If hypothetical odds are to be believed – and if we can’t believe in hypothetical odds, what do we have left? – the second half of Georgia Tech’s season could be a rough ride. But if you believe in a computer model that picks about 77 percent of its games accurately, start making reservations for Charlotte.
For a story that ran in Wednesday's paper and myajc , I asked Jay Kornegay, the director of the Westgate Las Vegas SuperBook, to provide odds on the remainder of Tech’s games. Kornegay graciously complied, offering the caveat that the lines are guesses, tough to make and are bound to change dramatically when the actual games approach. That said, if the games were to follow the line (Tech winning games in which it is favored, losing a game in which it is the underdog), the Jackets would finish 8-4, beating Duke and North Carolina, losing to Pitt, beating Virginia and then losing to N.C. State, Clemson and Georgia.
A 4.5-favorite against Duke.
A three-point favorite at North Carolina.
A two-point underdog at Pittsburgh.
A three-point favorite at home against Virginia.
A one-point underdog at N.C. State.
An 8.5-point underdog at home against Clemson.
A 16-point underdog at Georgia.
It’s interesting to note that four of the games have a line of three points or fewer. Georgia and Clemson received considerable respect from Kornegay (or, conversely, Tech didn't receive much). Last year, the Bulldogs, without Aaron Murray and playing at Tech, were a 2.5-point favorite last year against the Jackets. Clemson was a 10.5-point favorite against the Jackets last year at Death Valley.
A 5-3 record would put Tech in the picture for the Coastal title, although it’d be tenuous. Probably a little more guess than I’m quite ready for at this point.
The N.C. State and Pittsburgh lines seem off, particularly the latter as the Panthers seem to be reeling, but I'm not the expert. Obviously, home-field advantage is taken into account, as it is for Tech home games.
Another way of looking at it from Atomic Football, which gives algorithm-based predictions with win probability percentage.
Duke, 32-28, W, 62.7 percent.
North Carolina, 37-35, W, 54.9 percent
Pittsburgh, 29-28, W, 52.4 percent
Virginia, 33-25, W, 71.7 percent
N.C. State, 33-25, W, 71.7 percent
Clemson, 38-26, L, 29.0 percent
Georgia, 40-28, L, 20.1 percent
The Atomic Football model gives Tech a 10-2 record, including 7-1 in the ACC, which would get the Jackets to the ACC title game. The probability of winning out is .5 percent.
If you’re wondering, Atomic Football has picked 73.3 percent of games correctly this season. It was 76.9 percent last year. (This is interesting – I got the accuracy numbers from a website that tracks how well a number of different models pick games. Last year, there were 58 different models, including the betting line, that ended up between 78 percent and 73 percent, and it’s relatively close to that year after year. It’s like it’s impossible to accurately predict 79 percent of games straight up. When that happens, maybe computers take over the world.)
Anyway, food for thought.
Most recent 20