Bargain or bust? The keys to avoiding free-agent mistakes.

Ah, we've reached the golden time of the baseball calendar. Next week, we'€™ll start to see playoff games. And if you'€™re the fan of a team that'€™s going to the playoffs, get ready for the heart-pounding roller-coaster cliche machine that is the playoffs. And if you'€™re not a fan of a playoff team, well ... I guess it'€™s time to get ready for the forthcoming free-agent season!

This year'€™s free-agent class is heavy on top-of-the-rotation arms; maybe your favorite team will decide it needs a shiny new ace at the top of the rotation. Free agency is fun because for the winter, you get to envision what your team would look like with David Price or Jason Heyward or even Alejandro De Aza! And that warm thought keeps you from thinking about the sticker shock (how many zeroes on that contract? And for how many years?) and the fact that there'€™s a very real chance that once that seven-year contract reaches year seven, that warm feeling won't be so warm anymore. Just ask Yankees fans about CC Sabathia.

When a free agent signs, there are usually two reactions that you can guarantee will come: "€œGreat pickup for the team. But they'€™ll be regretting those last two years."€ It'€™s something of a given that big-ticket free-agent contracts are a bit of an overpay, usually in the form of a couple of extra years in which the team signing him hopes that he will hold on to his form. Or at least some remnants of it. And once in a while, you end up with a player who actually does play at the same high level all throughout his contract. If it was your team that signed a player to one of those deals, you will be looking for the justification that this is the magic unicorn who can do it.

There are plenty of theories as to how to avoid the free-agent trap. Pitchers are bad bets because they get hurt so much. Older players are injury risks. Superstars are overpriced. One-year wonders are more likely to be a flash in the pan than a nascent spark of brilliance. But probably the most correct theory is that baseball players have a much shorter shelf life than we'€™d like to think. Consider that after the 2010 season, the top 10 players in baseball according to Baseball Reference's WAR were Josh Hamilton, Evan Longoria, Robinson Cano, Roy Halladay, Adrian Beltre, Albert Pujols, Ubaldo Jimenez(!!!), Brett Gardner, Felix Hernandez and Carl Crawford. I sure hope none of those guys are signed to big, long-term contracts.

If free agents in general are a hit-and-mostly-miss game, let'€™s at least see if there are types of free agents who, while not guaranteed, are a little more likely to beat the odds. Maybe there are bad free-agent bets and slightly less bad free-agent bets.

Warning! Gory Mathematical Details Ahead!

I used wins above replacement (WAR) data from Baseball Reference from 1999-2014. I looked at all players who posted a season of at least 2 wins, whether that player was actually eligible for free agency after that year or not. I did the same for starting pitchers. (Relievers are notoriously fickle because they don'€™t pitch very much, compared to starters, and to how much of a sample it generally takes before a pitcher'€™s performance is reasonably aligned with his talent level on a number of stats.)

A 2-win player is roughly an "€œaverage everyday player."€ He'€™s the kind of player who would easily hold down a regular spot on a team. He'€™s the kind of guy who would merit a news item when he signs a free-agent deal.

But what are the chances that he'€™ll keep that up next year? Or the year after? Or in five years? Below, I generated a graph for each player age. The top line represents all of the players who had a 2 win season at age X, and then whether in the next season they again reached at least 2 wins. The line below it is the percentage who had a 2-win season in the second year after, and then the third year, and so on.

The graph for hitters is first, then pitchers.

We see here that if a player, whether pitcher or batter, had a good season this year, he's around an even-money bet to do so again next year. However, we also see that it'€™s no guarantee. Rates hover in the 60 percent range for hitters in their 20s (who probably aren'€™t in the free agency basket to begin with at 25) to around 40 percent for guys on the wrong side of 30. For hitters, we see that older players become more and more risky, as we might intuitively expect. Sadly, for free-agent shoppers, hitters who are "€œgetting up there"€ are more likely to lose "€œit"€ and by the time most hitters reach an age where they are eligible to test the waters, they'€™ve started to enter that danger zone.

But what is going on with pitchers? If a pitcher survives to about age 32 and is still a reasonably good pitcher, his chances of continuing to survive as a good pitcher are actually better than when he was 28. There'€™s a little dip in the middle of that graph around a pitcher'€™s later 20s and early 30s where past performance isn'€™t so indicative of future success, and it'€™s worth asking why. We know that hitters tend to peak in their late 20s, and it'€™s reasonable to expect that pitchers also have a similar physical peak around then. When they come off of that peak, though --€“ perhaps losing a mile per hour of velocity here or there --€“ all hope is not lost. Pitching is about much more than just being able to throw the ball 95 mph. Pitching is a rich symphony of sequencing and location and movement.

There are a number of pitchers who have gotten by on limited "stuff"€ because they had a good command of these mental skills. Perhaps by 32, pitchers have naturally passed their physical prime and there's been a winnowing process. The ones who are even still around in our sample are around specifically because they can survive without their prime physical gifts. The ones who aren'€™t there anymore are those who will opted for retirement at 33. It's something to think about if your favorite team might be in the market for David Price, Johnny Cueto, or Jordan Zimmerman, all three of whom will be 30 next year.

There are other things that we can learn from these graphs as well. For one, we can learn that baseball players in general are risky assets to buy! We see that a significant percentage of guys who had a good year this year can'€™t be counted on to do the same next year. Teams can'€™t predict the future, but it makes sense to sign the players who are good now and assume that they'€™ll stay good. One of the frustrating things about free agency is that a lot of contracts go bad. And if roughly a third of hitters and half of pitchers don'€™t survive the first year continuing to be at least average regulars, there'€™s an awful lot of room for heartache. Research that I've done elsewhere shows that teams generally get about twice the value, in terms of wins per dollars spent on their cost-controlled, internally developed players, even after you factor in the costs that it takes to develop them. So the first lesson is not to shop in the free-agent bin.

But if you must, it's worth considering the other implication of these findings. Players who are hitting free agency in their late 20s and early 30s often request deals that are four and five years (sometimes longer) in length. It makes sense that they would want a contract that would give them financial security (and an eight-digit salary) for the next few years, especially if this is their first time venturing out on the free-agent market. Further down the line, players who are 34 and 35 are asked to be content with two-year deals, often at a reduced price compared to what a similar player might get at the same level of talent, but a younger age.

There'€™s a general idea that the older players are less likely to hold their value --€“ and these data show that to be true --€“ but when thinking strategically, it€'s important to ask "how much less likely?"€ Looking at 29-year-old vs. 34-year-old hitters as a bellwether, it seems that the risk for a contract year going "€œbad"€ (i.e., producing less than two wins of value) is about 10 percent greater. But even for 29-year-olds, the chances that by the fifth year of a contract, you'€™ll have to swallow a bad year are over 70 percent. You're probably going to get stuck with a dud year at some point. It might just be a matter of deciding when you want that year to be.

For pitchers, we actually see that the 34-year-old might actually be a better bet. Older free agents come with a little more risk, but also often come with the benefit of not asking to be tied into a contract for as long. If your favorite team signs a 34-year-old to a two-year deal in the offseason, it might not be a bad signing. Sure, it might go sour, but anything can go sour.

Oh Wait, They Have Names!

It'€™s important to remember that, when dealing in analyses like these, we'€™re talking about the "€œaverage"€ player. All that the analysis knows here is a player'€™s age and that he was a better-than-two-win player last year. David Price is a 30-year-old pitcher, but teams should take more than those couple pieces of information into account when deciding whether he'€™d be a good investment or not. If there's a lesson to be learned from all of this, I think it'€™s that fans (and perhaps teams) tend to overestimate the amount of risk involved in signing older free agents, particularly free-agent pitchers. I think the reason has to do with how humans process information.

There will be an older free agent or two who simply doesn€'t pan out next season. (If I knew which one was which, I would be a millionaire.) We'€™ll note that "age caught up with him and moved on." It'€™s a quick, ready-made explanation. But there will also be a couple of 30-year-old free agents who don'€™t pan out, either (I'm looking at you, Hanley Ramirez.) Sometimes, we can see the cause (injury, lost another mile on the fastball, etc.) But sometimes a guy like Ramirez, still in his physical prime, just ... breaks down.

There'€™s been plenty of analysis of Hanley'€™s problems this year, but most of those pieces end with an exasperated shrug and some form of "€œBut this shouldn'€™t have happened"€ plus a glimmer of hope that maybe this is just some sort of one-year hiccup. In baseball, we have trouble believing that there are things other than age (and freak injury) that can cause a player to collapse (and that bodies can show signs of aging before the mid-30s). Sometimes it is an injury, but sometimes it'€™s a decline in a subtle, but very important skill or because he reached a point in his life where he just stopped caring.

Language is a very powerful thing. We have a name for when a player loses his battle with Father Time, but no real name for what it'€™s called when he just ... loses it. It'€™s easy to be skittish of something that you have a name for. But when that threat is a big amorphous blob that we can'€™t quite define, it's harder to be threatened by it. Yes, older players are risky, but when assessing risk, you always have to ask the question "compared to what other alternatives?"€ and then ask whether that risk is properly priced in the market. If it'€™s not, then like all mis-pricings, there's room to take advantage.

So, if your favorite team signs a guy who'€™s had a good career, but seems a little over the hill and you'€™re mad because there was a much younger, shinier model on the shelf, your team might have actually been making the smart play.