I really like this idea. But I'm going to push back a little...presumably moving from one tier to the next implies a big gap in ability - how do you estimate the size of that gap? Is it ceiling and floor? And then probability of floor and ceiling? With computer models this is easier because the models spit out a number and we could cluster similar values (say values with 5 or 10 or 15 points of each other). But it's not clear to me how Trevor decides on 'gap size'.
It's a good point, Jeff. Appreciate the comment. I'm admittedly speaking for Trevor here and I can clarify with him. Knowing him, this is a subjective endeavour - the gaps are not necessarily enormous between tiers, but in evaluating, seeing shelves of prospects emerge, even if subtle in nature. I don't think it's specifically an exercise in floor and ceiling as that also varies prospect by prospect, and then we'd be getting into risk-reward analysis, which for sure factors in, but not the clear way these tiers are divided. I may have muddled the waters more than clarified them, admittedly! I think it's a cool ongoing, evolving discussion to be honest, and just another way to look at the prospect pools from year to year. I could see NHL squads having similar lists which helps them gauge when, if, and in what range they'd like to move up or down (or stay put) in the draft (and then value of draft picks analysis also factors in). Live diving in to such things, so feel free to pick my reply apart and add further thoughts. Cheers.
Hey Chris, I think you're right that NHL scouting teams do this kind of tiering. I had little interaction with the scouting teams in either Calgary or Ottawa when I worked for them, but a little more with Calgary than Ottawa. And in 2007 when they were choosing at #24 they had a group of 5 or 6 players (I don't know which players) who they thought were similar prospects and they would choose from whoever was still available from that group. They ended up taking Mikael Backlund. Presumably, they had the players ranked within that tier...but the way they talked about it was that it didn't really matter who they got from that group.
But I'm always a bit skeptical when analysts of any kind split 'objects' into groups but can't explain to me how they do it. I don't have a problem with subjective - all that means is that you might do it differently from me. But if you can't explain to me what criteria you use to group the objects, then I don't trust that you have a process that is repeatable...even for you. Monday you might group them one way and Tuesday another. And that has always seemed like a problem to me.
So, if Trevor was to say
Tier 1 - top 3 forward or top pair D with 80% probability.
Tier 2 - top 6 forward top 4 D with 80% probability.
Tier 3 - Middle six or ¾ D with 80 % probability
...and so on.
That would make sense to me. Subjectivity could arise in two forms. First, I accept Trevor's criteria but might disagree about who goes in each tier based on our subjective opinions about players. Or second, I use different criteria for the tiers, but could tell you what they were.
But in either case, you would understand my thinking. At this point Trevor's thinking is opaque to me.
I've never bought the "I can't define pornography, but I know it when I see it..." school of thought. So, I don't buy "I can't define gaps, but I know them when I see them...". But my guess is that Trevor could already tell me his criteria...or if not, could tell me after giving it a few minutes thought.
Oh for sure! I can explore it with him as well. He's completed over 250 prospect profiles, so absolutely, in his mind, he has it mapped out why prospects are in the various tiers/shelves/levels. It's a neat exploration to ponder. It's crunch time for us getting ready for our Draft Guide with not a ton of spare time, so if not now, then the offseason, it'll be something to dive deeper into. Cheers! Really appreciate you weighing in on these posts. I've been following the draft in-depth for decades and still looking to further develop, progress. Keep in touch.
I really like this idea. But I'm going to push back a little...presumably moving from one tier to the next implies a big gap in ability - how do you estimate the size of that gap? Is it ceiling and floor? And then probability of floor and ceiling? With computer models this is easier because the models spit out a number and we could cluster similar values (say values with 5 or 10 or 15 points of each other). But it's not clear to me how Trevor decides on 'gap size'.
It's a good point, Jeff. Appreciate the comment. I'm admittedly speaking for Trevor here and I can clarify with him. Knowing him, this is a subjective endeavour - the gaps are not necessarily enormous between tiers, but in evaluating, seeing shelves of prospects emerge, even if subtle in nature. I don't think it's specifically an exercise in floor and ceiling as that also varies prospect by prospect, and then we'd be getting into risk-reward analysis, which for sure factors in, but not the clear way these tiers are divided. I may have muddled the waters more than clarified them, admittedly! I think it's a cool ongoing, evolving discussion to be honest, and just another way to look at the prospect pools from year to year. I could see NHL squads having similar lists which helps them gauge when, if, and in what range they'd like to move up or down (or stay put) in the draft (and then value of draft picks analysis also factors in). Live diving in to such things, so feel free to pick my reply apart and add further thoughts. Cheers.
Hey Chris, I think you're right that NHL scouting teams do this kind of tiering. I had little interaction with the scouting teams in either Calgary or Ottawa when I worked for them, but a little more with Calgary than Ottawa. And in 2007 when they were choosing at #24 they had a group of 5 or 6 players (I don't know which players) who they thought were similar prospects and they would choose from whoever was still available from that group. They ended up taking Mikael Backlund. Presumably, they had the players ranked within that tier...but the way they talked about it was that it didn't really matter who they got from that group.
But I'm always a bit skeptical when analysts of any kind split 'objects' into groups but can't explain to me how they do it. I don't have a problem with subjective - all that means is that you might do it differently from me. But if you can't explain to me what criteria you use to group the objects, then I don't trust that you have a process that is repeatable...even for you. Monday you might group them one way and Tuesday another. And that has always seemed like a problem to me.
So, if Trevor was to say
Tier 1 - top 3 forward or top pair D with 80% probability.
Tier 2 - top 6 forward top 4 D with 80% probability.
Tier 3 - Middle six or ¾ D with 80 % probability
...and so on.
That would make sense to me. Subjectivity could arise in two forms. First, I accept Trevor's criteria but might disagree about who goes in each tier based on our subjective opinions about players. Or second, I use different criteria for the tiers, but could tell you what they were.
But in either case, you would understand my thinking. At this point Trevor's thinking is opaque to me.
I've never bought the "I can't define pornography, but I know it when I see it..." school of thought. So, I don't buy "I can't define gaps, but I know them when I see them...". But my guess is that Trevor could already tell me his criteria...or if not, could tell me after giving it a few minutes thought.
Best Jeff
Oh for sure! I can explore it with him as well. He's completed over 250 prospect profiles, so absolutely, in his mind, he has it mapped out why prospects are in the various tiers/shelves/levels. It's a neat exploration to ponder. It's crunch time for us getting ready for our Draft Guide with not a ton of spare time, so if not now, then the offseason, it'll be something to dive deeper into. Cheers! Really appreciate you weighing in on these posts. I've been following the draft in-depth for decades and still looking to further develop, progress. Keep in touch.