Snap quick!: TV photos require fast work

Q. You love that fetching face on TV, so you fetch your camera to snap a picture. Will her loveliness be preserved for your photo album? ­S. Mann

 A. Better check the camera's exposure time first, since at 1/250th second– a typical indoor setting– it'll be four times faster than the top-to-bottom scan ("refresh rate") of the TV screen, says UVA physics professor Louis A. Bloomfield in How Things Work: The Physics of Everyday Life. In that case, you'll capture only maybe her scalp, or chin, or horizontal mask-strip of the eyes, a quarter-screen high. "While the TV's scanning process easily fools your eye into seeing a complete image, a camera records exactly what it sees."

Slower film speeds and longer camera shutter times will more fully assuage your facial beauty fetish.

Q. Hospital prankster slips a cadaver presumably "sleeping"– into the MRI unit for imaging of the brain or torso. Would the technicians necessarily notice? ­Dr. Kildare

 A. Assume a "fresh" body was found, on its way to the morgue, no missing parts, no strange-looking fluid pockets, no chemical preservatives pumped in yet that might make some tissues show up abnormally bright or not at all, poses University of Wisconsin medical physicist emeritus James A. Sorenson.

Now the differences would be less obvious. After cell death, water gets redistributed, causing changes in tissue contrasts. For instance, the contrast between white matter and gray matter of the brain would be suppressed after cell death and deterioration. "An experienced technician might notice these changes. But they would be subtle."

"Dead or alive" would be easier to tell on side-by-side images, where any body movements or pulsatile flow might show up as "ghost images." Here the moving structures appear displaced from their true body location.

It is virtually impossible to remain still for the entire MRI scan, and of course the heart and blood will move regardless, says Sorenson. So researchers have worked hard to suppress ghosts, since they interfere with image interpretation. "Still, I think an experienced observer on careful inspection wouldn't be fooled, except perhaps on images with no prominent arteries, such as the hands and feet. In a vascular scan designed to show only the blood vessels, there would be nothing to see, because there would be no moving blood to create an image. That would be a 'dead giveaway.'"

Q. There's a hot little betting system that'll let you go to a casino roulette table with $1000 and come home a $60 winner nine times out of 10. Interested? Just don't hold us responsible for that 10th visit. ­Lemuel Gullible

A. You mark down 1-2-3, because you'll be winning your money $6 at a time. Start by betting $4– always the lowest + highest numbers– on either red or black. There are 18 reds on the wheel, 18 blacks, and two for the house, giving you 47 percent chance (18/38) of winning your $4. If you win, cross off the 1 and 3 and bet $2 (the remaining number). If you win again, you've crossed off all numbers and netted your first $6, so you write down 1-2-3 and start over.

If you lose your $4 bet, write down 4– always write down amounts lost add the 1 and 4, and bet $5. If you win the $5, cross off the 1 and 4, then bet $5 (the 2 + 3) again. If you win on this, all the numbers are crossed off and you're on to a fresh 1-2-3 set. Each loss pushes your bets a little higher until finally when you win, you cover your losses, while netting $6.

There's a catch, of course: You could run out of money before crossing off all the numbers!

A computer simulation by John McGervey (Probabilities in Everyday Life) beat the house an average of 99 times out of 100 on these 1-2-3 sequences, each winning $6. Winning 10 sequences got to the goal of $60 on a given "visit," and on 90 sequences (9 visits) the computer won $540. But on every 10th visit, on average, the 1-in-100 loss caught up with the computer, scattering its $1,000 to the winds.

There are scores of similar schemes– all useless. Even after winning on 9/10 visits, the computer ended up $460 in the hole. Any random scheme could have done just "as well"!

Send Strange questions to brothers Bill and Rich at strangetrue@compuserve.com.