It’s a tough question in any sport. How do you put a single number on a player’s value, whether that number is in terms of runs, points, goals, wins, or dollars? I recently took a look at the free agent safeties with the goal of establishing a market value for a player based on his expected performance.

A first-round pick in 2007, Landry has exceeded expectations. In his first three seasons he averaged 2.0 +Expected Points Added (+EPA) per game, which put him squarely in the upper echelon of all safeties. (+EPA puts a point value on a defender’s direct contribution to the game outcome, measuring the impact of tackles, sacks, turnovers, passes defended, and other plays.)  Then Landry had his breakout year in 2010, notching 4.4 +EPA per game, which was second in the league, sandwiched between likely Hall of Famers Ed Reed and Troy Polamalu. His other stats, like +Win Probability Added, Success Count, and Tackle Factor were also career bests.

In his eight games this season, Landry regressed from his 2010 peak, in part no doubt due to his injuries. But he was still a force in the Redskins secondary, averaging 2.8 +EPA per game, good for 8th in the league for all safeties. By comparison, O.J. Atogwe averaged 1.3 +EPA per game and Reed Doughty averaged 1.4 +EPA per game in 2011. The bottom line is that since his second season Landry has averaged 2.9 +EPA per game, which is a reasonable expectation of future performance.

So what does that mean in terms of salary? To answer that question, we can evaluate the relationship between safety +EPA and salary. I chose +EPA per game as the measure of performance because it fit exceptionally well with actual salary values. Due to the complexity of bonuses and incentives, I chose “cap hit” as the most relevant measure of player cost.

The bottom line is that top safeties tend to cost slightly less than $2 million dollars per point of +EPA. The correlation between cap hit and +EPA is nearly 0.7, which is remarkably strong given the natural uncertainty between expected and actual performance. What this suggests is that advanced stats such as +EPA produce a reasonably good estimate of player value.

Landry would be worth about $6 million per year in cap hit. For comparison, Troy Polamalu recently signed a four-year, $36 million deal. Polamalu averaged 3.2 +EPA per game since his second season, which would make him worth closer to $6.4 million per year than $9 million. The difference might be explained by the timing of Polamalu’s signing, which was very recent. Salaries inflate as the league’s revenues steadily climb.

Landry is four years younger than Polamalu, but has a history of fragility, including his lingering Achilles injury, and hasn’t yet proven to have the on-field consistency that the Steelers star has shown. Landry will likely be offered a five-year contract, and I would estimate he would be worth $35 million over that timeframe, or about $7 million per year depending on how high the team caps are raised for 2012. The upshot is that he’d typically be worth about 6 percent of whatever the cap is.

Would that a good deal for the Redskins? I would say so. They have the cap space, especially considering Landry already had a very lucrative rookie deal. And without Landry, safety becomes a weak spot for the Redskins. Atogwe and Reed have played as well this season, but they are both a notch or two below Landry. Neither one has the ability to get into the backfield the way Landry can. Drafting a replacement doesn’t appear to be an option because there are no top shelf safeties projected to come out in the next draft. Plus, the team has pressing needs at several positions and can’t afford to spend a high pick on a position that has traditionally been a strong suit. The free agent market at safety will have a few notable players available, including Jim Leonard, Michael Griffin, Quintin Demps, and Jordan Babineaux, but none are at Landry’s level. The lack of substitutes will likely push Landry’s market value up into the $8 million range, and that price would still offer value to the Redskins.

Brian Burke is the creator of Advanced NFL Stats, a Web site about football, statistics and game theory.