Everyone gets the basics of life under the salary cap by now. You need to get the best team you can assemble for the total money you spend on said team. To do that, you look for players who will produce more than what they'll cost to acquire.

There are two categories of players who meet that description (which are not mutually exclusive): ones whose cost is suppressed by the league's CBA structure (players on entry-level contracts and restricted free agents), and ones who are undervalued by their own and other general managers. I want to look at one group in that second category, and show why, how and when the Leafs can get value there.

I want to sift through the remains of broken warriors.

What?

Injured guys. Guys who have been injured. Come on, let me be poetic about this.

1. Executive Fear

The worst thing to be called in hockey isn't an epithet, or an insult, or even "enigmatic." It's "injury-prone." That shrugs off almost any value a player might provide, because that value becomes tentative and hedged with only-ifs. And there's good reason to beware what injuries can do to players. They can cut down prospects at the root, stop primes early, and make the end of careers sadder and less effective. This fear is deep and it is by no means baseless.

There's another fear that impacts injured players: the front-office fear of looking like an idiot. NHL general managers operate in a job where, in the simplest terms, 97% of them fail each season. True, not every non-Cup season is considered a failure, but that's only when GMs are perceived as having advanced toward being the 3% guy at some point in the next few years. Put another way, eighteen of the thirty current NHL GMs have been hired since the last lockout, which ended at the start of 2013. As everyone knows, there's not a lot of security in this business.

In an industry where failure is much more likely than success, and there's immense turnover as a consequence, GMs—consciously or not—have to prepare to position themselves for when failure comes. It's a lot easier to explain away careful, conservative decision-making than swinging for the fences and striking out. You know how we all laugh at Jarmo and the Jackets for signing Nathan Horton? To an uninsured contract he's only played 35 games of (out of a potential 246)? That they wound up trading for David Clarkson to try and escape from? Yeah. Jarmo has survived the mistake—possibly because it was made so early in his tenure, and the Jackets are a budget team—but it was a disastrous-looking move.

So when you sign an injury-prone player, and he gets injured again, it looks like the ultimate unforced error. If you had a professional sword of Damocles over you, you too would probably be wary of decisions that could turn out to look bad, independent of whether you thought they actually were bad.

And finally, there's something elementally frightening to human nature about disabling injury. The list of hockey players devastated by injury goes from Bobby Orr to Mike Bossy to Marc Savard and on and on. Does anyone want to spend time thinking about Orr limping to the end of a Hall-of-Fame career, or the best scoring winger of the 80s being finished at age thirty, or Savard talking about how he struggles to stand bright rooms? No, because those things are wincingly sad and unpleasant to contemplate. I honestly believe this aversion plays a real, if unconscious, role in NHL decision-making.

So in short, there might be a class of players that NHL GMs are afraid to believe in. Let's look into that.

2. Injuries, Recoveries, Contracts

It's time for everyone's favourite segment: an unwieldy spreadsheet!

After some thought and some wise advice from Katya, I've decided to put my spreadsheet in a Google doc, so people can play around with it, and to put the full explanation of all of the things I was doing in a document there. If you have questions, submissions, objections, concerns, queries, approbation, or misgivings, go check it out and you'll at least understand the twisted thinking behind my little experiment—it lays out why I included and excluded who I did, why I tracked what I did, why I chose the year I did, and so on. If you don't care about any of that stuff, look below the link, and you'll find my conclusions.

Short explanation: I looked at all of the significant players who had at least 20 man-games lost to injury in 2009-10, then looked at their performance in three metrics in 2008-09 vs their performance in 2010-11 (i.e. the year before and the year after they had a seriously injury-marred year.) I removed players who seemed to me to be minor-league/fringe players before the injury or who did not provide at least twenty games of data in 2008-09, plus the odd case of Kurt Sauer. The idea was to track, roughly, the rehab process of badly-injured players in their subsequent games missed, raw points, average ice-time, and 5v5 CorsiRel. I then noted what the players AAV was in the year they got hurt, whether the players changed teams by the start of 2011-12, and the details of their first post-injury contract, if applicable.

Some sourcing notes: for the man-games lost to injury numbers, I have relied on the excellent and excellently named site Man Games Lost, to which I am a happy subscriber. If you want to find this stuff, that's the place to go, though you'll need to pony up $5 to find the stuff I used here. However, it does rely (for the year I studied) on team descriptions of why the player wasn't playing, and there can be some tracking issues when a player changed teams mid-season. I have tried to check these against other resources whenever possible, but it did mean I had to infer a couple of times why a player was out (I assumed it was due to injury if I wasn't sure, to err on the side of caution.) Most of the rest of the information came from War-On-Ice (please keep the site up, guys) and General Fanager.

Okay, enough warm-up: here's the data, here's the explanation.

So what can we take away from this?

If you look at the group of surviving forwards—of whom there are 44—they averaged a loss of about 6.6 points in their first full year back from injury, though this is skewed by the enormous decline of Marc Savard (who had a 78-point drop); if you take him out, the remaining 43 guys were about 4.8 points worse. They averaged about six seconds less of total ice-time, and they were about a tenth of a point worse in relative possession on average (though the numbers are wild.) And yet, 20 of them (45%) changed teams before the start of 2011, whether by trade or free agency. On their next contracts, they averaged a raise of about $95,000, which is surprising considering many of them were reaching UFA status for the first time.

The story is similar on defence. Our 17 defencemen, on average, dropped 3.2 points, 42 seconds of ice time, and 1.22 percentage points in CorsiRel. On their next deals, they pulled an extra $77K over their prior salaries. 10 of them (59%) were playing for different teams by September 2011.

In other words, guys who missed considerable amounts of time in 2009-10 were, on average, not that far down from their prior form in 2010-11, but they nonetheless were often on the market. Losing four or five points isn't nothing, but it's also far from catastrophic--and remember, that's a raw count, so any games missed are baked into the result. I don't want to get carried away with the salary numbers in subsequent years, because there are many factors and the players were often multiple years removed from their injuries at the time. But in only a few cases did players get considerable raises soon after their injuries—most notably Paul Martin and, bizarrely, Rostislav Klesla. I think it's fair to say that injuries make teams more willing to let players go, and less likely to pony up for them—even though the evidence suggests they're usually still solid NHL players.

But wait, this is the survival group. What about the group that had been NHLers, then left the league?

There are five of them. Three forwards over 35 (Pavol Demitra, Jere Lehtinen, and Kirk Maltby), one 30-year-old third-pair defender (Martin Skoula), and one fringe fourth-line forward, aged 24 (Nick Tarnasky.) Lehtinen and Maltby, to look at their numbers, had essentially nothing left in the tank even before being injured; Demitra went to the KHL and put up a very impressive season (61 points in 54 games) before dying tragically in the Yaroslavl Lokomotiv plane crash. Tarnasky's career high in points was 10, and he struggled to hold down an NHL job even before his injury. Skoula also went off to the KHL, where he played decently for several seasons. In other words, I think the injuries in all of these cases at most nudged these players towards leaving the NHL, when they were already on the way out anyway—excepting Demitra, who almost certainly was physically able to keep playing an NHL role.

There are also two defencemen who took a year-plus to rehab before returning to an NHL role: Andrei Markov and Sheldon Souray. Souray rehabbed in the minors for a season, then returned to being a top-four defender in 2011-12, though he was nearing the end of his career. Markov spent three seasons looking like he was going to become the best player in the "out-of-the-league" part of this study, and then returned miraculously to form; he's been on Montreal's top pair for the past four seasons and has only missed two games in that time.

At most, I think these groups tell us to be more wary of players who are recovering in their 30s, but even then, most of the players who were good before taking their injuries recovered to become good in again. The list of under-35 players who went from very good to significantly worse and never recovered is short—you could argue the only name in these sheets that justifiably meets that standard is Savard.

So most rehabilitations of good NHL hockey players are largely successful. What does that do, besides making us feel good about the resilience of the human body?

3. Two Robberies And A Fuck-Up

June 29, 2010: Nashville trades Dan Ellis, Dustin Boyd, and future considerations for Sergei Kostitsyn and future considerations

The Habs were tired of Kostitsyn. He was often injured, moody, and bad with the press. The Predators picked him up for a career backup and a minor-league forward, then signed him for $550,000. Kostitsyn scored 50 points that season, and 47 in the next one. Dustin Boyd played ten games for Montreal before his NHL career ended for good. Dan Ellis was Dan Ellis.

Kostitsyn was obviously at unusually low value in the Habs' eyes for several reasons; he fought with his coach and Carey Price. But being seriously injured certainly did him no favours, and very possibly pushed him from "a headache" to "more trouble than he's worth." The Preds thereby got a scoring forward for a song and signed him for pennies. The Preds did what smart teams should hope to do: find players who are undervalued. And as the evidence above shows, injuries devalue players more in the eyes of their managers than in reality.

February 9, 2011: Toronto trades Francois Beauchemin to Anaheim for Joffrey Lupul, Jake Gardiner and a 2013 4th

The 2011 Anaheim Ducks had fond memories of Francois Beauchemin being their stalwart third defender on the 2007 Stanley Cup team. They also had 27-year-old scoring winger Joffrey Lupul, missing time with injury and with two and a half seasons left on his contract at $4.250M. Hmm.

Brian Burke had an uneven record as Toronto's GM. This trade, however, was an absolute theft. Jake Gardiner has blossomed into the Leafs' best defenceman (admittedly that hasn't always been saying much.) But Joffrey Lupul went on to produce 108 points in 113 games throughout the remaining life of the contract. His health continued to be imperfect, but he was an extremely productive player despite that. And this was for the salary dump part of the deal!

When a trade turns out as lopsidedly as this one did, it's worth a look at the psychology involved, and I'm certain the Ducks thought they were selling a falling stock. The Leafs did their due diligence, evaluated Lupul's injuries, and concluded correctly that there was value to be found there.

I think one of the more significant things is that Lupul was by no means a best-case scenario. He did miss 47 games over the last two years of the deal, which is an unusually bad result health-wise. But the Ducks really wanted rid of him, and the Leafs wound up getting a guy who made the All-Star game.

Note this is not to say anything in favour of the extension the Leafs signed Lupul to later on. That extension broke the cardinal rule of signing a merely-good player with a terrible injury history, which is to give term well into the player's thirties, and it looks stupid now. But the acquisition of him was brilliant, and is one the Leafs off the future should hope to emulate.

Of course, there are good ways to get injured players, and bad ones.

July 5, 2013: Columbus signs Nathan Horton for $31.7M over seven years

It had to be this, didn't it?

The Blue Jackets wanted to make a splash under new GM Jarmo Kekalainen, and they did it by bidding high on a player who was about to undergo shoulder surgery. He did, came back in January 2014, then was diagnosed the following October with a degenerative back issue that effectively ended his career.

First person who needs me to explain why Horton is different than Stamkos gets bopped on the head with a mallet.

Defenders of Jarmo (are there any?) might note the injury that stopped Nathan Horton was not the injury he had when he was signed, and that they can't have been expected to foresee it. But Horton had a rough injury history across the board, and had not been all that productive in the prior two seasons. He was also 28. Despite that, he's exactly the kind of player this article would recommend acquiring. But I wouldn't recommend signing him for seven years. If the market was setting Horton's price that high and that long, then the market inefficiency didn't exist in his case. In signing at the length they were, the Blue Jackets were taking the risk involved in signing Horton and multiplying it. The fact they didn't insure the contract is an aside in an article concerned mostly with cap strategy, but it suggests they were pretty cavalier about the downside of signing Horton.

More importantly, they were doing it for a player who wasn't that incredible. The best case scenario for a player at age 28, barring a Jiri Hudler moment, is that he matches his best seasons. Horton's best seasons, from which he was five years removed, were 62 points. The risk/reward calculation was very lopsided. Remember kids, term past age 30 is for superstars.

With that caveat, I think there's considerable evidence that teams are lower on injured players than they should be on a regular basis. That means there's something to be exploited—because most injured players under 35 recover well. As long as the term risk can be controlled, players in the process of recovering from injury can give you above-average play for a lower price in contract or in trade than you'd otherwise have to pay.

To sum up, I'll paraphrase a slogan attributed to the Rothschild banking family: when there's blood on the ice, buy players.

When you're ready.

4. Timing and Horseshoes

FiveThirtyEight, whom we somewhat distrust for hockey analysis but who are aces for politics and baseball, had an interesting article after the MLB trade deadline last season. The article tracked at what point it made sense for teams to trade wins in the future for wins in the present. It found that as you get closer to championship status, adding more Wins Above Replacement benefits you at an increasing rate. In other words, the better you are now, the more aggressive you should try to be adding even more talent in the short-term. Strike while the iron is hot, is the idea.

Conversely, as you get worse, wins in the present become less and less valuable to you, and you should be more and more willing to trade wins in the present for even fractions of wins in the future. While the expected championships article was about baseball, all of this is conventional hockey wisdom too; look at the traditional sorting of deadline teams into buyers and sellers.

This means that expected championships, and what changes teams should make to pursue them, form a horseshoe. On one side are the teams who are in full rebuild, eyes only on the future, trading for picks and prospects; on the other side are the teams who are all-in, hoping to win a championship in the current season. According to the FiveThirtyEight article, the worst thing to do is nothing; to tread water and thereby be meh now and meh in the future. Don't get stuck in the mediocre middle.

Let's follow this thinking a little further. If we know there is a point where it makes sense to sacrifice the potential for future wins in order to increase the potential for present wins, that means that any risk—not just deadline trades--with a chance of adding present wins becomes a better risk to take as your team improves. This definitely does not mean that any risk becomes good. It just means that smart gambles get smarter and borderline gambles start to be smart, if they have a chance of improving the team in the short-term.

When you're on the verge of a Cup, you want to add talent. But it gets increasingly hard to find guys who can put real separation between themselves and the replacement-level players you can call up from your AHL affiliate. You'd like to get a guy who can offer Wins Above Replacement, and you should be willing to take more risk in getting him, because your window is opening.

Further, if your team is likely good enough to make the playoffs—something that should be the case, if you really are moving up the horseshoe--you have much more time prior to April to manage a player's recovery conservatively. A team that knows it's going to be in the postseason doesn't have to rush players back in January. There is a double benefit here to being a top-drawer team: you get more value from adding good players, and you have more freedom of action when you get them.

The problem with smart gambles is they're still gambles, and given that executive fear discussed in Part I, gambles are a terrifying thing to contemplate when the most probable way for you to end your season is explaining to the owner why you didn't win. You could do all the right things, get knocked out in the first round, and whoops-a-daisy, you're unemployed.

Keeping that in mind, we can see why NHL GMs would be leery of adding injured players, even if the risk were an intelligent one. Injured players, as shown in Part II, can be a cheaper way to add quality above replacement level—they can kick Colin Greening out of your top nine, for example, and if healthy, make your team better. But the "if healthy" feels like too big a hill to spend money climbing.

This is too bad for mid-level players trying to rehabilitate after an injury, who suffer for being the kind of gamble that's statistically intelligent but professionally dangerous. The benefits that come from them working out are good but unspectacular, whereas the risks seem pronounced. These guys still get signed, usually, but the market drives their price down considerably.

That's where we come in.

5. The Ideal Candidate

Not all franchises are created equal. As Katya has noted, it can be easy for Leafs fans to forget that being able to conduct a rebuild is a luxury. You have to be sure your team can afford to embrace years in the wilderness without being financially devastated. This fear is part of the reason for some of the bizarre decisions the Canucks have made in recent seasons, and is partly why the Sharks have (wisely) not traded Joe Thornton or Patrick Marleau. Not many owners would be ecstatic to sign a coach for $64,000,000 and then have him declare that "there will be pain" upcoming. The Leafs did it and nobody blinked. This speaks volumes about our position in the NHL.

Looking back at the Horton-for-Clarkson trade again: this was a deal facilitated by the fact that the Leafs have much more real money than they need and the Blue Jackets do not. We committed to spending nearly $30,000,000 in real dollars to free up cap space via the LTIR designation. The Jackets were in the unenviable position of trading a total loss for a merely catastrophic loss. We took the loss the Jackets were so desperate to get rid of and we swallowed it whole.

Backtracking a moment: as we all know, the CBA has a designation called Long-Term Injured Reserve (described here by Clark), which permits teams to exceed the salary cap when replacing players injured more than ten games or twenty-four days. This designation is very helpful, as the Leafs have shown time and again in using it for players like Horton and Stephane Robidas. It is not a get-out-of-jail-free card, as ostensible "Dr." Aglikepull recently explained, and it is not a license to be stupid. But it's beneficial, and the Leafs are in a better financial position than any team in the NHL to exploit it, because they have so much more money than the cap normally allows them to spend. This means that risks that are smart in the abstract are even smarter for the Leafs to take, because we have a greater capacity to absorb cap-related injury risk than any franchise in the league.

Of course, this doesn't make everything easy. LTIR, again, isn't a panacea, and injuries are hard to predict. There's always the Marc Savard scenario. There will always be a chance of this blowing up in our faces.

But the Leafs have an edge here, too. The grass is always greener in Toronto. The Leafs should be giving way to no other team in medical knowledge on staff. Part of the way to manage injury risk is for the team to know all it can going in about what the player's injuries are and how the rehabilitation process should go—and the Leafs are better-positioned to manage that process, too. Further, as mentioned in Part III, a team that's in a go-for-it position can afford to be more patient in maximizing its team quality come playoff time, because it isn't as pressed to ensure it makes the playoffs. If the Leafs do rebuild effectively, they should end up both best-suited to absorb the risk and best-suited to maximize the benefit.

Most of all, what injured players offer is a potential for higher-level production at a reduced cost. That's what the Leafs should chase.

There's a wild card here, though, and it harks back to executive fear. Leafs ownership has been spectacularly trusting of Leafs' management since the hiring of Brendan Shanahan in spring 2014, and extremely generous financially. They've opened the corporate wallet multiple times for huge amounts, and have tolerated a 27th-place team openly conceding it was going to spend another season in the basement. Thus far, they've been dream owners for a front office.

This would seem to suggest that the Leafs, in this respect like the others, will be in a good spot to take smart risks when their time comes. But ownership can be fickle, and it's hard to know how they'll feel two or three or four years down the road, especially if the Leafs' revenue streams have gone through a downturn. MLSE's money men may want to see more of a return on their investment and become more interventionist, and what impact that might have is unpredictable. Rich corporations are rarely infinitely patient, and without the trust of ownership, the Leafs won't be able to take the kinds of risks described here.

Still, the early signs have been uniformly positive, the Leafs have long been a license to print money, and the Leafs' FO seems to enjoy universal confidence right now. May it deservedly continue, and put us within range of a Cup.

6. Our Future Appetite

All these advantages notwithstanding, the Leafs are not yet the ideal candidate. They are, hopefully, beginning the ascent from the bottom of the NHL, but they're building to add wins in two to four seasons. Adding boom-or-bust reclamation projects right now doesn't further that goal—either we're taking too much term risk or we're (at best) getting temporarily cheap value in a year where it doesn't do us as much good. We're at the wrong spot on the horseshoe.

But if the rebuild goes according to plan or well by accident, we could get to the right spot quite soon. I'm a loud and frequent proponent of the theory that good rebuilds don't have to be slow (laid out gorgeously by Arvind here); suffice it to say, the Leafs should be aiming higher, sooner than snarky Edmonton fans would have us believe. If the players the Leafs have assembled can be the core of an elite team, we could soon be looking at how to divvy up our remaining cap space for our best shot. And if—when—we get there, we should start getting hungry for rehab projects.

This doesn't just mean signing faint hopes to PTOs, which is essentially scrounging around for free wallets. It means recognizing that within reason, it can make sense for the Leafs to bid the highest on a player coming off a damaged year. It means that we can look at deals some teams want to unload and actually want to acquire them on their own account—because the Leafs can afford to be smarter and bolder than other front offices. That value discussed in Part II is there for the capturing, and when the time comes, we should be the team to capture it.

I'll end this article by explaining the title I chose. Glass damage is a type of insurance coverage, and the essential idea of it is the same as this article: just because something is fragile doesn't mean it's a bad risk.