Over the course of an individual season, players can benefit from a high batting average on balls in play (BABIP), thereby getting "lucky" over the course of the season.

A ground ball can be just out of a fielder's reach, or a fly ball can be "perfectly placed" where an outfielder just can't quite get it. But over time, this luck will even out, and even the lucky ones will regress back towards the mean. 

Players with high BABIP compared to the league average (.300) and their other hitting metrics -- hard hit ball rate and line drive rate -- will eventually regress to mean and stop getting lucky over the course of seasons. A player who has a .382 BABIP one year can be expected to regress back towards the average the next. And at the same time, a player who has .238 BABIP will likely improve the next season, as he gets luckier himself. 

Of course, there are other factors that come into play.

Bryce Harper and other hard-hitting sluggers will continue to have plus-.350 BABIPs year after year, but there is a strong correlation between a hitter's BABIP and their line drive percentage (LD%).

numberFire's Jim Sannes looked at the correlation last year, finding upwards of a .45 correlation between the two. And again this year there is a significant correlation, .48 to be exact, so using the best-fit line, we can find a hitters' expected BABIP and in turn figure out who should improve at the plate and who will regress. 

Using the best fit line BABIP = 0.572(LD%) + 0.1873, I calculated every qualified hitter's expected BABIP and took the difference between their expected and actual BABIP. Those with highly negative expected-actual BABIPs should regress, while those with positive expected-actual BABIPs should improve next season -- but we'll get to them in the next article. You can find the entire data set, via FanGraphs, here. 

So here are five players who should regress next season based on their expected BABIP

1. Dee Gordon, 2B, Miami Marlins

BABIP: .383 | Line-Drive Percentage: 21.5% | Expected BABIP: .310

Dee Gordon was a monster for the Marlins this season, hitting .333/.359/.418 with a .337 wOBA and 113 wRC+ and posting a 4.6 fWAR. But with a .383 BABIP, 73 points higher than expected, tied for the highest among qualified players, Gordon is expected to regress next season even with his speed on the basepaths. Just 21.5 percent of Gordon's hits were line drives -- which go for hits about 68.5 percent of the time -- compared to a ground ball percentage that's second in the Majors at 59.8 percent. And even with Gordon's speed, batters hit, on average, just .239 on ground balls, so with that high of a ground ball rate, regression is bound to happen. 

Further with a low hard hit rate of 17.6 percent, good for second worst among qualified hitters in the Majors, Gordon absolutely will see some regression. Batters typically hit around .700 on hard hit balls, compared to just .400 on medium hit balls and .150 on soft hit balls. 

Projecting a substantial regression, Steamer projections for 2016 predict Gordon's BABIP to fall to .332, and in turn a regression to .284/.325/.368 with a 303 wOBA, 90 wRC+ and 2.0 fWAR, not too shabby, but not close to his career-highs this season. 

2. Odubel Herrera, OF, Philadelphia Phillies

BABIP: .387 | Line-Drive Percentage: 23.5% | Expected BABIP: .322

Odubel Herrera burst onto the scene this year after being selected by the Phillies in the Rule 5 draft from the Texas Rangers. Herrera played well enough to earn a roster spot and soon took over in center for Philadelphia when Ben Revere was traded. The left-hander hit .297/.344/.418 with a .333 wOBA, 110 wRC+ and 3.9 fWAR, but unfortunately for the Phillies, a big regression is in Herrera's future. Herrera had an astounding .387 BABIP, 65 points higher than expected, and with a just-above average line drive percentage, Herrera will have trouble sustaining it in 2016. 

And similar to Gordon's, Herrera's below average hard hit rate at 26.3 percent along with an above average medium hit rate, 54 percent, which isn't a good thing in this case, will further regress him downward. Steamer projects Herrera to hit .266/.313/.365 with a .298 wOBA, 86 wRC+ and 1.5 fWAR as his BABIP drops a full 60 points to .327. And while as a young player Herrera may be able ward off some of the regression by improving at the plate, it's still coming.

3. Francisco Cervelli, C, Pittsburgh Pirates

BABIP: .359 | Line-Drive Percentage: 21.0% | Expected BABIP: .307

After bouncing around the New York Yankees' system the past 8 years or so, Francisco Cervelli broke out last year, and continued his performance this year, hitting .295/.370/.401 with a .341 wOBA, 119 wRC+ and 3.8 fWAR. But at some point, Cervelli has to come back down to earth, and 2016 just may be that year. After posting BABIPs of .408 and .359 the past two years, Cervelli is due for some serious regression. His expected BABIP this season was 52 points lower than his actual BABIP, and with just an average line drive percentage and hard hit percentage, Cervelli should regress next season.

Steamer projections peg a small regression for Cervelli's BABIP, dropping to .324, along with a slash of .268/.338/.376 with a .315 wOBA, 102 wRC+ and 2.1 fWAR, all down from last season. 

4. Nick Markakis, OF, Atlanta Braves

BABIP: .338 | Line-Drive Percentage: 20.8% | Expected BABIP: .306

After posting his highest BABIP since 2008, Nick Markakis showed some signs of eternal youth at the age of 31 as he hit .296/.370/.376 with a .327 wOBA, 107 wRC+ and 1.6 fWAR. But unfortunately for Braves fans, with a below-average line drive and hard hit percentage, Markakis is due for a regression. Markakis' expected BABIP was just over .300 at .306, and it should definitely drop next season.

Steamer projections peg his BABIP at .298, signaling a substantial regression from his .338 BABIP and 107 wRC+. Steamer projects Markakis to hit .269/.340/.369 with a .310 wOBA, 96 wRC+ and just 0.6 fWAR, down a full WAR from last season. 

5. Ian Desmond, SS, Washington Nationals

BABIP: .307 | Line-Drive Percentage: 15.6% | Expected BABIP: .277

MLB teams be warned. As a 30-year-old free agent, Ian Desmond will be looking for a sizeable contract, but with further regression possibly in his future, teams might not want to go about giving him tens of millions over the next few seasons. Even with a low BABIP of .307, the model still expected an even lower BABIP, due to a poor line drive percentage. With Desmond's below average line drive and hard hit rate, he could be looking at further regression if he can't bring either of those up next season. Desmond's line-drive percentage was a career low, but still for his career, Desmond is below average in terms of line drives. 

And while Steamer is a bit more optimistic than the model, predicting an increase in Desmond's BABIP to .316, there's still not a whole lot of improvement Desmond can make. After hitting .233/.290/.384 with a .294 wOBA, 83 wRC+ and 1.7 fWAR last season, Steamer projects modest improvements: .246/.302/.401 with a .306 wOBA, 91 wRC+ and 1.9 fWAR. But with his low line drive percentage and a high strikeout rate -- 29.2 percent last season -- Desmond could be in for a startling regression.