The Data-Based Drafting Thread (what players would a Potato pick?)

[Hollywood Burrows]

There's a chunk of posters on this forum who need to be force-fed Melvin's work, clockwork orange style. I volunteer to hold their eyes open.

 

[Melvin]

It might be worth noting that Boqvist production is very unequal (for a d-man) in terms of goals vs assists.

Someone pointed out on another board that Boqvist had the highest goals/gm ever, for a U18 d-man in the Super Elite J20 league. On the other hand his assist numbers were extremely pedestrian.

14 goals and 10 assists (25 games) is an extremely odd stat line for a d-man in any league. People can infer whatever they wish out of that, but I see his assist totals as something of an anomaly. Usually any type of high impact offensive defenceman is going to have as many assists as goals no matter how in love with their own shot they are. Given his shot, and the obvious lack of team-mates who were on his level- i'm also guessing his coach told him to shoot at every opportunity. All in all, lack of production only really applies to Boqvist's pro numbers imo.

That's an interesting point and not something I have looked into. Definitely worth the research after the draft. Thanks!

 

[krutovsdonut]

melvin

maybe you did this somewhere, but i'd be curious to see combined rankings putting multiple years of top 20 picks in the same sort set to see how the potato compares them.

one of my issues with this draft is whether the number of dmen ranked highly who i think have question marks reflects a lack of top forwards. i'd like to see how the forwards in this draft compare to years past.

i still think the most interesting question the potato poses for me is whether it would be smarter to focus on chasing high scoring players while drafting and accepting you will get a lot of forward prospects, and then to try to trade some for other positions.

 

[Melvin]

melvin

maybe you did this somewhere, but i'd be curious to see combined rankings putting multiple years of top 20 picks in the same sort set to see how the potato compares them.

one of my issues with this draft is whether the number of dmen ranked highly who i think have question marks reflects a lack of top forwards. i'd like to see how the forwards in this draft compare to years past.

i still think the most interesting question the potato poses for me is whether it would be smarter to focus on chasing high scoring players while drafting and accepting you will get a lot of forward prospects, and then to try to trade some for other positions.

I am not sure precisely what you are looking for but this is what the top-ten forwards looks like.

Elias Pettersson
Connor McDavid
Elias Lindholm
Patrick Kane
Mikael Granlund
Auston Matthews
Aleksander Barkov
Nicklas Backstrom
Sam Gagner
Patrik Laine

No forward in 2018 is close to this group. Svechnikov ranks - people will love this - right in between Nail Yakupov and Ryan Strome, around 30th in my dataset.

It is just not a good draft for point-producing forwards.

On the other hand, both Dahlin and Bouchard rank highly among the defenders:

Victor Hedman
* Rasmus Dahlin
* Evan Bouchard
Oliver Ekman-Larsson
Dmitry Kulikov
Ryan Ellis
Zach Bogosian
Cam Fowler
Adam Larsson
Juuso Valimaki

Having said that I think the algorithm sort of breaks at the extremes. I think it is more useful at finding players in the later rounds than it is at comparing the elite of the elite.

 

[Peter10]

Having said that I think the algorithm sort of breaks at the extremes. I think it is more useful at finding players in the later rounds than it is at comparing the elite of the elite.

Maybe a combination works best. For the first round use the concencus from all the draft guide rankings and then in the 2nd round continue with Potato

 

[krutovsdonut]

I am not sure precisely what you are looking for but this is what the top-ten forwards looks like.

No forward in 2018 is close to this group. Svechnikov ranks - people will love this - right in between Nail Yakupov and Ryan Strome, around 30th in my dataset.

It is just not a good draft for point-producing forwards.

thanks, this is the detail i was looking for. it is what i was sensing. the positioning next to yakupov is a bit unfair, but it does match my sense that this is a thin draft at the top for forwards.

how many franchise forwards does the potato miss i wonder? kind of subjective, but i doubt many elite players slip through.

 

[Melvin]

thanks, this is the detail i was looking for. it is what i was sensing. the positioning next to yakupov is a bit unfair, but it does match my sense that this is a thin draft at the top for forwards.

how many franchise forwards does the potato miss i wonder? kind of subjective, but i doubt many elite players slip through.

One player who slipped through would be Brock Boeser. He is an interesting case as a player whose stats were pretty meh, and whose skating is not great either. That pick is looking like a pretty clear scouting win right now.

 

[Melvin]

Maybe a combination works best. For the first round use the concencus from all the draft guide rankings and then in the 2nd round continue with Potato

I mean one of the things you would need to be good at is predicting whether or not a player would be available for a later selection. If the highest ranked player is a player you expect to be able to select with a later pick then you obviously don't want to take him too early. You can pretty easily improve upon the results by being mindful of this.

But I wouldn't dismiss its first round selections that easily. Look at its picks for the canucks in the first round compared to who we actually took..

 

[bossram]

It is definitely easier to pick forwards in this way than defenders, that's for sure. Although this methodology makes it impossible to blow a 3rd overall on someone like Gudbranson, it will also not be able to find someone like Hampus Lindholm. It's a trade-off, of course.

Well, the Potato drafting system looks clearly superior anyway. With that many elite forwards, you could just trade for defense.

 

[Melvin]

Based on Bob's list, I am expecting our first three picks to be Bouchard, Hallander and Nathan Dunkley.

 

[Melvin]

After Round 1 ....

1. Rasmus Dahlin
2. Evan Bouchard
3. Andrei Svechnikov
4. Jesperi Kotkaniemi
5. Isac Lundeström
6. Filip Zadina
7. Filip Hållander
8. Noah Dobson
9. Jacob Olofsson
10. Oliver Wahlstrom
11. Sean Durzi
12. Martin Kaut
13. Ryan Merkley
14. Akil Thomas
15. Joe Veleno
16. Nicolas Beaudin
17. Jonatan Berggren
18. David Gustafsson
19. Rasmus Kupari
20. Rasmus Sandin
------------------------------------

Canucks picks as drafted by the potato:

7 - Evan Bouchard

 

[Melvin]

Potato picks for Vancouver -

7 - Evan Bouchard (Quinn Hughes)
37 - Filip Hallander (Jett Woo)
68 - Justin Almeida (Tyler Madden)
130 - Linus Nyman (Toni Utunen)
186 - Nathan Dunkley (Artyom Manukyan)
192 - Kyle Topping (Matthew Thiessen)\

 

[bossram]

Potato picks for Vancouver -

7 - Evan Bouchard (Quinn Hughes)
37 - Filip Hallander (Jett Woo)
68 - Justin Almeida (Tyler Madden)
130 - Linus Nyman (Toni Utunen)
186 - Nathan Dunkley (Artyom Manukyan)
192 - Kyle Topping (Matthew Thiessen)\

Pretty interesting. I'd rather have Hughes than Bouchard, but taking Bouchard at 7 is still pretty defensible.

I think it's more than likely that the Potato comes out even or better compared to the rest of Vancouver's actual picks though. That's saying something.

 

[Phenomenon13]

I am not sure precisely what you are looking for but this is what the top-ten forwards looks like.

Elias Pettersson
Connor McDavid
Elias Lindholm
Patrick Kane
Mikael Granlund
Auston Matthews
Aleksander Barkov
Nicklas Backstrom
Sam Gagner
Patrik Laine

No forward in 2018 is close to this group. Svechnikov ranks - people will love this - right in between Nail Yakupov and Ryan Strome, around 30th in my dataset.

It is just not a good draft for point-producing forwards.

On the other hand, both Dahlin and Bouchard rank highly among the defenders:

Victor Hedman
* Rasmus Dahlin
* Evan Bouchard
Oliver Ekman-Larsson
Dmitry Kulikov
Ryan Ellis
Zach Bogosian
Cam Fowler
Adam Larsson
Juuso Valimaki

Having said that I think the algorithm sort of breaks at the extremes. I think it is more useful at finding players in the later rounds than it is at comparing the elite of the elite.

I find it insane that svechnikov is ranked alongside such disappointments. He's ranked a lot lower than I thought he would be. Is there a reason for this? He put up almost a goal per game no?

How does he compare to someone like Seguin or hall?

 

[David Bruce Banner]

Very enlightening to read and follow, Melvin. Thanks for all the hard work. I'll be interested in seeing how the Potato 2018 draft pans out.

I highly recommend that some of the stat heads here to read Leonard Mlodinow's awesome and engaging book The Drunkards Walk: How Randomness Rules Our Lives, for further insight on Potato drafting, dart throwing monkeys and how we often misinterpret data because humans really like to see patterns when sometimes there are none...

People systematically fail to see the role of chance in the success of ventures and in the success of people like the equity- fund manager Bill Miller. And we unreasonably believe that the mistakes of the past must be consequences of ignorance or incompetence and could have been remedied by further study and improved insight. That's why, for example, in spring 2007, when the stock of Merrill Lynch was trading around $95 a share, its CEO E. Stanley O'Neal could be celebrated as the risk-taking genius responsible, and in the fall of 2007, after the credit market collapsed, derided as the risk-taking cowboy responsible — and promptly fired.

 

[Melvin]

Very enlightening to read and follow, Melvin. Thanks for all the hard work. I'll be interested in seeing how the Potato 2018 draft pans out.

I highly recommend that some of the stat heads here to read Leonard Mlodinow's awesome and engaging book The Drunkards Walk: How Randomness Rules Our Lives, for further insight on Potato drafting, dart throwing monkeys and how we often misinterpret data because humans really like to see patterns when sometimes there are none...

I read this book a few years ago. It's an excellent recommendation. I should give it another read.

 

[Melvin]

I find it insane that svechnikov is ranked alongside such disappointments. He's ranked a lot lower than I thought he would be. Is there a reason for this? He put up almost a goal per game no?

How does he compare to someone like Seguin or hall?

Well those disappointments were ranked pretty highly too. Yakupov and Svechnikov have basically the same point totals in the same league.

Hall and seguin both had 30+ more points in a dozen more games.

I think players whose goal totals are more impressive than their point totals tend to be a little overrated (see Jake Virtanrn,) but we will see. He is still highly regarded, just not the stand out #2 that scouts have him.

 

[Melvin]

Now that the draft is complete, here are some things I'm planning to look into this summer:

1) Weighting Goals vs. Assists for defensemen vs forwards.
2) Adding D-1 stats to see if incorporating more years of data leads to better picks.
3) Continuing to look into BMI research.

Does anyone know of any books I can buy for the older drafts that I might be able to get information from that better reflects at-the-time data? I have reason to believe some of the information on NHL.com is not accurate for older drafts.

Looking for something like this but if I could get pre-2010 that would be fantastic.

Also, if you have any other research topic for me throw them my way.

 

[Phenomenon13]

After Round 1 ....

1. Rasmus Dahlin
2. Evan Bouchard
3. Andrei Svechnikov
4. Jesperi Kotkaniemi
5. Isac Lundeström
6. Filip Zadina
7. Filip Hållander
8. Noah Dobson
9. Jacob Olofsson
10. Oliver Wahlstrom
11. Sean Durzi
12. Martin Kaut
13. Ryan Merkley
14. Akil Thomas
15. Joe Veleno
16. Nicolas Beaudin
17. Jonatan Berggren
18. David Gustafsson
19. Rasmus Kupari
20. Rasmus Sandin
------------------------------------

Canucks picks as drafted by the potato:

7 - Evan Bouchard

I find it very interesting that players like Hughes and Boqvist aren't in the top 20.


Where does a player like Jake Wise or Bode Wilde fit? I feel that the USHL is a little under represented in general.

It is also interesting that players like Lundestrom, Hallander and Olofsson are rated so highly. They play in Mens Leagues but are seen as more of the two way type of players with less offensive upside than Wahlstrom/Berggren. Do you feel it overvalues this type of player and goes for more "safe" picks than swing for the fences?

 

[Melvin]

I find it very interesting that players like Hughes and Boqvist aren't in the top 20.


Where does a player like Jake Wise or Bode Wilde fit? I feel that the USHL is a little under represented in general.

It is also interesting that players like Lundestrom, Hallander and Olofsson are rated so highly. They play in Mens Leagues but are seen as more of the two way type of players with less offensive upside than Wahlstrom/Berggren. Do you feel it overvalues this type of player and goes for more "safe" picks than swing for the fences?

I have received a lot of comments that the USHL and USDP are under-represented but the numbers just don't bear it out. Perhaps it has changed very recently but it will take a few years for that to show up in the data if it's true. Bode Wilde's numbers are comparable to that of Tyler Motte, Sean Malone, Kevin Lynch and Troy Terry. Overall he was ranked around the 50's. Wise was higher, around 30 or so.

I don't know if it overvalues so-called "safe" players but I will say that it is set up in such a way where it is primarily designed to find NHL players and doesn't place a ton of emphasis on finding elite players vs. finding more complementary players. I use NHL time-on-ice as my primary success measure so it could potentially overrate players who play a lot but aren't considered elite. I have thought about experimenting with this in such a way where your Tyler Mottes are considered to be more of a failure and such I could skew the model to favor the more boom-or-bust types but I am not sure this is a good strategy. I suspect that simply trying to find NHL players with every pick and assuming you get elite players with your top picks and non-elite regulars with your other picks is not going to get you a substantially different profile.

 

[vancityluongo]

I've said this to Melvin before, but I think even though it does heavily overweight just making the NHL as successful, that's essentially the point, no? It's a baseline model.

No would say that Ashton Sautner is or ever has been a better prospect or hockey player than Hunter Shinkaruk, but the former may end up playing more NHL games. So if the next Hunter Shinkaruk-like prospect can't provide more NHL utility than Ashton Sautner, then he's not a worthwhile player to draft at the baseline level. The reward of drafting a Shinkaruk and having them pan out is obviously much, much higher, but this quantifies some of the risk of doing so. I think it's brilliant.

 

[Melvin]

I've said this to Melvin before, but I think even though it does heavily overweight just making the NHL as successful, that's essentially the point, no? It's a baseline model.

No would say that Ashton Sautner is or ever has been a better prospect or hockey player than Hunter Shinkaruk, but the former may end up playing more NHL games. So if the next Hunter Shinkaruk-like prospect can't provide more NHL utility than Ashton Sautner, then he's not a worthwhile player to draft at the baseline level. The reward of drafting a Shinkaruk and having them pan out is obviously much, much higher, but this quantifies some of the risk of doing so. I think it's brilliant.

Yes, exactly. If one can out-perform the baseline by going more boom-or-bust then that needs to be demonstrated.

 

[Tobi Wan Kenobi]

Pretty interesting. I'd rather have Hughes than Bouchard, but taking Bouchard at 7 is still pretty defensible.

I think it's more than likely that the Potato comes out even or better compared to the rest of Vancouver's actual picks though. That's saying something.

lol...I think it's time to start trusting Judd Brackett dont you think? Give me a break

 

[Melvin]

lol...I think it's time to start trusting Judd Brackett dont you think? Give me a break

The entire point of this thread is to measure the extent to which we should be "trusting" him.

 

[Phenomenon13]

Yes, exactly. If one can out-perform the baseline by going more boom-or-bust then that needs to be demonstrated.

I would suggest adjusting your model to weigh more heavily for boom or bust types in the first two rounds of the draft and then using your current alogirthmn. I feel your algorithm would outperform a typical draftingbgm either way. I'm not sure how the thresholds or numbers would work out but that's just my opinion.


I think a more offensively talented guy like wahlstrom hitting their potential could be worth perhaps two or three hallanders meeting his. Again, it's riskier but the benefits are much greater. It's a fine line.