A Thousand Times No

Once the puzzle grids are completed on the competition floor, they move on to their next life in the hands of the judges. Puzzles are graded the old-fashioned way: by hand, one square at a time. The vagaries of hasty handwriting are such that it the job of deciphering the contestant’s intent is best left to a fellow human being (preferably not one named “Chad”).

The conference rooms in the Brooklyn Bridge Marriott are named after famous Brooklyn residents, and judges were assembled together in “Walt Whitman,” down a hallway and around the corner from the main ballroom. Gathered around a series of circular tables, we feverishly worked to mark up each puzzle grid with the information required to assign it a score.

Each judge uses a colored, felt-tip marker to write upon a puzzle page. First, a judge looks at the back of the page for the contestant number, written there by each contestant before the round begins. (Contestants are also asked to write their names to help resolve any numbering issues.) The judge then rewrites the contestant number on the front of the puzzle, accompanied by the judge’s own initials. Then the “grading” begins.

The judges are given a sheet with the correct answers shown in a puzzle grid. Using this answer sheet, a judge looks through a competitor’s puzzle and circles every incorrect letter and every blank square. Circles are used rather than X’s or slashes so that the letters can still be reviewed in the event of a scoring discrepancy. When a judge is finished checking an entire puzzle, two numbers are written at the top of each sheet: The number of incorrect letters, and the number of incorrect words.

Why two numbers? Knowing the number of letters missed won’t simply tell you how many words were incorrect. If one letter is missed in the upper left corner of the puzzle, and another letter is missed in the lower right corner of the puzzle, each letter would result in two words being incorrect (one down, one across). 2 letters, 4 words missed. However, if you missed two letters immediately to the right and left of each other in the upper left corner of the puzzle, you would have only missed one word across. 2 letters, 3 words missed.

Where there’s a complete bloodbath — like Puzzle #5, David J. Kahn’s “Up-scale,” which Will Shortz dubbed “The Bastard” — it’s not uncommon to have puzzle after puzzle come in with far less than half the squares filled correctly. In that case, judges are asked to denote how many letters and words were correct, rather than incorrect. The inversion is noted with a small minus sign placed in front of the letter and word scores: -43 letters, -17 words means that 43 letters and 17 words were correct in the puzzle. Puzzle #5 was one big House of Minus.

Puzzle grids with an abundance of blank squares take more time to grade because they involve checking square-by-square, rather than word-by-word. From a judge’s perspective, a brain-smarting puzzle like this year’s Puzzle #5 gives new meaning to the phrase “This hurts me more than it hurts you.” (A chorus of groans went up in Walt Whitman when the last few batches of Puzzle #5 were brought into the room for grading.)

Judges pile their graded puzzle grids in the center of each circular table. At regular intervals, an official picks up all the graded grids from the Walt Whitman room and carries them down the hall to the Jackie Gleason room where the scoring tabulations take place. In the scoring room, graded puzzle grids are sorted according to contestant number. Then the graded puzzles are passed on to one of three tables, where judges with laptop computers work to enter the grading information into a database that tabulates and records each contestant’s score.

Contestants finish each puzzle in widely varying amounts of time, which throws a twist into the flow of puzzle sheets. The rolling arrival of puzzle grids — from floor referees to the grading stations, and from the grading stations to the scoring room — means reams of paper are in constant motion throughout the tournament. The first few completed puzzles from the tournament’s top tier of competitors will ebb in early, there’s often a pause, and then several minutes later a second wave of puzzles begins to come in. After a few more minutes, the pace tails off again, and the puzzles generally come in at a steady rate until the time limit is reached and all remaining puzzles are collected.

Meanwhile, in the grading room, the grading of the puzzles proceeds at a steady pace. The first puzzles to arrive tend to be the easiest to grade, containing few or no mistakes. As the later finishes come in, the rate of error increases; with the more difficult puzzles, it’s surprising to reach “the bottom of the stack” and pull out a puzzle grid that’s completely filled.

During the push to finish the puzzle grading for each round, officials who have been performing other duties such as refereeing and sorting will sit down to join the work at the grading stations. When an official starts grading puzzles late in the round, the other judges will often cherry-pick a few nearly completed puzzles from the unmarked stack and hand them to the new grader, easing them into working with an unfamiliar puzzle solution.

Judges try to give the contestants the benefit of the doubt. “Is that a ‘C’ or an ‘L’?” we’ll ask each other, calling a huddle to peer at a letter together. “That’s a ‘T’, not an ‘I’, you can see the indent in the paper in the black square where they crossed it.” Credit is given for vaguely alphabetic glyphs and mostly-erased letters. With blank squares, however, the answer is always clear; I heard sympathetic cries of disappointment from judges as they graded early finishers whose perfect work was marred by a lone oversight.

Of course, no account of puzzle grading would be complete without mention of inadvertently humorous answers. Vanna White has spoken of how a “Wheel of Fortune” contestant once guessed “Done with one hand” as a solution, rather than the correct answer, “Gone with the Wind”. Although the ACPT contestants are more of a “Jeopardy!” than a “Wheel” crowd, they still find plenty of unusual ways to fill in the blanks. Among the answers we saw:

• “It’s between Quebec and Sierra in a phonetic alphabet”: RAMBO for ROMEO
• “Extreme liberal”: (right arrow)-IST for (left arrow)-IST
• “Aristotle output”: THE OPERA for THEOREMS
• “Stage name of U2 guitarist David Evans”: THE FOGY for THE EDGE

Puzzle #2, Mike Shenk’s “Change of Venue,” involved not only solving the puzzle, but building a word ladder next to the puzzle, starting with the word “VENUE”. Since the word ladder itself was used to give clues, there were a many stabs in the dark:
(just for grins, I’ll leave it to you to fill in the ladder based on the answers)

• Clue: “2nd word in the ladder”: DONE GODDESS for LOVE GODDESS
• Clue: “3rd word in the ladder”: BALLS OF FIRE for BILLS OF FARE
• Clue: “4th word in the ladder”: SEDUCED BY for REDUCED BY
• Clue: “5th word in the ladder”: COAL TITS for COAL PITS
• Clue: “6th word in the ladder”: WAS A MEAL for HAS A MEAL
• Clue: “7th word in the ladder”: SEEDY BUNS for SEEDY BARS
• Clue: “8th word in the ladder”: PAR OPPOSER for WAR OPPOSER
• Clue: “9th word in the ladder”: ENGLISH HORN for ENGLISH PORT

This all led to the last word in the ladder, which was also the final Across clue: “Person responsible” — MOVER.

“Don’t worry,” one of the judges had warned during grading. “By the time you get home and take a look at these puzzles, you still might struggle with the toughest ones, even though you’ve been staring the answers in the face all day.” Of course, thanks to “The Bastard,” the warning proved to be prophetic — one week later, I still haven’t managed to work my way through Puzzle #5.

As one of the longest Across answers in the tournament would say: “Don’t that beat all?”