: you'd either need to have done intense studying, or have read the not-really-widely-read-on-circuit books in question, to get the early clues, because they don't often come up or are minor in the work itself; beginning at around Spoon River, though, we hit works that either really expects a really top-level player to have read (Daisy Miller, Great Gatsby) or are very major characters in canonical works (Herzog, Rhinoceros) that one might simply have learned by hearing a lot of packets. Is it a hair arbitrary exactly where the power goes? Maybe, but that's the right basic area, within a clue or so.

[evilmonkey]

If you did power by this definition, it would almost necessitate the elimination of common link tossups

Why?

The definition I was referring to was the power definition that differentiated "serious knowlege" from "not-so-serious" In your tossup, if someone happened to read the books referenced in the opening line, they could FIFTEEN that question. That's not serious knowledge of Daisys, thats having read the book that the question writer decided to use in the lead-in. You don't study different Daisys.