NHL Draft Software: 2018 Top100 End of the regular season rankings

[Ainec]

well done

best ranking I've seen

 

[Andrei79]

Impressive work my friend.


Sucks for all of us that you'll be in the next mock, but now I can steal all your picks.

 

[blindpass]

I can see how it can be difficult to grasp how I am using the subjective evaluations so here are further explanations; I rate players attributes which is converted into an overall score, this score is sort of used as an expected value or you could say a barometer to remove anomalies in statistical performance (if the player overperformed or underperformed base on what you would expect). Those evaluations are also seen by the software sort as a projection of the player skills, affecting the final score again at the end, you can often see players with good stats but bad skating fall in the draft, so I particularly value skating so the software can match that. So it matters that I watch the players in order to have the appropriate subjective evaluations. If you were to be using the software you might obtain different results, but it should normally tend to look like mine.

Is there a reason why you think subjective opinion + "software" is superior to subjective opinion alone?

 

[37 others]

I have been entering everything myself, i'll be looking into EP's API/a webscraper it could save me a lot of time.

Nice to see people sharing my interest, I believe there is true potential I that type of approach. I am not sure I fully understood your last question; I just started out less than a year ago on my own sort of as challenge I set myself (thinking I could have better scouting results with a well programmed software) and here I am now. Do you mean how did I achieve those type of results with my algorithm?

No, just about how you got into it in general. I've used EP's API before and I'd be happy to help you out.

 

[Ainec]

Is there a reason why you think subjective opinion + "software" is superior to subjective opinion alone?

all the other mocks you will see are subjective opinions

you will see that a lot of them do not deviate far from the monthly public lists and final rankings..

 

[ProspectsFanatic]

Is there a reason why you think subjective opinion + "software" is superior to subjective opinion alone?

Things I stated in the past in that regard to people I entered in contact with:

"I firmly believe that this software can be a useful tool to support the evaluation of prospects, in fact, I would even go as far as to say that it has the potential to give better results than the standard way of evaluating players. When I started doing the exercise of creating my own rankings of draft eligible players I quickly realize how complex of a mental process it is considering all the variables of evaluation that have to be taken into account when evaluating a player (speed, skill, shot, defensive game, compete level, age, offensive production, quality of teammates, league difficulty, ice time, height, weight, tournament play, etc, etc, while assessing to each criteria their degree of importance), inevitably some criteria of evaluation become overvalued or undervalued. Once you start comparing two players the mental process increase in complexity, recalling every relevant information about each player, gaging each gap between the players and assessing which accumulation of gaps outweigh the other, the increase in complexity inevitably creates more room for misjudgment. At that point, if two scouts are arguing about two prospects a fight of will must take over to a certain extent since the analysis of the objective subtleties that differentiate the players can't be addressed in his totality. As a matter of fact, we face ourselves against the limitation of our brain even when isolating simple variables easy to grasp, even the ones who are already objectively defined for us can lead to misevaluation, let’s take the age of player for example, the brain can't associate a precise value for each day of the year, instead, for the sake of simplification, the brain will tend to reduce the number of values by regrouping to a single value weeks or entire month or possibly closer months with one another. This can be seen as a marginal difference that might not have a considerable impact, but those marginal differences can be found everywhere, they add up and their sum can make noticeable differences, even more so if you have a bias towards a specific player your brain will unconsciously selectively favored the regrouping of values in a manner that will confirm your bias opinion. But most importantly, this software eases the process of evaluation by isolating each variable of evaluation in order to come to a final value, the brain performs much better at assessing the most precise value possible to one single variable at the time than taking everything into account at once in order to come to a final value. For those reasons, I firmly believe that if the exercise of assessing the most exact value possible to each isolated variable is done properly with the help of a software it will undeniably give results of greater accuracy."

"The human brain has limitation in regards to his precision (exact value attributed to a prospect playing in a certain league at a certain point in time considering his specific day of birth), consistency in evaluation (I give the exact same value to that player as for each other player in that same context, if it is a slightly different context the value is slightly altered proportionally) and processing power (quantity of variables factored in to precision at the same time, looking at all that data at once), those are three major limitations which software does not have, I believe that there is potential there to be exploited in the scouting hockey world."

 

[FrankGallagher]

Excellent concept, this is the most impressive thing I have seen on here. The eye test alone and stats alone both have limitations, putting them together covers a lot of blindspots each has.

 

[ProspectsFanatic]

No, just about how you got into it in general. I've used EP's API before and I'd be happy to help you out.

Well, yea, like I said earlier, I started one day just for the sake of trying, there isn't much more to it, I like those type of mental challenge, I don't have a background in programing, but I always had ease with numbers. I will look into API for EP if I need help I let you know, thank you for the offer.

 

[AveryStar4Eva]

If I even own an NHL team I'm hiring you as my head scout. Awesome list with tons of work put into it

 

[blindpass]

Things I stated in the past in that regard to people I entered in contact with:

"I firmly believe that this software can be a useful tool to support the evaluation of prospects, in fact, I would even go as far as to say that it has the potential to give better results than the standard way of evaluating players. When I started doing the exercise of creating my own rankings of draft eligible players I quickly realize how complex of a mental process it is considering all the variables of evaluation that have to be taken into account when evaluating a player (speed, skill, shot, defensive game, compete level, age, offensive production, quality of teammates, league difficulty, ice time, height, weight, tournament play, etc, etc, while assessing to each criteria their degree of importance), ...

I can see the value of keeping detailed notes as an aid to memory. As much as putting a concrete number for a few subjective dimensions helps you hold on to information you might forget, it necessarily glosses over intangibles and subtleties as well. Forcing all the numbers through some ad hoc formulas (correct me if they are more than that) doesn't seem better to me than going with an expert's subjective opinion in the first place.

That said your rankings seem as good as the next guys.

 

[blindpass]

If I even own an NHL team I'm hiring you as my head scout. Awesome list with tons of work put into it

Apparently the snake oil impresses some.

 

[ProspectsFanatic]

I can see the value of keeping detailed notes as an aid to memory. As much as putting a concrete number for a few subjective dimensions helps you hold on to information you might forget, it necessarily glosses over intangibles and subtleties as well. Forcing all the numbers through some ad hoc formulas (correct me if they are more than that) doesn't seem better to me than going with an expert's subjective opinion in the first place.

That said your rankings seem as good as the next guys.

Personally I can't analyze stats with that much precision on my own; (1 row is 1 player)

and this is only for one league during one year, so this is like only the 1/10th of the maths. Not sure aid memory and some concrete numbers help you account for everything with that much precision. I am sure there is some algorithm in the technology you use that help you do things more effectively in your day to day life.

 

[Poppy Whoa Sonnet]

Can you compare TOTAL scores across years? Or is it meant to be an index showing relative value within each year's draft?

 

[snipes]

Fantastic work, nice to see the algorithm is as high on Wahlstrom as I am and low on Tkachuk as well.

Interesting the hit Zadina takes in your software.

 

[ProspectsFanatic]

Can you compare TOTAL scores across years? Or is it meant to be an index showing relative value within each year's draft?

Theoretically you could, but since I am always fine-tuning the algorithm the comparison isn't perfect. With a brief overview, outside the top2 last year seems better though.

 

[ProspectsFanatic]

Fantastic work, nice to see the algorithm is as high on Wahlstrom as I am and low on Tkachuk as well.

Interesting the hit Zadina takes in your software.

I think I figured out the main reason why Zadina is ranked lower than on most rankings (many people have him at #3 even challenging #2, and I have Zadina at #6). When people think of Zadina they mostly remember WJC20 tournament where he was dominant, thats when they saw him play, so consequently, they probably account that performance for a good chunk of his overall evaluation. The software only accounts his WJC20 play for around 20% of his draft eligible year score and maybe 12% if you also account the 2 years before. The reality is that he didn't really match that level of play in the QMJHL or Czech leagues, in regards to the QMJHL his stats are solid, but you would expect more out of a 3rd overall being born in November, he wasn't even really able to outproduce a teammate like Somppi. Otherwise, scouting wise, very good edge work, but lacking acceleration/top speed gives further argument to position him below 3rd overall.

To sum it up, to me by looking at the broader picture I feel the software is making the right read, people might make a flawed assessment of Zadina by basing their evaluation too much on how he performed in the WJC20. That being said, I can see people having Zadina as high as #3 (he isn't that far from #3 base on the software results), but I don't think there should be any discussion between Svechnikov and Zadina, Svechnikov is clearly the superior player in my view.

 

[JK2K]

Information is a useful tool.
Depending too much on this stuff gets you where Florida and Arizona are today.
I think some teams use it because other teams are and don’t want to seem like they are missing anything.
Give me an old scout who can see the intangibles in a player over some data filled software package any day.

 

[ProspectsFanatic]

Information is a useful tool.
Depending too much on this stuff gets you where Florida and Arizona are today.
I think some teams use it because other teams are and don’t want to seem like they are missing anything.
Give me an old scout who can see the intangibles in a player over some data filled software package any day.

I also see it as a tool too, the algorithm can't cover everything and always make the perfect estimate, I believe you should thoroughly revisit your judgement of players when the software indicates the opposite by a fair margin, but when the players are close to one another you should go with your gut feeling, since the imperfection of the software may have skewed the results in the wrong way. That being said, this software doesn't only look at numbers it also factors in subjective evaluations made in scouting, so it should normally tend to be close to what you already think. Finally, I wouldn't generalize too much Florida/Arizona, it is no accident that the NHL is trending towards more and more analytics.

 

[JK2K]

Brian Burke said something like “ show me analytics that measure a player will to block a shot with his face”
I agree with him. Numbers are good, but on the periphery.

 

[Superlative Soup]

2017
5. Elias Pettersson
33. Nicolas Hague
55. Josh Brook
64. Max Gildon
95. Ostap Safin
135. Calle Sjalin
181. Alexander Chmelevski
188. Pavel Shen

2016
5. Matthew Tkachuk
64. Vitaly Abramov
140. Tim Gettinger
154. Jesper Bratt
184. Stepan Falkovsky
194. David Bernhardt

2015
23. Boeser (Thats an important one, easy to say you called it, I actually found my old crappy list https://bit.ly/2H0Y5X9 ; no one drafted below him is in front of him on my list. List matching my MTL picks having Sprong and Kylington high)
66. Filip Ahl
114. Dmytro Timashov (I think)
144. Cooper Marody
149. Nikita Krorostelev
174. Gustav Bouramman (I think)
210. Nathan Noel
Thats probably the year I followed the less for the later rounds ending up with pretty bad results, not fully sure about the selections though.

2014
6. Nikolaj Ehlers (damn, most people would have went Ehlers or Nylander to be fair I believe)
24. David Pastrnak (oh god)
36. Roland Mckeown
66. Brayden Point (ok that year would have been insane)
126. Ondrej Kase (wait... wasn't over)
156. Adam Ollas Mattsson
186. Vladimir Tkachev

2013
9. Valeri Nichushkin (sorry for that)
24. Hunter Shinkaruk (same mistake)
85. Pavel Buchnevich
115. Eric Roy (damn you missed the good ones, late 3rd Bjorkstrand & Lehkonen with 2nd)
145. Lucas Wallmark
175. Nikita Tryamkin (got him 1 year before!)
205. Brendan Harms

Did not follow that draft enough before that, but your former Vancouver Giants Gallagher would be on your team for sure (but I may have messed up some other picks tho who knows?).

Pastrnak-Point-Boeser
Ehlers-Buchnevich-Tkachuk
Bratt-Kase-Pettersson

Not bad , trade whatever extra for a D. You got 3 good Ds in the early rounds of 2017, thats a good bank to start securing the future too.

That'll drive a guy to drink.

 

[blindpass]

Information is a useful tool.
Depending too much on this stuff gets you where Florida and Arizona are today.
I think some teams use it because other teams are and don’t want to seem like they are missing anything.
Give me an old scout who can see the intangibles in a player over some data filled software package any day.

Don't mistake this for serious analytics or machine learning! That can't replace expert judgement yet, but used right it is a great complement to it.

This isn't that, from what I can see.

Subjective ratings are being fed into an ad hoc formula and being tweaked when the results aren't "right". It is just a roundabout way of doing subjective rankings, which is why it looks an awful lot like the consensus rankings. There's no reason to think that where it deviates that it is on to something.

 

[ProspectsFanatic]

Don't mistake this for serious analytics or machine learning! That can't replace expert judgement yet, but used right it is a great complement to it.

This isn't that, from what I can see.

Subjective ratings are being fed into an ad hoc formula and being tweaked when the results aren't "right". It is just a roundabout way of doing subjective rankings, which is why it looks an awful lot like the consensus rankings. There's no reason to think that where it deviates that it is on to something.

If I understand correctly my work is just a big waste of time; creating complex formulas and entering all that data, to only in the end, end up tweaking the results to match the consensus rating. Seeing that your last 5 posts on this forum have been dedicated to removing any form of relevancy to my work I doubt you will consider what I am about to say, but I will say it anyway; maybe, just maybe... it looks an awful lot like the consensus rankings because when people evaluate a prospect they look at his stats and values it, coming to similar conclusion to the software, and also maybe the evaluation I entered also matches other people evaluation of prospects, so in the end, we end up we similar results. Crazy idea right? Saying it is awful lot like the consensus is just a false claim to begin with so I shouldn't need to defend myself; tell me who has Jenik and Shafigullin in the first round? Have you ever heard of Wassenius before? Well, he is 63 on my list. I never saw Kravtsov in the top10 or Tkachuk outside of it before my list went up. Zadina at 6 is quite low too, and I knew many would disagree, I wouldn't naturally place Zadina there, but this is where he ended up, I try to make sense why the software gives those results subsequently, because I trust the software and I think it is well programmed.

The subjective ratings aren't there to tweak the results they are there to attempt to mimic the overall process of scouting evaluation and take everything into consideration. It is impossible to know a player defensive game/competitiveness purely by looking at stats, I don't have a choice to add this data in a form of subjective evaluation to have a complete overall picture. There is no other way around it. Skating is very important for projection too. Saying I am using this to cheat out everything without any proof of it is unfair and malicious.

Please go find someone else work to discredit I have heard enough of you here.

 

[tony d]

Thanks for posting this, appreciate the effort that went into this.

 

[FlameChampion]

Do you know what players like McDavid, Matthews, Eichel, Laine would of ranked using your scale? Just curious for comparison sakes.

 

[emptyNedder]

First--yours is my favorite of all the "top prospects" lists on HF. Specifically because you use data.

I was glad to see you have Wassenius as a high 3rd rounder. I have mentioned him several times. And yes, I am a stat watcher. Because stats are objective. In fact, my only question to you would be related to this.

It is impossible to know a player's defensive game/competitiveness purely by looking at stats, I don't have a choice to add this data in a form of subjective evaluation to have a complete overall picture.

I agree on the aspects you noted and really like that you take strength of league and line mates into consideration. Those both make a big difference. Where I am not so sure is in the "subjective" measures of skating, shooting, puck control, and playmaking. Because the difference in any of those individually or all combined must surely show up in goals/assists. Over any season of 30 or more games luck should be mitigated. With luck minimized, an excellent skater with a very good shot will score more goals; an exceptional puck handler with excellent playmaking ability will have more assists. So either the subjective measures become redundant or introduce bias to explain why players that scouts really like aren't more productive.

Again I really like that you have created a model and I think it will consistently outperform subjective rankings. I think you might "trust" it a little more by letting the data stand on their own.