What is the value of a draft pick?

102419748_10_std

The 2014 NHL entry draft is nine days away, and there has been a lot of discussion about the possibility of the Oilers, Flames, Panthers and Islanders possibly trading up or down. We rarely see much movement within the top-five picks, so it it unlikely these teams trades spots, but in the past we have seen many teams trading up or down once you get past the first ten picks.

Most of those trades occur in the 2nd, 3rd, 4th and 5th rounds, and often we debate if it was a good move. The Columbus Blue Jackets’ director of hockey administration and resident analytics guru, Josh Flynn developed a chart that would rank the value of a draft pick, and it gives us some insight into whether those draft day deals involving picks are good moves or not.

Flynn was not the first to build a value chart, but the ideas behind his interested me.

Rob Mixer wrote an article about Flynn’s chart here. The NFL has had a value chart for years, but you can’t compare it straight up to the NHL, because NFL players are 20-22 of age and they go directly to the NFL. The majority of NHL draft picks spend two to four more years developing in junior or the minors before they crack an NHL lineup. So the success rate of players would be much different, not to mention NFL has 30 more roster spots, so more chance for players to make team. But as a pick for pick value, I think it could be close.

Flynn explained to Mixer how he came up with the value system for his chart.

“I wanted to come up with something to help guide us through
pick-for-pick trades, but the problem was that, in the NFL, players are
going straight into your lineup,” Flynn told BlueJackets.com. “I decided
to think about value of picks this way: what is each pick worth in
terms of what the asset will produce for the club that owns it?

“Can you define a pick’s ‘success’ based on how many games the player
plays? You don’t want to rank picks based on how many points the player
produces, because then you get into roles and positions and it gets too
complicated. I said ‘you know what, we’re not going to look at it
objectively.’ When Tyler Wright (former
director of amateur scouting) was here, I’d tell him: when I compare
you to other players based solely on games played, you and Pavel Bure
are the same player. And no offense to Tyler, but I’d rather have Pavel
Bure with my pick. What I ended up doing was grading the
players based on what they became at their peak, and tweaked it based on
how long they sustained that level of play.”

When you look at a period of 10-12 years and there’s a 25 percent
success rate in the second round, pick 46 may have produced six guys and
pick 45 produced a few less, but you’d still rather have the 45th
pick.” Flynn said. “We have a tool that we can go into the draft with
and guide our decisions on trades with a real, solid idea of what a pick
is worth.”

The value chart used by the Blue Jackets is born from a complex
system that uses data compiled from a dozen years of NHL drafts – dating
back to when the draft was nine rounds deep – and from that data, Flynn
developed a rubric of grading picks on a 1-10 scale (for example, picks
1-210 in a seven-round draft) based on what they produced over time.

Flynn then “clustered” the picks together to create a firm value for
each one, and then, places them into a formula that determines the
pick’s final value on the chart.

“Basically, if pick 65 in 2006 produces a player who never played an
NHL game, that pick is going to get a zero,” Flynn said. “On the flip
side, you can figure out that the No. 1 overall pick is generally going
to get a higher rankings on the value chart because it has always
produced an NHL player.


“The system is designed to show what picks are intrinsically worth,
and it helps us weigh draft positions against each other in terms of
value. The chart is based on what we think each pick will produce, and
it’s backed up by a dozen years of information.”

Here is what NFL chart looks like.trade-value-chart

It is interesting to look at this if NHL valued picks the same. Many have wondered what it would cost for the Oilers to move up from #3 to #1. If we use that chart, and only include draft picks, it would take the #3 and the #21 (total of 3000), for the Panthers to consider it.

I don’t think the Oilers, or any other team, would give up the 21st pick just to move up two spots in the draft, unless there was a Sidney Crosby-type player available. And because it involves 18 or 19 year old kids, there is always going to be wrong decisions. This is just a guideline.

In 2002, the Blue Jackets traded up from #3 to #1 with Florida. The Panthers received the option to switch picks with Columbus in 2003 and ended up moving from #4 up to #3 where they took Nathan Horton while the Blue Jackets picked 4th and took Nikolai Zherdev.

If you were Craig MacTavish would you offer the #3 this year and the option to switch draft spots next season with the Panthers in exchange for the #1? Would Dale Tallon?

Scott Howson was the GM in Columbus when they started using their value chart. Columbus has been using it since 2009. He used it when they traded down from #16 to acquire more picks, and then traded back up from #26 to #21 to take John Moore.

Howson was with the Oilers at last year’s draft when MacTavish twice traded down to acquire more picks. For fun, based on above chart chart did the Oilers get good value for their picks?

The Oilers traded the #37 to Los Angeles (value of 530) for picks  #57 (330), #88 (150) and #96 (116). If you go solely on the numbers the Oilers traded 530 “points” and received 596 in return. Based on this chart that was a good trade for the Oilers.

The Oilers then traded the 57th pick to St.Louis (330) for #83 (175), #94 (124) and #113 (68). They traded 330 points for 367. If you judge the trade solely on the value of the picks, MacTavish made another solid deal.

Howson was in meetings when I posted this, so I couldn’t get an answer whether he uses a similar with the Oilers, but I will ask him next time we chat.

Here are the actual players that were taken with those picks.

#37- Valentin Zykov (Kings)
#57- William Carrier (Blues)
#83- Bogdan Yakimov
#88- Anton Slepyshev
#94- Jackson Houck
#96- Aidan Muir
#113- Ben Betker

It is much too early to say which team will benefit the most from these deals, but I plan on using this chart during next Friday’s draft as a guideline to see which teams make good value trades based on his data. I know it isn’t exact, because it is football, but the NHL one likely has similar differences in picks from round 2-7.

I’m sure there are other charts out there, and if you know of any let me know. I’ve researched draft picks in the past and from  the 1996-2005 drafts my data suggested teams had a 60% chance of getting a good NHL player in the first round, a 23.7% chance in the 2nd round, 15.8% in
the 3rd, 8.6% in the 4th, 7.1% in the 5th,  9.1% in the 6th and 9.7% in the 7th round.

It sound like Flynn’s chart is more precise to the value of each pick rather than just the round, and I will be closely monitoring this year’s draft to see which teams made good value trades.

What makes it fun is that every team scouts and fan base will value a player differently, regardless of what the chart says. You can play the percentages, but if you are a gambler you know that sometimes you go with your gut despite what the odds tell you.

PRESEASON…

If you missed it the Oilers released their preseason schedule yesterday.

Date Opponent Time
Sunday, Sept 21st  Calgary Home (split squad) 6 p.m.
Sunday,
Sept 21st
Calgary Road (split squad) TDB
Wednesday,
Sept 24th
Winnipeg Road 7 p.m.
Sunday,
Sept 28th
Chicago Road (In Saskatoon) 4 p.m.
Monday,
Sept 29th
Winnipeg Home 7 p.m.
Wednesday,
Oct 1st
Arizona Home 7 p.m.
Thursday,
Oct 2nd
Vancouver Home 7 p.m.
Saturday,
Oct 4th
Vancouver Road 7 p.m.

It would be nice if the game in Calgary on the 21st was an afternoon game, which would allow for both games to be televised on Sportsnet. Fans could watch a battle of Alberta double header. Fans of both teams will be starving for hockey by then, and I’m sure the audience numbers would be decent.

It also looks like the Oilers will make most of their cuts late in camp. The Oilers play five games in seven days and I’m guessing Dallas Eakins will want extra bodies in camp so no one has to play more than three games in the final week.

Recently by Jason Gregor:

  • slats432

    Thank god we got good value in 2003 when we got 1030 points in 22 and 68 for our 17th pick (950).

    That said, it is good to at least have a metric to measure by.

  • HardBoiledOil 1.0

    the pick is worth whatever a team is willing to part with. although i am not a fan of trading down, the Oilers essentially turned two 2nd rounders into 6 lower round picks so time will tell if quantity turned out to be a better move than quality, though the players available in most of the 2nd round, perhaps #37 on down weren’t that great in my opinion.

    • pkam

      Actually, the Oilers traded a 2nd rounder (#37) into 2 3rd rounders (#83 and #88) and 3 4th rounders (#94, #96 and #113). That #57 was a 2nd rounder coming back from trading the #37.

      • Jason Gregor

        They actually turned two 2nd rounders into two 3rd rounders and three 4th rounders.

        They acquired five picks in total and traded away two picks.

        You have to include the 57 pick as something they gave up.

        • pkam

          Using your logic, we give away 530 (#37) and 330 (#57) to acquire 175 (#83), 150 (#88), 124 (#94), 116 (#96) and 68 (#113).

          So at the end, we give away a total value of 860 to acquire a total value of 633. And your conclusion is that the Oilers did pretty good?

          That #57 was not ours at the beginning of the draft and was not ours at the end of the draft. It is just a pass-by during the process.

          If we use the chart you provided, we trade away 530 and get 633, net gain of 103 or a #99.

          • Jason Gregor

            They are separate transaction. Very simple to see.

            The Oilers made two trades. You can’t say they only traded away 57, but never received it. Good grief. This isn’t complicated.

          • A-Mc

            No i think he’s saying that the Oilers went from having #37 at the start of the draft, to finishing with 83/88/94/96/113.

            Yes there were a couple transactions to make that happen (namely the 57 trade) but because 57 wasn’t ours to begin with, if you’re judging the oilers on the day as a whole: they started with 37 and ended with the 5 other picks.

            THEREFORE

            37 (530) –> 83(175), 88(150), 94(124), 96(116), 113(68)

            530 value turned into 633 value

            As a math equation, the value transaction on the day looks like this:
            -530 + 330 + 150 +116 – 330 + 175 + 124 + 68 = 103

            the +330-330 cancel out (which is the value of the 57 pick) so its like the 57 never happened unless you want to evaluate the 2 independent trades on their own as opposed to all trades on the day as a whole

          • Jason Gregor

            Go back and read what he wrote…

            “Using your logic, we give away 530 (#37) and 330 (#57) to acquire 175 (#83), 150 (#88), 124 (#94), 116 (#96) and 68 (#113).

            So at the end, we give away a total value of 860 to acquire a total value of 633. And your conclusion is that the Oilers did pretty good?”

            That is so wrong it is funny. You can’t say they only gave away 57, but didn’t receive it. I’m done. This isn’t hard to add up.

            Based on the value of the chart the Oilers won the trades. Add it up any way you want. But no one can pretend that #57 isn’t part of the trades. The 2nd trade doesn’t happen without 57.

          • pkam

            We didn’t give away #57 for the 2nd trade, we give away part of the return of #37 for the 2nd trade. We didn’t have the #57 pick without the 1st trade. All the return is the result of the #37 trade.

            If we don’t know there are 2 trades, all we know is we traded #37 for 2 3 rounders and 3 4th rounders.

            For example, if we trade away a player to get a pick from team A, then use that pick and another prospect to get a player from team B. Will you say we give away a player, a pick and a prospect to get the player, or we give away a player and a prospect to get the player?

          • A-Mc

            You guys are saying the same thing, you’re just not seeing it. Look at his previous posts regarding the transactions. He’s trying to say the net gain for the Oilers, on the day, was 103 pts, or a #99 pick worth of value. You were saying the same thing. The only issue is that there is a miscommunication happening.

            At the beginning of the day the Oilers had #37 only. By the end they had 83,88,94,96,113; a value Add of 103 pts. On this we can all agree.

          • I believe Gregor is right. You cannot look at the transactions as a whole; what if he couldn’t find a second trading partner? Judge the first trade on its merits, and the same with the second.

          • A-Mc

            I think you’re right.

            We started with the #37 only. In the end we ended up with #83,88,94,96,113.

            That 57 can’t be counted because it wasn’t value we started with.

  • Quicksilver ballet

    Interesting stuff. It certainly reflects the incremental value between the 1 thru 4 spots in this upcoming draft. Ekblad certainly the consensus No.1, followed next with Reinhart and Bennett in a photo finish for those 2 and 3 slots.

    This pretty much confirms the Oilers would be leaving something on the table selecting a lesser rated player in that 3 slot, no?. If we’re going to apply this blueprint to next weeks draft.

    • pkam

      So if Ekblad is gone and Reinhart is still available you think we should draft Reinhart so we actually take something off the table instead of leaving something on the table?

    • Jason Gregor

      Can’t say it confirms that, because it is one team’s value rankings. This isn’t a guaranteed system. It is a guideline that they use. It can’t be the same from year to year. When Crosby is #1 he is worth more than any other #1 and the gap would be different from him to the picks below him.

      Every year we see some of the best player drafted in the 2nd round, or late first round.

      And because central scouting, Red Line report or Bob McKenzie rank a player in a certain spot, doesn’t mean it is correct. None of them had Shea Weber, Duncan Keith or PK Subban as top-ten picks in their draft class, but they would be now.

      Go back and look at rankings and compare them to how players turn out five or ten years later.

      The pre-draft rankings are wrong just as often as NHL teams are.

      I don’t think anyone can say with accuracy that Bennett, Reinhart will automatically be better NHL players than Draisaitl, or vice versa.

      The other factor is how a team develops the player once they draft him. There are many factors after the draft that impact what type of NHL player their pick becomes.

      I don’t think you should say the ranking confirms anything. It gives us something to consider, but it is far from a certainty. There are many uncontrolled variables that go into drafting, and much of them happen after the player is picked by a team.

  • A-Mc

    Pretty cool post Gregor. Thank you!

    You’ve given me something to print off and nerd out on as i get loaded while watching the draft in 9 days.

    I hope Vancouver or Calgary make pick trades that don’t add up favorably so i have yet another thing to rag on them about.

    GOilers!

  • Wow! by this metric Alex Plante is worth 580 more points than PK Subban. I get that it’s trying to quantify the value of a particular numbered pick given all the past data in order to help GM’s make smart decisions about moving up or down or making trades. But it’s really no substitute for a scouting department that can pull a Subban out at 43, a Benn out at 44, or a Weber out at 49.

  • vetinari

    Great article. I can’t see Howson abandoning the model just because he rejoined the Oilers and it does give you something to use to objectively weigh the value of a draft pick, especially when you are trading multiple picks from different rounds between teams.

    It is also interesting to note the drop off in value between certain picks, such as pick #7 being worth only half the value of the #1 pick, or that picks #23/#24 are worth about half the value of the #7 pick.

    Also, while I am not saying that MacT should do this, but according to the chart, he could theoretically trade the #3 pick for the #4 pick plus the #50 pick (or close to it) and get equal value in return.

    Looking forward to watching whether the Oilers make any pure “pick for pick” swaps at the draft and if the points totals end up favouring the Oilers.

  • RyanCoke

    This is getting retarded, you guys are all saying the same thing except for pkam. Gregor would agree with the math because it factors in the gaining and losing points. pkam is not factoring in gaining the 57 pick. It is 2 trades yes but the second trade would not have happened without the first one but at the end of the day we started with 1 pick and got 5. Everyone agrees on this including Gregor. Only person not getting it is pkam lol

    • A-Mc

      he is actually, but for some reason there is a miscommunication. When i read what pkam is saying, he’s saying the same thing gregor and I are saying but for some reason pkam is wrong and gregor is right (According the trashes/likes).

      This is the post in contention made by Gregor:

      “They actually turned two 2nd rounders into two 3rd rounders and three 4th rounders.

      They acquired five picks in total and traded away two picks.

      You have to include the 57 pick as something they gave up.”

      Depending on how you read this, it’s inaccurate. It should actually read “They acquired SIX picks in total and traded away two picks; but 1 of the picks coming in were traded away as part of the process”. We did NOT trade 2x 2nd rounders for 2x 3rds and 3x 4ths. We traded 1x 2nd rounder for 2x 3rd and 3x 4ths. That #57 pick was never ours and it came as PART OF the #37 trade.

      If the Oilers started with #37 and 57 and we traded them both for the 3rd/4th rounders, the value equation doesnt add up because THAT FIRST TRANSACTION would be: #37 = 88, 96. Or in value terms: 530 –> 150 + 116. which is an insane loss in value (530 vs 266).

      Is this making sense to people? Pkam is only arguing that one post Gregor made

    • pkam

      This is what JG said: “They actually turned two 2nd rounders into two 3rd rounders and three 4th rounders.”

      If you think carefully, we didn’t turn 2 2nd rounders into two 3rd rounders and three 4th rounders. We turned one 2nd rounders into two 3rd rounders and three 4th rounders.

      If you really want to count both transactions, then we turned two 2nd rounders into one 2nd rounder (#57 coming back from the 1st trade), two 3rd rounders and three 4th rounders.

  • Quicksilver ballet

    Thanks for that explanation Jason. I see the Oilers are using a metric of their own to try and tilt the odds a little more in their favor.

    In usual QSB form, if you could allow me one more metaphor sir. Isn’t this like giving a visually impaired kid at the marble table 10 marbles. Instead of 7, like the rest of the visually capable kids at the same table. With their scouting track record the last 15 yrs. That eventual outcome isn’t in doubt.

    • A-Mc

      Yes =)

      That #57 comes in as well as goes out. When Gregor says we traded 2x second rounders for 2x third rounders and 3x forth rounders, he’s missing the addition of #57 in what he’s saying and is therefore inaccurate.

    • Jason Gregor

      Huh?

      Read the article. It states very clearly what was traded.

      The Oilers traded the #37 to Los Angeles (value of 530) for picks #57 (330), #88 (150) and #96 (116). If you go solely on the numbers the Oilers traded 530 “points” and received 596 in return. Based on this chart that was a good trade for the Oilers.

      The Oilers then traded the 57th pick to St.Louis (330) for #83 (175), #94 (124) and #113 (68). They traded 330 points for 367. If you judge the trade solely on the value of the picks, MacTavish made another solid deal.

      I’m amazed that people think when you acquire a pick, but then trade it later then it was never their property.

      The Oilers can’t make the 2nd trade, if they didn’t own the #57 pick. Simple.

  • DieHard

    Bottom line:

    We traded
    Valentin Zykov

    For
    Bogdan Yakimov
    Anton Slepyshev
    Justin Houck
    Aidan Muir
    Ben Betker

    We can see in 3-4 years how this all worked out.
    We have a chance of being ahead.

  • BLAKPOO

    Obviously any tool you can employ to make your job easier would be nice, but setting a hard value for a pick based on draft order alone is a terrible idea.

    If anything, scouting teams should be assigning values to individual draft eligible players, based on their own criteria ( which could be a combination of talent, ceiling, size, position, system depth, etc. ) then slotting them on a similar chart based on their predetermined values. Then when offers come up to swap picks, they can look at the values of those specific players to determine if they have a chance to come out ahead.

    A generic table, like the one shown, says getting the #6 and #7 overall in exchange for McDavid at #1 is a win. Which is obviously ridiculous. Every pick at every level would have the same inconsistencies.

    Unless tailored and valued by a specific group, based on a specific draft year, and those draft eligible players, a “value chart” is misleading and useless.

    • pkam

      When you trade up or down, you may know who is available at the earlier pick, but you have no idea who you may get for the later picks. So how can values assigning to players going to help?

      • BLAKPOO

        Huh?

        Trading down is a crapshoot however you slice it. There literally is no tool you could use to positively confirm who will be available at a later pick.

        Unless it’s a time machine.

        At least if you have values assigned to players, you could compare your list to “projected” lists published by third party services; like ISS and Central Scouting. Then you could see if you have players valued higher or lower than their projected draft positions, and hope to pick up players you potentially have rated higher than everyone else with lower picks.

        Of course, this is pretty much the fundamentals of scouting.

        I was initially referring to adapting a table to the process that allows you to calculate the potential value of picks on your list in comparison to other lists. If someone offers you multiple late round picks for a higher round pick, you could average out the values of the 3 players closest to that draft position on your list and compare the sum of those values to that of the player you’d most likely give up with the higher pick.

        It’s by no means a way to guarantee better or more accurate results, it’s merely a tool that might be beneficial to the process.

        • pkam

          How do you know if the 3 players closest to that draft position are still available or not by then? And it is also possible some players higher up in your list that you don’t think will be available then become available.

          Since there is no way to determine which players will be available for the later picks, what good is values assigned to player? Might as well assign value to the draft position since both are a pure speculation at the point of trading the draft.

          • BLAKPOO

            At this point, it literally becomes the difference between a guess and an educated guess.

            There’s always different ‘cliffs’ that seperate talent. You might have a strong draft through the first 30 picks one year, and a strong draft through to only the first 18 the next. Or maybe one year the 40-60 class is stronger than that of another year.

            Why would you choose a generic tool that won’t compensate for these shifts, when you could use one where you assign your own values, based on all your scouting info, to allow for more accurate choices?

            Not much common sense in your argument here.

          • pkam

            My argument is that you have no idea what players will be available for the later picks when you make a trade, so how practical is assigning values to players?

            If you think the year is a deep year, you can always adjust the value of the picks.

            Not saying this is perfect, but at least doable. Your suggestion is not unless you have a crystal ball.

  • RyanCoke

    Now how does this graph work when you factor in the common phrase that whoever ends up with the best player wins the trade? Like whoever ends up with the highest pick won the trade. Because even though there is a net gain we still lost the 37 pick which was the highest pick even though we got 5 picks in the end. Like what if some one ended up with the first overall pick but had to trade the #6 and #16 and #25 overall. The math says the #6 and #16 and #25 won the trade but I would argue that the guy with the first overall pick won the trade especially in a conner mcdavid draft or crosby draft. I would say the same thing if you through in 2 more picks in the second round.

    • pkam

      This chart is not about players, but draft picks. Nobody knows who the team is going to draft with that pick and it is hard to tell how good a player the 18 years old will turn out to be.

  • What is the value of a draft pick in a deep year like next year or the year before?

    Would Nurse be considered as good or better then Ekblad who could go number 1 overall?

    How would you rate that if the draft is deep and good?

  • pkam

    Three oiler fans walk into a bar.
    The first oiler fan (teir 1) orders a Crown Royal neat.
    The second oiler fan (teir 2) trades the Crown Royal for 2 shots of CC.
    The third oiler fan (shirtless)trades the 2 shots of CC for 4 shots of the house brand.
    I like the way this works.

    Walks away happy cuz fan 1 paid for everyone’s drinks.

  • The Soup Fascist

    A third round and fifth round selection leave on a Via train from Edmonton Thursday at noon, travelling at 80 km/hr and a second rounder and fourth rounder leave by car from Edmonton on Friday at 8:00 am travelling at 120 km/hr?

    Who arrives in OKC first?

  • OldSchool

    Whne looking at Prospects we ask what peers have they performed like?

    .20+ GPG and .30 PPG 18 year old Seasons in RSL/KHL last 15 years.

    SEMIN .21 GPG .36 PPG

    Ovechkin .24 GPG .45 PPG

    Malkin .23 GPG .61 PPG

    Kuznetsov .38 GPG .73 PPG

    Tarasenko .21 GPG .45 GPG

    Slepyshev .27 GPG .35 PPG

    Yakupov .41 GPG .82 PPG

    Yakimov .21 GPG .38 PPG 58% FO%

    of the 8 .20 GPG and .30 PPG forwards in thre last 15 years out of russia we have 3 and 2 were picked din the 3rd round last year.

    KHL site says Yakimov is curently 6′ 4.375″ and 222lb and scoring goals like Semin/Malkin/ Tarasenko at same age.

  • OldSchool

    I have to be missing something here? The NFL is a deep deep draft every year into the 2nd and 3rd rounds at least. If you were to put a value comparison between an NHL draft and an NFL draft. A late 2nd or even an early 3rd in the NFL would be equivalent to a late 1st in the NHL. A lot has to do with the shear #’s of available prospects as well as a later age when its more easily known the talents the player possess’s. How can these two leagues be compared with such a divergent set of circumstances?