Tuesday, August 14, 2007
Select company
Jiri Tlusty, Trevor Lewis, Jack Skille, Marc Staal, Andrei Kastsitsyn, Joffrey Lupul, Stanislav Chistov, Igor Knyazev, Jens Karlsson, Raffi Torres, Artem Kriukov, Krys Kolanos, Henrik Sedin, Kris Beech, Scott Kelman, Martin Skoula, Ty Jones
Pop Quiz: what do Sam Gagner (and Alex Plante) have in common with this list of players?
- Skaters picked in the Top 20 of the NHL Draft (list is 1997-2006)
- 2nd skater selected from their amateur club in the same draft
For NHL teams, the draft is an educated guessing game, where you hope that your theories and your system and your decision-making process will earn you a success rate over time that's better than that of your peers (the luck will even out, always does). There's no magic formula, but there are pitfalls to be avoided, and this looks to me like it might be one. Because looking backwards, the last time an NHL team hit the jackpot using a Top 20 pick to select the 2nd best draft-eligible player on a team was 1995, and no I'm not referring to Terry Ryan.
[**This works in both directions: when there are other elite-level prospects on a junior/ amateur team, the #1 guy is going to get more looks as well, and therefore (per the Saw Him Good theory and the last 10 years of results) be drafted higher than he ought to have been. Excluding the first 4 guys on the list (starting with Kastsitsyn), here's the list of the #1 skaters, picked before their teammates (draft position in brackets):
Zherdev (4th), Bouwmeester (3rd), Svitov (3rd), Kovalchuk (1st), Sjostrom (11th), Klesla (4th), Vorobiev (11th), Orpik (18th), D. Sedin (2nd), Brendl (4th), Saprykin (11th), Henrich (13th), B. Ference (10th)
Kovalchuk, Bouwmeester, and Sedin are hard to argue with in retrospect. But other than probably Klesla, everyone picked from 4th on would have to be considered busts relative to their draft positions.]
# posted by Matt @ 4:29 pm 
Comments:
Briefly looking at the players outlined I'm already doubtful of your proposed effect. I'd wager the underlying mediating factor between the first and second group is that the first group on average was drafted in the bottom half of the top 20, while the second group was on average drafted in the first half. As most who follow the draft will tell you, there is a generally substantial drop-off in the quality of an NHL prospect between those two ranges.
If you're somehow hoping that Gagner turns out to be a bust I would suggest you spend more time worrying about Calgary's own Swedish darkhorse from the same draft.
As most who follow the draft will tell you, there is a generally substantial drop-off in the quality of an NHL prospect between those two ranges.
I fail to see how this counters anything. The fact that there is a "generally substantial drop-off" in quality only supports the "Saw Him Good" theory. The whole point is that NHL GM's, for the most part, do a terrible job of drafting players. If GM's can't find quality in the 10-20 range of the first bloody round, they should just put all the names in a hat and randomly pick.
There's very little question that they should pick the highest rated player (composite scoring of major scouting services' ranks) or if there's a "tie" just throw a dart. There's no evidence that picking a Riley Nash is a "smart" pick under any circumstances. It's extremely unlikely for off-the-board picks to turnout better than all the players rated in front of them.
Drafting seems to be another big ego thing for management types. Ie- I know hockey better than you guys, nya nya. It's pretty hard to argue that anyone is consistently making steals in the first round.
I remember watching Brendl play for the Hitmen. God he was good. Too bad he never learned to play even the slightest bit of defense.
I honestly think people post stuff just to angry up the blood. I don't know what measure anyone wants to use, but I would like to challenge ANYONE to take a look at the NHL Entry Draft over any 10-year period and make an argument against this statement:
NHL teams generally do an excellent job drafting the most talented players early in a given draft.
Exception being the Russian/Euro Bure era stuff that buggers up the curve.
I don't know which point you are arguing against, LT. Are you saying GM's doing a good job of drafting talent, or not?
He is asking you to support your contention that GMs "for the most part do a terrible job of drafting players." After all, a player's draft rank is a fairly reliable estimator of his ultimate placing within his draft cohort; the draft is not a total crapshoot. #1s do fare better than #2s, #2s better than #3s, nos. 3-5 better than 5-10, 5-10 better than 10-15, etc. Based on my own observations the overall noise doesn't start to dominate the signal until the end of the first round at earliest.
the last time an NHL team hit the jackpot using a Top 20 pick to select the 2nd best draft-eligible player on a team was 1995, and no I'm not referring to Terry Ryan.
Petr Sykora?
I think that it's worth noting that going back to 1990 we add the following names to the list:
Terry Ryan (8th), Jarome Iginla (11th), Brad Church (17th), Petr Sykora (18th), Ethan Moreau (14th), Jason Allison (17th), Darius Kasparaitis (5th), Cory Stillman ( 6th), David Cooper (11th), Markus Naslund (16th), Turner Stevenson (12th) and Brad May (14th).
Seven of these twelve guys are still playing in the NHL, which means at least a ten-year career, and a pretty good pick. Two of these guys ended up being franchise quality players for a portion of their careers (Iginla, Naslund), and they weren't drafted in the top ten. Also seems pretty good. Of the five players not in the NHL, two of them had careers and in the case of Kasparaitis, a pretty good one.
I think that four of these players (Sykora, Iginla, Moreau and Naslund) had the best career of anyone drafted after them in the first round which seems like a decent if not overwhelming success rate.
Every one of these players played more than 0 games at the NHL level, which is more than you can say for a few players in each draft class, though it's not a great measure of success.
Now, I don't know if this totally debunks what you're saying, but looking at the six drafts from 1990-1995 the same logic doesn't seem to apply. Given that the highest pick was 5th and there are only three picks in the top ten, this seems like an above average group.
I guess people just got worse at drafting?
Scott.
Scott, well done. That absolutely undermines -- if not explodes -- my little theory, especially since you have to assume that back in those days, talent evaluation would have been (relatively speaking) less dependent on data and more dependent on simply watching the guys.
As to Lowetide's point -- though I think he was mainly raging at Andy, not me -- he's right. Sure, you can find busts no problem, but as Cosh notes there is a plenty strong correlation between draft position and quality of career.
All that said: here's a shiny penny that says all sorts of 1st-rounders -- Voracek, Couture, and Ellerby among them -- end up being better players than Gagner. If I was an NHL GM, I wouldn't touch a London Knight with a 10-foot Koho, at least not anywhere near the slot where they're ranked.
Does anyone know of a internet thingie that has up to date all time stats? Hockey db seems to go only to 2004 for NHL.
In terms of which stats? I know HockeyDB has been updating the year-to-year stuff fairly regularly throughout the season.
I'm looking forward to seeing Gagner play without Kane this year. If Gagner puts up similar numbers, perhaps he was the straw that stirred the Knights offence.
Forgot to add...Gagner made Team Canada at 17 for the World Jrs. His play aside, that's a huge feat in itself...obviously the coaches saw something they liked
Based on my own observations the overall noise doesn't start to dominate the signal until the end of the first round at earliest.
I never said anything against that. Jon did. I said if they couldn't make good picks in the first round, they might as well just pick names out of a hat.
He is asking you to support your contention that GMs "for the most part do a terrible job of drafting players."
My support is right here. It's a post by LT (and RiversQ) called "Saw Him Good."
Post a Comment
<< Home
Briefly looking at the players outlined I'm already doubtful of your proposed effect. I'd wager the underlying mediating factor between the first and second group is that the first group on average was drafted in the bottom half of the top 20, while the second group was on average drafted in the first half. As most who follow the draft will tell you, there is a generally substantial drop-off in the quality of an NHL prospect between those two ranges.
If you're somehow hoping that Gagner turns out to be a bust I would suggest you spend more time worrying about Calgary's own Swedish darkhorse from the same draft.
As most who follow the draft will tell you, there is a generally substantial drop-off in the quality of an NHL prospect between those two ranges.
I fail to see how this counters anything. The fact that there is a "generally substantial drop-off" in quality only supports the "Saw Him Good" theory. The whole point is that NHL GM's, for the most part, do a terrible job of drafting players. If GM's can't find quality in the 10-20 range of the first bloody round, they should just put all the names in a hat and randomly pick.
There's very little question that they should pick the highest rated player (composite scoring of major scouting services' ranks) or if there's a "tie" just throw a dart. There's no evidence that picking a Riley Nash is a "smart" pick under any circumstances. It's extremely unlikely for off-the-board picks to turnout better than all the players rated in front of them.
Drafting seems to be another big ego thing for management types. Ie- I know hockey better than you guys, nya nya. It's pretty hard to argue that anyone is consistently making steals in the first round.
I remember watching Brendl play for the Hitmen. God he was good. Too bad he never learned to play even the slightest bit of defense.
I honestly think people post stuff just to angry up the blood. I don't know what measure anyone wants to use, but I would like to challenge ANYONE to take a look at the NHL Entry Draft over any 10-year period and make an argument against this statement:
NHL teams generally do an excellent job drafting the most talented players early in a given draft.
Exception being the Russian/Euro Bure era stuff that buggers up the curve.
I don't know which point you are arguing against, LT. Are you saying GM's doing a good job of drafting talent, or not?
He is asking you to support your contention that GMs "for the most part do a terrible job of drafting players." After all, a player's draft rank is a fairly reliable estimator of his ultimate placing within his draft cohort; the draft is not a total crapshoot. #1s do fare better than #2s, #2s better than #3s, nos. 3-5 better than 5-10, 5-10 better than 10-15, etc. Based on my own observations the overall noise doesn't start to dominate the signal until the end of the first round at earliest.
the last time an NHL team hit the jackpot using a Top 20 pick to select the 2nd best draft-eligible player on a team was 1995, and no I'm not referring to Terry Ryan.
Petr Sykora?
I think that it's worth noting that going back to 1990 we add the following names to the list:
Terry Ryan (8th), Jarome Iginla (11th), Brad Church (17th), Petr Sykora (18th), Ethan Moreau (14th), Jason Allison (17th), Darius Kasparaitis (5th), Cory Stillman ( 6th), David Cooper (11th), Markus Naslund (16th), Turner Stevenson (12th) and Brad May (14th).
Seven of these twelve guys are still playing in the NHL, which means at least a ten-year career, and a pretty good pick. Two of these guys ended up being franchise quality players for a portion of their careers (Iginla, Naslund), and they weren't drafted in the top ten. Also seems pretty good. Of the five players not in the NHL, two of them had careers and in the case of Kasparaitis, a pretty good one.
I think that four of these players (Sykora, Iginla, Moreau and Naslund) had the best career of anyone drafted after them in the first round which seems like a decent if not overwhelming success rate.
Every one of these players played more than 0 games at the NHL level, which is more than you can say for a few players in each draft class, though it's not a great measure of success.
Now, I don't know if this totally debunks what you're saying, but looking at the six drafts from 1990-1995 the same logic doesn't seem to apply. Given that the highest pick was 5th and there are only three picks in the top ten, this seems like an above average group.
I guess people just got worse at drafting?
Scott.
# posted by 
: August 15, 2007 9:08 am
: August 15, 2007 9:08 am Scott, well done. That absolutely undermines -- if not explodes -- my little theory, especially since you have to assume that back in those days, talent evaluation would have been (relatively speaking) less dependent on data and more dependent on simply watching the guys.
As to Lowetide's point -- though I think he was mainly raging at Andy, not me -- he's right. Sure, you can find busts no problem, but as Cosh notes there is a plenty strong correlation between draft position and quality of career.
All that said: here's a shiny penny that says all sorts of 1st-rounders -- Voracek, Couture, and Ellerby among them -- end up being better players than Gagner. If I was an NHL GM, I wouldn't touch a London Knight with a 10-foot Koho, at least not anywhere near the slot where they're ranked.
# posted by : August 15, 2007 10:07 am
: August 15, 2007 10:07 am Does anyone know of a internet thingie that has up to date all time stats? Hockey db seems to go only to 2004 for NHL.
# posted by : August 15, 2007 11:49 am
: August 15, 2007 11:49 am In terms of which stats? I know HockeyDB has been updating the year-to-year stuff fairly regularly throughout the season.
I'm looking forward to seeing Gagner play without Kane this year. If Gagner puts up similar numbers, perhaps he was the straw that stirred the Knights offence.
Forgot to add...Gagner made Team Canada at 17 for the World Jrs. His play aside, that's a huge feat in itself...obviously the coaches saw something they liked
Based on my own observations the overall noise doesn't start to dominate the signal until the end of the first round at earliest.
I never said anything against that. Jon did. I said if they couldn't make good picks in the first round, they might as well just pick names out of a hat.
He is asking you to support your contention that GMs "for the most part do a terrible job of drafting players."
My support is right here. It's a post by LT (and RiversQ) called "Saw Him Good."
Post a Comment
<< Home
