Tuesday, March 31, 2015

12 ways to fix the NHL's draft lottery

With the playoff race coming down to the wire and the battle for top seeding in each conference raging on, the most anticipated game on last night’s schedule may have been the one between the two worst teams in the league. The Coyotes and Sabres met in a rematch of last week’s showdown, one that ended with Sabres fans cheering an overtime goal against their own team.

All of that is ridiculous, and it’s a situation the NHL had to know was coming. By going with a draft lottery system that guarantees the last-place team a top-two pick in a draft with two franchise players, the league all but guaranteed that things would get silly as we came down to the wire. The Sabres have been widely accused of tanking their entire season to ensure they get one of those two picks, and the Coyotes have given them a run for their money by trading away just about everyone with a pulse as the season wore on. Meanwhile, fans of other miserable teams like the Oilers and Maple Leafs watch with envy, wishing their teams could drop into the race for dead last. The whole thing is a mess, and the league should be embarrassed.

But while it’s easy enough to second-guess the NHL’s handling of the situation, I’m going to invoke my long-standing rule of hockey criticism: You don’t get to complain unless you can offer up a better idea. Luckily, when it comes to assigning draft order, there are plenty of other options available to the league. Some of those involve tweaks to the current way of doing things, others offer more radical changes, and still others discard the existing system entirely and come at the problem from a whole new angle.

You won’t like every idea below, but you’ll probably think at least a few of them would be improvements over what we’re stuck with now. So here are a dozen ways the NHL could go about the business of handing out draft picks, ranked in order from the least to most disruptive.

Option 1: Just keep things the way they are

We’ll start with the status quo. As a reminder, here’s how the system works today: All 14 teams that miss the playoffs are entered into a lottery, with the worst team getting the best shot at winning (20 percent) and the odds dropping from there. Whichever team wins moves up to the first overall pick, with everyone else slotted in based on the reverse of the final standings (meaning nobody can drop more than one spot).

And maybe that’s fine. After all, not everyone thinks the current system is a disaster. And even those who’d agree that this year’s silliness is a bad look for the league might argue that that’s mainly a result of a fairly unique set of circumstances — i.e., two blue-chip franchise players in the same draft. That’s a reasonably rare situation — the last time we’ve seen this sort of top two was probably back in 2004, when Alexander Ovechkin and Evgeni Malkin came in — so maybe there’s no need to panic over a problem that won’t be anywhere near as pronounced most seasons.

Biggest advantage: It would certainly be the easiest solution.

Biggest objection: The league has already decided against this approach, as we’ll see in the next section.

Option 2: The same basic system, but we draw more than one winner

What if we kept the same general idea behind the current system — a weighted lottery involving all the non-playoff teams, with the same set of odds tilted in favor of those that finished further down in the standings — but drew more than one winning number? Like, purely for the sake of argument, three?

That would mean you’d have three lottery winners, and they’d get the first three picks in the draft. Other teams could drop as many as three spots, so the team that finished in last place overall could end up picking as far down as fourth.

A system like that wouldn’t eliminate tanking, and it wouldn’t discourage fans from rooting against their own team to get better odds. But it would at least reduce the incentive to mount an all-out campaign to finish dead last, and it would prevent a situation like this year, when the reward for being terrible is a guaranteed chance to draft a franchise player.

Biggest advantage: The NHL has already decided to do it. The system I just described was approved last summer, and will go into effect next season.

Biggest objection: Other than not going far enough, the biggest problem with the league’s new system is that the NHL didn’t implement it for this season. That was understandable — you could argue that it’s unfair to shift the odds around on short notice — but in hindsight, the league should have pushed to get the change done before the offseason started. Don’t like it, bad teams? Make some moves and don’t finish dead last.

Option 3: Keep the weighted lottery, but establish a floor

This is a relatively new idea; I saw it for the first time a few weeks ago in a post on a Coyotes blog. Essentially, you’d keep the existing system, but establish a minimum points threshold that teams would have to hit to be eligible for the draft lottery. Maybe it’s a firm number, or maybe it shifts based on how the league as a whole is performing, but the core idea remains the same: If a team can’t meet that minimum threshold of on-ice respectability, then it loses its eligibility for the lottery.

Biggest advantage: Teams couldn’t blatantly tank if they knew they had to maintain some minimal degree of competitiveness. And you’d occasionally get late-season situations in which bad teams had to win their way into the lottery, which would be fun to watch.

Biggest objection: We’d probably spend the entire offseason arguing over what exactly the cutoff should be. And while we all hate tankers, sometimes a team finishes last because it really did get hit by a massive wave of bad luck; those teams could end up missing the cutoff even though they weren’t trying to tank.

>> Read the full post on Grantland


  1. Love the Gold system and The Wheel, but there's actually another option out there which almost sounds like a joke, but which is quietly brilliant: abandon the lottery and make the bottom teams swap their first round draft picks halfway into the season. In this case "bottom teams" could mean anything from "all teams at least xx points out of the playoffs" or just the "x lowest teams in the standings."

    For instance, let's say you're setting the cutoff at "the six worst teams in the league". Once all teams have played 41 games, their points total at the end of that 41st game is recorded - the six lowest all swap their first-round picks (30th swaps with 25th, 29th swaps with 26th, and 28th swaps with 27th). You then play out the rest of the season and set the draft order as normal, sans lottery. Assuming the standings don't change by the end of the season, the worst team's draft pick (now held by the 25th-place team) goes first, followed by the second-worst team's draft pick (now held by the 26th-place team), so on and so forth.

    I love this idea, because it has a number of advantages going for it.
    -It discourages tanking. The incentive to lose largely goes away when you don't hold your own pick, and no one wants to be the GM who just gave somebody else the first overall pick through their own ineptitude.
    -It encourages winning, even amongst the bad teams. When you win, not only are you reducing the value of your own draft pick (which you no longer own, so you'd rather it be less valuable), you're also pushing the other bottom-feeders even lower, meaning the pick that you DO hold probably just got even more valuable.
    -It still allows legitimately bad teams to get better (since even if you bottom out you're guaranteed a Top 6 pick out of the deal) - you might not wind up with Connor McDavid, but the consolation prize would still be decent.
    -It's a relatively minor change (compared to sweeping overhauls like the Gold system or The Wheel), meaning it actually has some chance of happening. Everything stays the same, you just abandon the lottery and institute the pick swap.

    This system shares a lot of its strengths with the Gold system, actually - like Gold's idea, it would force any tanking GM hoping to get a high pick into a balancing act where they would want their team bad enough to be in the running for the top pick, but not so bad that they can't start winning games in the second half to secure their position. In fact, deliberately getting the top pick becomes almost impossible under this system, because it would require *someone else* who you have no control over to finish last on your behalf.

    -The swap would occur regardless of who currently owns the picks being swapped. This prevents teams from trying to cheat the system by trading away a 1st overall pick for huge returns just before it gets turned into a 6th overall.
    -Any team who is forced to trade their pick in this manner is barred from ever reacquiring that pick via trade, exercising a right to swap picks, RFA compensation, etc. Once again, this is done to prevent short-circuiting the system by a team who tanks, then gets a hold of their original pick to ensure they reap the rewards. (Obviously this applies only to the team's original pick and does not prevent the team from acquiring other first round picks).

    1. Ottawa offered to implement this exact idea with San Jose in 1993 (where they'd trade 1st round picks), but San Jose refused.

  2. Regarding the weakness of the 'Gold System' - it's worth noting how late in the season that teams actually get mathematically eliminated from the play-offs. Obviously, Buffalo was effectively eliminated as soon as they got about 15 points out of the final play-off spot (if not in the pre-season), but they were mathematically eliminated much later - likely around game 50.

    This makes the 'switch' much harder to pull off, as you would have to tank the first ~50 games of the year, then start trying - it's difficult to see how this would be feasible. It's not as if you can keep a large number of NHL-ready players in the AHL for the first 2/3rds of the season - even if you ignore waivers, can you think of a worse free agent pitch?

    1. I would argue that that's the whole point. You don't want someone deliberately creating a terrible team and losing on purpose in order to gun for the first overall pick; you want them to try and ice the best possible team they can and try to win down the stretch. Even if they fail to actually get rewarded for their efforts, the fact that they're trying means you accomplished your mission.

      Without actually checking the standings to see if this is accurate, I think the team that would most likely earn the first overall pick under the Gold system this year would be Edmonton and - I hate myself for saying this, because Edmonton's front office is utterly horrible and does not deserve any type of reward for the atrocities they've wrought - that's actually a good outcome, because Edmonton seems to be the only one of the bottom feeder teams that didn't wind up there deliberately and is actually trying to still win games. Contrast Arizona and Buffalo, who've traded away everyone who threatened to be any good the moment they showed a spark of life, with Edmonton, who is icing a lineup that includes three first overall picks. Even if Edmonton is awful, at least they're awful honestly and not because they're deliberately sabotaging their own team in the hopes for a better pick (not this year, anyways). Managerial incompetence is preferable to (and much better for the game than) deliberate game-throwing.

    2. That's correct... as of yesterday, the Oilers would be leading the Gold standings.

      Full standings here: https://twitter.com/DownGoesBrown/status/583064717498793985

  3. Hi Sean,

    Huge fan of Down Goes Brown! Keep up the good work!

    I think a combination of Option 10 and Option 12 would be interesting. All draft-eligible players would become free agents. Teams could start bidding on players as soon as they're mathematically eliminated from the playoffs. This would give them a head-start on other teams, giving them an advantage over teams like NY and LA. We would also keep all the advantages of Option 12.

  4. Option 13:

    All 14 teams that miss the playoffs go into a weighted reverse lottery and the league draws all 14 positions. Ex. 17th place team gets 14 ping pong balls, 17th place = 13 ping pong balls, 14th place = 12 ping pong balls,....29th place = 2 ping pong balls, 30th place = 1 ping pong ball. Then you televise the lottery and draw each of the 1st round draft positions reversing from the 17th overall pick down to the first overall pick. I don't know what the odds would be but there would definitely be less incentive to tank, but poorer teams would still have slightly better odds of the top pick (but the risk of falling all the way to the 17th pick is there too) and there would be much more interest in watching the lottery.

    Rory (not McIlroy)......(nor Fitzpatrick)

  5. Crap....I gotta proofread better, and post while sober. Redoing my previous post....

    Option 13:

    All 14 teams that miss the playoffs go into a weighted reverse lottery and the league draws all 14 positions. Ex. 17th place team gets 14 ping pong balls, 18th place = 13 ping pong balls, 19th place = 12 ping pong balls,....29th place = 2 ping pong balls, 30th place = 1 ping pong ball. Then you televise the lottery and draw each of the 1st round draft positions reversing from the 17th overall pick down to the first overall pick. I don't know what the odds would be but there would definitely be less incentive to tank. Poorer teams would still have slightly better odds of the top pick (but the risk of falling all the way to the 17th pick is there too). And there would be much more interest in watching the lottery.

    Rory (not McIlroy)......(nor Fitzpatrick)

    1. This comment has been removed by the author.

    2. I didn't realize that Rory's idea included a lottery to determine the 'reverse' standings as opposed to assigning a number of ping-pong balls based on actual order of finish--which is a plan that I have been suggesting for years. Under this system the odds of winning the #1 pick would be as follows:

      standings this plan current difference
      30th(BUF) 13.33% 20.00% -6.67%
      29th(ARZ) 12.38% 13.50% -1.12%
      28th(EDM) 11.43% 11.50% -0.07%
      27th(TOR) 10.48% 9.50% +0.98%
      26th(CAR) 9.52% 8.50% +1.02%
      25th(NJD) 8.57% 7.50% +1.07%
      24th(PHI) 7.62% 6.50% +1.12%
      23rd(CBJ) 6.67% 6.00% +0.67%
      22nd(COL) 5.71% 5.00% +0.71%
      21st(DAL) 4.76% 3.50% +1.26%
      20th(SJS) 3.81% 3.00% +0.81%
      19th(FLA) 2.86% 2.50% +0.36%
      18th(WPG) 1.90% 2.00% -0.10%
      17th(OTT) 0.95% 1.00% -0.05%

      As you can see, the odds don’t change significantly for any non-playoff team *except* the league’s worst. There is little incentive to finish 30th as opposed to 29th or 28th. Problem solved. You’re welcome, NHL.

  6. A proposal to save the draft-position tourney from players not caring: make it short. Maybe guarantee a home game each, for fan interest, then hold a single elimimation tourney from eight teams, in a non-NHL building where fans would have fun coming to see the big show.
    A second, crueler proposal: if the players hate it, make winning the way to end it. Have the losers progress at each round. The teams winning each round would get equal odds in the lottery, and the losers would compete for the next round's odds, which would get progressively worse. Cruelest part: high-stakes final to determine the very worst team. Cruel, but fun.

    1. Longtime lurker, occaisional commenter. Keep up the good work, DGB!

  7. If we're going to discuss really radical proposals, here's one that combines a few techniques. Do a conventional last-to-first draft order, but institute a relegation/promotion arrangement with the AHL. Setting the number of teams relegated would be tricky- since the low end of the standings is less fluid than the bubble of playoff qualification, relegating too few teams makes it easier to hang out one spot ahead of the cutoff; on the other hand, relegating too many would create large imbalances as underqualified AHL teams are thrown in over their heads. A solution to this is to randomize the number of non-playoff teams relegated, biased to the low end. For example, flip a coin repeatedly and count how many times tails comes up before heads does- thus the worst team is relegated on average every 2 years, the next-worst averages every 4 years, up to the best non-playoff team averaging over 16,000 years between relegations.

    Since the relegation cutoff is unknown, you can't ever be certain you're safe. Every position you drop in the standings doubles your odds of getting the boot, which ought to be sufficiently worrisome that nobody does it on purpose, even when a good prospect is available.
    The worst teams in the league (including the promoted former AHL teams, if any) get the best draft picks.
    Everyone in the standings is playing for something- the bottom teams are trying to avoid relegation, the mid-range teams are fighting to qualify for the playoffs, and the powerhouses are struggling for home ice in the postseason.

    Good luck getting any owner to agree to a system that may one day relegate their team.
    Some AHL teams may play in arenas that are too small for NHL play, so having sufficient facilities would have to be a prerequisite for promotion [or a grace period could give a newly-promoted team some years to bring their venue up to spec].
    Promoting a team that has an affiliate agreement with a current NHL team would raise the possibility for management sabotaging one in favor of the other.

    1. If there had to be a relegation option, a full non-playoff option wouldn't work. If anything, the AHL/NHL relegation option would only work of there's eligibility ratings added. Even the "best non-playoff team is relegated once every 16,000 years, the chance of a high non-playoff team in the NHL like the Kings or Bruins being relegated after a elimination on the final day of the season while a seventh-seeded AHL playoff team makes it to the NHL would be far worse for competitive balance.

      By contrast, I'd use an eligibility rate for a promotion/relegation: Only teams that dip below 60 points in the NHL are eligible to be relegated (which is usually the worst of the worst in the modern era- usually the worst teams around 70 points when the loser point comes into play), and only teams that had over 100 points in the AHL and/or the Calder Cup champion are eligible for promotion (making sure only the best of the best AHL teams in any given year have a chance.)

      As a result if it happened this season- Buffalo may have put themselves out for McDavid- but in the process, they just knocked themselves out and welcomed Manchester into the NHL- and suddenly, Arizona has to sweat as Utica, Grand Rapids, and San Antonio are playing their final three games to see if they can get 100 points and relegate the Coyotes.