01-04-2022, 10:33 PM(This post was last modified: 01-04-2022, 10:36 PM by hhh81.)
An Introduction: Who am I? What is this? What can you expect?
For the last few seasons, I’ve been working with a scoring system to narrow the field for a few SHL Awards. Initially it was just the Bojo Biscuit, for best defensive defenseman, then I expanded it to the Jeff Dar (best two-way forward) and then developed/refined something for the Scott Stevens (best defenseman) in S61. I felt like some things got lost in the shuffle when I didn’t use a more objective* first round of analysis. A lot of people fixate on a guy or two who jump out to them when first looking over the index, and then let their decisions all go from that first person they liked. I wanted something more consistent for myself, that made sure I was considering all potential options at a base level, before diving a little deeper into the context to choose my finalist nominees.
*There really aren’t many fully objective measures when it comes to this job; I made some judgement calls initially and keep doing so. I don’t claim to be an expert, but I think I’m pretty good at awards at this point haha.
I feel really good about my process for nominating people for those awards. Over the weekend, I thought, “wow, I should pass this along to drokeep and SMJHL Awards group. I’m not sure what sort of process they are using, and they might find some value in an approach like this, especially with the more defense-focused awards (Metzler and Holmes). I know dro is new, so I wonder how he’s approaching leading the team.”
Then I saw the S62 Metzler nominees, and that greatly sped up my process of sharing.
I thought it might be an interesting media for me to apply what I’ve been calling the “Bojo score” to the SMJHL this season and see if I came to the same conclusions on finalists as the SMJHL Awards Committee did. For anyone who might be a little newer on the site, I’ve been on the SHL Awards committee for 6 seasons now (since S57) and was previously SMJHL Awards Committee Head from S48-S57. Suffice to say, I’ve been doing site awards for a very long time, and think I do a pretty good job.
Without further delay, here’s my process. I don't think it’s perfect by any means, but I think it works well. I’ll break this media into two parts: Part 1, I outline my process and share who were the “top” players using my Bojo Score. Part 2 dives deeper into context and comes to some conclusions on who I think should be finalists for the S62 Metzler trophy.
(Revised) Bojo Score Metrics
What makes a good defensive defenseman? You could ask 5 people and get 5 different answers. It’s not a clear one-size fits all answer (that’s part of why we’re here). I endeavored to try to capture a lot of different important pieces with my Bojo Score—value the physical game, blocking/suppressing shots, generally keeping the goal out of your net, not turning the puck over, extensive use on the penalty kill, etc. What I value/assign points to has changed a bit as the seasons have gone on, but now I’ve narrowed it down to 7 categories I think carry part of the bigger picture value of “good defense”.
1. Shorthanded time on ice (SHTOI) – simply put, if your team isn’t using you to kill penalties frequently (or at all), I don’t think you are the best defensive defenseman.
2. Goals Against per 60 (GA/60) – on average, how many goals does your team let up when you’re on the ice?
3. Takeaway ratio (Takeaways/Giveaways) – a higher TA/GA means you’re forcing more turnovers than you’re causing. This is a good thing for a defender.
4. Goal differential (GF/60-GA/60) – does your team have a net positive scoring when you’re on the ice?
5. Shot differential (SF/60-SA/60) – same as above.
6. “Effective” Hits (Hits/PIM) – in an ideal situation, you want to hit the other team so they lose possession/control without taking a penalty
7. Block Quality (Shots Blocked/Corsi Against) – what percentage of the opposition’s shot attempts did you block?
I think using HIT/PIM and SB/CA is a little more beneficial than raw hits/blocks; it puts some context to the raw stats. A player making 10 hits a game but averaging 6 PIMs a game for a team with a 72% penalty kill probably isn’t as effective as a player who takes 4 hits a game but rarely ever puts his team down a man. Previously I used both raw hits/blocks and these other stats in my score—I realized this was double counting to some degree, and after some feedback tried to find ways to streamline the metrics.
From here, I set thresholds for points: hitting the upper mark gets 1 point, hitting the lower mark gets 0.5 points for the category.
Some of the cutoffs probably make sense; at least a minute of SHTOI is a reasonable expectation of a defensive defender, a 1.0 or better TA-GA is seen as ideal for a player, positive goal/shot differentials are also a good thing to see generally. As for some of the others, I would say the hit and block cutoffs are the most subjective. I used the hit metric the first season looking at the data and finding natural cutoff points (possibly for the top 10%? I can’t remember now), and the 15% high mark for SB/CA came from a conversation awhile ago in SHL Awards of “elite” shot blocking. If people have a different cutoff point to justify, I’d be happy to listen. This is my media, and I’m sticking to it!
The SHTOI cutoff brought us from 86 down to 51 defenders to consider. From this, the top 12 SMJHL defenders in the Bojo Score are as follows:
The highest score was a 6.5/7, and 6 teams are represented in the top group. I wanted to do a top 10, but to be inclusive of tied points, 12 qualified. Interestingly, 7/12 come from the league’s top 3 teams—Anchorage, Vancouver, and Newfoundland. This contrasts to a more open representation in the SHL, where of the top 14 Bojo Scores in S62, 12 different teams had representation. This makes me wonder if this will be overly skewed toward players on the best teams. I’ll file that away for part two of the analysis.
Next time, I’ll break down this top group and share how I’d go from this to nominating 3-4 players for the award. I might even tell you all who I think should have won!
Quote:1100 words, give or take, plus at least a few hours of analysis. Ready for grading! I'll release my full Bojo Score results with part 2, if that gets me more $$$.
Thank you karey, OrbitingDeath Ragnar, and sköldpaddor for sigs!
only thing i don't agree with is the TA/GA stuff, simply because in FHM if you're on a good team as a defensive dman you'll often have more takeaways than giveaways in general (in my experience), and that applies to basically all positions, i'm guessing it's because of how possession works in the game but who knows
01-05-2022, 12:27 AMPythonic Wrote: only thing i don't agree with is the TA/GA stuff, simply because in FHM if you're on a good team as a defensive dman you'll often have more takeaways than giveaways in general (in my experience), and that applies to basically all positions, i'm guessing it's because of how possession works in the game but who knows
Hello! I don't have the spreadsheet saved on my personal laptop (don't tell my boss), but the numbers don't really support this statement. In just the most informal review of the SHL's top teams, most of Chicago, Baltimore, and Texas' top defensemen had more giveaways than takeaways (in the case of Corey Kennedy, drastically so) this past season. I went through to look at total takeaways, giveaways, and ratios for the last two seasons for the SMJHL teams. There wasn't a super clear pattern to "good teams better ratio, bad teams worse ratio" or even a clear path on the inverse. If I remember to email that document to myself tomorrow, I'll share those numbers. Also with SHL Awards, I've found many times players on bad teams actually end up with more total takeaways, as good teams carry and control possession more often, so they have more opportunities to lose the puck.
Thank you karey, OrbitingDeath Ragnar, and sköldpaddor for sigs!
01-05-2022, 12:27 AMPythonic Wrote: only thing i don't agree with is the TA/GA stuff, simply because in FHM if you're on a good team as a defensive dman you'll often have more takeaways than giveaways in general (in my experience), and that applies to basically all positions, i'm guessing it's because of how possession works in the game but who knows
Hello! I don't have the spreadsheet saved on my personal laptop (don't tell my boss), but the numbers don't really support this statement. In just the most informal review of the SHL's top teams, most of Chicago, Baltimore, and Texas' top defensemen had more giveaways than takeaways (in the case of Corey Kennedy, drastically so) this past season. I went through to look at total takeaways, giveaways, and ratios for the last two seasons for the SMJHL teams. There wasn't a super clear pattern to "good teams better ratio, bad teams worse ratio" or even a clear path on the inverse. If I remember to email that document to myself tomorrow, I'll share those numbers. Also with SHL Awards, I've found many times players on bad teams actually end up with more total takeaways, as good teams carry and control possession more often, so they have more opportunities to lose the puck.
ah shit i said good teams instead of bad teams lmaooo my bad
I like what you have done here. I'm smiling that you say in part II that you spent 15 hours on this.
On one hand, if we each spent 15 hours analyzing each of the 12 awards, that would come to 180 hours, or slightly more than 22.5 days on an 8 hour per day job. We obviously cannot do that. The Metzler was not assigned to me as the leader of the discussion for the award either. However, I do want to use what you have done here to change how I approach the Metzler in S63.
The only place that I would differ from you is in the use of SHTOI in the equation. My player, a forward, actually led the entire SMJHL in SHTOI as a rookie. (Smiling again at how I might pronounce SHTOI and what that might mean that I led the league in it) Anyway, all that shorthanded time had no bearing on how good my player was defensively relative to the entire league. It says that my team was penalized often and that I was one of the best options available on my team to attempt to kill a penalty.
I don't want to condense the awards process down entirely to a formula either. I'm trying to find a balance between your analysis and what we actually did on the awards committee. I'm looking and listening. When I have the time, I'll look at a modified version of the Bojo score and add in the players who were actually nominated to see the results. Is there any weighting in the formula? I might tinker with the weighting somewhat if there is. In any event, this is just a sincere effort to improve the process. I can use some of the same criteria for another award that is more my personal responsibility on the team.
01-09-2022, 07:26 AMSabresFan Wrote: I like what you have done here. I'm smiling that you say in part II that you spent 15 hours on this.
On one hand, if we each spent 15 hours analyzing each of the 12 awards, that would come to 180 hours, or slightly more than 22.5 days on an 8 hour per day job. We obviously cannot do that. The Metzler was not assigned to me as the leader of the discussion for the award either. However, I do want to use what you have done here to change how I approach the Metzler in S63.
As someone whose been involved in running awards on the site for 15 seasons now, I know this. Entirely fair thing to say. When I did it for my own nominations for the Bojo (and all other awards) I didn't spend anywhere near that much time. It took me as long as it did because I was writing it for media consumption, and had to explain my process verbally, so others can understand my logic. There are shortcuts I would use for my own ballot that I didn't want to use here, because y'all aren't in my head. I also had to do it in chunks over a few days breaking a flow state, which slowed me down as well.
01-09-2022, 07:26 AMSabresFan Wrote: The only place that I would differ from you is in the use of SHTOI in the equation. My player, a forward, actually led the entire SMJHL in SHTOI as a rookie. (Smiling again at how I might pronounce SHTOI and what that might mean that I led the league in it) Anyway, all that shorthanded time had no bearing on how good my player was defensively relative to the entire league. It says that my team was penalized often and that I was one of the best options available on my team to attempt to kill a penalty.
That's why it's part of the formula, and not the top stat used. I look to find the players that hit the most of the boxes I find valuable, especially when it comes to defensemen. The points come from cutoffs of minimums necessary in my mind--a player who played 1:47 SHTOI a game gets the same number of points as the player who played 2:47 SHTOI. It is the one stat I used as a full cutoff, because I don't think a player should get nominated for this award without playing at least 1:00 SHTOI a game. That doesn't overly benefit players on bad teams or hurt players on good teams. You would think teams would put their best PKers out for more time, even if you're taking drastically fewer penalties.
That's why this article came about, me taking such umbrage at you all nominating a player who played 0:38 of PK time a game as the best defensive defenseman, over any outstanding player from the other 13 teams. This is especially frustrating because the community still hasn't gotten any explanation from the committee behind their nominations. I spent much of my free time this past week going over the stats and can't figure out what led people to vote for Mikko Rashford. The award show hosts were confused, too, which I infer means the committee didn't provide any writeups for trashae and AgentSmith. I'm willing to give drokeep some grace because it's his first season, I've done the job, and it takes time to figure out the awards head job and all the little details.
01-09-2022, 07:26 AMSabresFan Wrote: When I have the time, I'll look at a modified version of the Bojo score and add in the players who were actually nominated to see the results. Is there any weighting in the formula? I might tinker with the weighting somewhat if there is. In any event, this is just a sincere effort to improve the process. I can use some of the same criteria for another award that is more my personal responsibility on the team. Bojo Score is my first filter to make sense of the field and narrow it down to a more manageable number to consider for the award. I also use similar systems for the Stevens (best defenseman) and Jeff Dar (best two-way forward).
I think I addressed the concern about condensing the awards process down to a formula. The Bojo Score concept is my first step to review and condense the field based on what stats I believe best reflect good defensive defensemen. It's not the final one. After that is what took me the most time--digging deeper into what the model is and isn't saying, and contextualize it better. Certain stats will trend higher for players on good teams, while others will trend higher for players on bad teams. This is the best way I have found to make sure I'm looking at both, so I ensure what I think will be the fairest process. There's certainly adjustments/weighting you can do around the cutoffs and what points each level gets. I've not formalized it in that way, because I want to see players who hit most/all of the boxes. I probably value the SB/CA and HIT/PIM more than most would, and might give that a little higher score than the goals/shots against stat/differentials, because a lot of that is so team dependent.
Maybe it would be valuable to assign the top group point total (1 in my case, maybe higher if someone doesn't like decimals) based on what the top 10% was that previous season. I think that has a flaw of not putting things into historical context, but could still be a valid and beneficial approach. I would still keep the SHTOI category the same--I think a full point for >= 1:30 SHTOI and 0.5 points for 1:00 =< x < 1:30 is a good metric.
Thank you karey, OrbitingDeath Ragnar, and sköldpaddor for sigs!
You caught my attention enough that I looked back to my own ballot to see who I voted for. I confess that I didn't put nearly as much time went into this article, but I can see at least where the discrepancy lies. I went from the suggestions put forth by the committee on this one and don't mind sharing that out of those choices, I had Rashford at the top. Then, I looked back at the Bojo score information reproduced below:
Then I discovered one of the limitations of the Bojo score. I'm going to run down your list of categories and get to where the formula breaks down in a moment.
I would throw out shorthanded time, as I said earlier. I wouldn't even consider it, because it relies on coaching decisions and the amount of penalty time the team receives. Rashford had a TA/GA > 1, although by a rather narrow 29:26. He only blocked 10 shots during the entire season, which would be the best place to point to in saying that Rashford did not deserve this award.
Rashford has a hits per PIM of about 2.5 or so, but this is a metric like shorthanded time that I would give less weight to. The most physical players are often taking themselves out of position, which is not something that the best defensive players are apt to do. The greatest defensive players tend to be more quietly efficient.
Now we are getting into the heart of the matter. The formula awards a +1 for a GA/60 < 3.0. Rashford, Metzler and O'Neil all had a GA/60 of 1.7, which tied for the league lead for defensemen. These players deserve more than a plus one for two reasons. First, GA/60 is probably the most defining statistic for a defensive player. The whole point of an effective defense is to prevent the other team from scoring. Second, the Bojo score is constructed to give a check box to every satisfactory statistic, but it does not further reward players who really excel in a category. The points need to be awarded based on standard deviations for the formula to mean much. That would simply take too much time. Rashford also edged out O'Neil for the lowest SA/60 in the league for a defenseman. In the Bojo score, Rashford gets a 1.0, the same score in each of these categories as a mediocre player. In a college grading system based on a curve, Rashford would get a 4.0 in each of these categories. I'm not out to fully "defend" the award; only to point out that GA/60 and SA/60 should be very important metrics in determining who the best defensive defenseman is.
Finally, the Bojo score needs to be adjusted to be useful. Currently it rewards, a player who is a C+ in all categories greater than it rewards a player who truly excels in many categories but has a lower score in one or two. The Bojo score distorts the picture if it is relied upon too much. It potentially rewards a jack of all trades and master of none.
Finally, If the award were being decided today, I would almost flip a coin between Rashford and Metzler. I might need to let FHM decide. FHM gives Rashford a 72 defensive game rating and Metzler a 58.
01-09-2022, 03:08 PMSabresFan Wrote: Finally, If the award were being decided today, I would almost flip a coin between Rashford and Metzler. I might need to let FHM decide. FHM gives Rashford a 72 defensive game rating and Metzler a 58.
Game rating is a function of deployed role and should have no factor in awards, just saying.
hhh81![[Image: unknown.png]](https://cdn.discordapp.com/attachments/895496887354945548/927738342261547028/unknown.png)
![[Image: unknown.png]](https://cdn.discordapp.com/attachments/813090165781168199/927746954690179123/unknown.png)
![[Image: AH23zKq.png]](https://i.imgur.com/AH23zKq.png)
![[Image: BAL_Player_Page.png]](https://cdn.discordapp.com/attachments/298599854899200001/1129921039950958682/BAL_Player_Page.png)
![[Image: BAL_Update.png]](https://cdn.discordapp.com/attachments/298599854899200001/1129921040164855819/BAL_Update.png)
![[Image: riU14F5.gif]](https://i.imgur.com/riU14F5.gif)
![[Image: Pythonic.gif]](https://sig.grumpybumpers.com/host/Pythonic.gif)
![[Image: Championship_Sig.png]](https://cdn.discordapp.com/attachments/256499331765501952/807015790431698944/Championship_Sig.png)
![[Image: Eo2nBCt.png]](https://i.imgur.com/Eo2nBCt.png)
![[Image: QtEp67y.png]](https://i.imgur.com/QtEp67y.png)
![[Image: 2sRs0Cq.png]](http://i.imgur.com/2sRs0Cq.png)
![[Image: embros.png]](https://media.discordapp.net/attachments/364939328046235651/904115330266333204/embros.png)
![[Image: ddIlIkT.png]](https://i.imgur.com/ddIlIkT.png)
![[Image: qzGJBye.png]](https://i.imgur.com/qzGJBye.png)
![[Image: sve7en.gif]](https://sig.grumpybumpers.com/host/sve7en.gif)
![[Image: 1tWWEzv.png]](https://i.imgur.com/1tWWEzv.png)
![[Image: 8zFnf2t.png]](https://i.imgur.com/8zFnf2t.png)
![[Image: 6Lj3x8E.png]](https://i.imgur.com/6Lj3x8E.png)
![[Image: xkAdpbO.png]](https://i.imgur.com/xkAdpbO.png)
![[Image: xnZrhKU.png]](https://i.imgur.com/xnZrhKU.png)
![[Image: 9YigPG2.png]](https://i.imgur.com/9YigPG2.png)
![[Image: bpYxJ69.png]](https://i.imgur.com/bpYxJ69.png)
![[Image: jay2233.gif]](https://sig.grumpybumpers.com/host/jay2233.gif)
![[Image: m59RPb7.png]](https://i.imgur.com/m59RPb7.png)
![[Image: FcWmVTl.png]](https://i.imgur.com/FcWmVTl.png)
![[Image: OX6Yrrn.png]](https://i.imgur.com/OX6Yrrn.png)
![[Image: hPSkjwC.jpg]](https://i.imgur.com/hPSkjwC.jpg)
