Eastside Hockey Manager - Part IX Maybe 1.5 Edition?

[GloryDaze4877]

IMO Zacha got overrated in the pack. 116 tops!

FYI, this post comes from a competing GM in our league. The guy that traded me Zacha actually lol.

Thanks for your input Brad.

 

[xECK29x]

I appreciate you taking the time to answer. This isn’t about busting your ass over ratings. I was just hoping you could shed some light on how you come up with the ratings. I thought maybe you had some sort of formula or template you used and was wondering why players with similar stat lines were rated very differently. Thanks

I didn't take it that way, don't worry I've gotten a lot worse from Canunks and Flyers fans . I actually hope Zacha is turning a corner, I like a lot of those Devils players.

 

[GloryDaze4877]

S% of 17.9 this season when his career is 10% (means he should have 1-2 fewer goals) has a PDO of 102.7 which is bound to regress. He's at most a middle top 9 player who can maybe fill top 6 roles due to injuries and isn't a top 6 forward on contending teams. Also his xGF this season (even with his hot streak) is 6.76 which is 356th amongst skaters (if you add in min TOI of 150 minutes) it's 257th amongst skaters and he has an xGF% of 47.73 (below 50 is bad). So in conclusion, he's on a hot streak, his underlying numbers don't support his current numbers and the Devils get outscored when he's on the ice.

Edit: Forgot some things.

Advanced stats are great, but in this case they have no relevance. I was asking why Zacha was underrated (imo) in the SIM based on the ratings of similar players. That would be based on their career stats in the NHL up to this season, not 20-21’s advanced numbers.

For example:

Zacha: (124 in SIM), 2015 6th overall pick,
255gp, 106 pts (35g/69a). His season totals are 24/24/25/32. Last two years he’s averaged more than 16 min TOI

Crouse: (134 in SIM) 2015 11th overall pick,
230gp, 63 pts (32g/31a). Season totals are 12/1/25/25. Last two years he’s been roughly 13 min per game.

Virtanen: (146 in SIM) 2014 6th overall pick, 279gp, 95 pts (50g/45a). Season totals 13/1/20/25/36. He’s averaged roughly 14 min per game the last two seasons.

So, what I was asking is why is Zacha’s SIM rating lower than Crouse’s when he’s produced 40+ more points over the same 4 year period. Why is it a lot lower than Virtanen’s when he’s produced more points in less games? Is it subjective? Or is there a statistical based reason?

Unfortunately, your advanced stats from 20-21 (while interesting), do nothing to answer the question I originally asked.

 

[Dr Quincy]

He's one that I've already bumped, but he was from all accounts trending towards busting vs his draft position. 124 has him as a bottom sixer, not a borderline NHLer and he had upside to be a middle 6 player.

He has been bumped to a solid 3rd liner with potential to be a top 6 player.

You can't please everyone. Thanks for your work.

 

[Dr Quincy]

nevermind

 

[GloryDaze4877]

nevermind

In answer to your question (that you deleted?), his draft position has nothing to do with his rating now. I originally used it just as a point of reference and that he has been a regular in the NHL since the age of 19.

When Eck replied, saying that Zacha appeared to be trending towards being a bust, I used some examples of players drafted in similar spots. Other than that, where he was drafted should have zero relevance on his SIM rating 5-6 years later.

 

[Beef Cake]

Advanced stats are great, but in this case they have no relevance. I was asking why Zacha was underrated (imo) in the SIM based on the ratings of similar players. That would be based on their career stats in the NHL up to this season, not 20-21’s advanced numbers.

For example:

Zacha: (124 in SIM), 2015 6th overall pick,
255gp, 106 pts (35g/69a). His season totals are 24/24/25/32. Last two years he’s averaged more than 16 min TOI

Crouse: (134 in SIM) 2015 11th overall pick,
230gp, 63 pts (32g/31a). Season totals are 12/1/25/25. Last two years he’s been roughly 13 min per game.

Virtanen: (146 in SIM) 2014 6th overall pick, 279gp, 95 pts (50g/45a). Season totals 13/1/20/25/36. He’s averaged roughly 14 min per game the last two seasons.

So, what I was asking is why is Zacha’s SIM rating lower than Crouse’s when he’s produced 40+ more points over the same 4 year period. Why is it a lot lower than Virtanen’s when he’s produced more points in less games? Is it subjective? Or is there a statistical based reason?

Unfortunately, your advanced stats from 20-21 (while interesting), do nothing to answer the question I originally asked.

Those aren't similar players to Zacha?

 

[WhalerTurnedBruin55]

A bit more progress with the Jersey Generator:

- Additional fonts (in .ttf format) can be added to the Generator by placing them in the "extra_fonts" sub-folder;
- I've added a two-tone effect option so that you can get a similar shine/two-tone jersey style as the LWL jerseys; and
- The text can be resized and moved around as different fonts have different default sizes and spacing.

I think it's just a case of adding the option to select different jersey designs or .svg files and then it should be ready to test. I can then work on the batch import and generation aspect.

Woah this is also very cool. Definitely would be down to test it out as well when you have a batch feature ready!

 

[GloryDaze4877]

Those aren't similar players to Zacha?

Stylistically, no.

Statistically, both are worse.

 

[Beef Cake]

Stylistically, no.

Statistically, both are worse.

How do you figure?

Zacha had 8 goals in 16 atoi last year (65 GP)
Crouse had 15 goals in 13 atoi last year (66 GP)
Virtanen had 18 goals in 13 atoi last year (69 GP)

 

[GagneHanson12]

I mean it takes literally 2 seconds to make edits yourself, and a kindergarten aged kid could do it in the editor after doing a min of google searching. I don't get why this is such a big deal lol.

 

[archibalduk]

Here's a new version of the Jersey Generator: Release v1.0.0-beta.4 · archibalduk/EHM_Jersey_Generator (github.com)

I think this is basically feature complete in terms of jersey generation. I need to sort out long names to ensure they fit within the jersey and I might add a setting to adjust the brightness of the two-tone colouring in case you want a darker/brighter effect. If anyone is interested and has any feedback then do let me know and I'll see what I can add for the next release. In the meantime, I'm going to get started on the batch generation side of the Generator.

The Generator allows for two types of jerseys: (1) custom layers and (2) preset images. The layers consist of .png templates for the foreground and .png templates for the trim. I've done 50% of the NHL so far and will try to get the remaining done for the next release. You can choose any combination of layers you want (e.g. you could use the Anaheim foreground layer and the Colorado trim layer) - and of course any colours. Alternatively, you can put a "preset" image in the 'jersey_presets' sub-folder which can then be selected in the Generator. The presets don't allow you to change colours other than the text colour (being the trim colour). If you try putting one of the .svg files from LeftWingLock in the 'jersey_presets' sub-folder then you will see how this works (see also the first screenshot below). Additional layers can be added to the 'layer_foreground' and 'layer_trim' sub-folders. There's also an 'extra_fonts' sub-folder where you can put additional .ttf fonts. Any additional presets, layers or fonts will show up when you next load the Generator.

Here's a very quick run-down of the settings:

• Font: Select the font for the name and number.
• Allow accented characters: Checking this box allows accents. Unchecking this box will convert any accents to unaccented text. Uncheck the box if you find that your selected font doesn't output any accented characters.
• Two tone effect: Enable or disable the two tone background colour.
• Adjust horizontal position / adjust vertical position: Moves the text left/right or up/down by the chosen number of pixels. A negative number moves left/up and a positive number moves right/down.
• Adjust text size: Make the text smaller/larger by the chosen number of pixels. A negative number makes the text smaller and a positive number makes the text larger.

Changelist:

• Added 'extra_fonts' sub-folder into which additional fonts can be placed (.ttf format);
• Added 'jersey_presets' sub-folder into which preset jersey designs can be placed (.png format or .svg format);
• Added 'layer_foreground' and 'layer_trim' sub-folders into which additional foreground and trim layer templates can be placed (.png format albeit .svg is possible but is untested);
• Added additional freeware fonts;
• Added layer template images based on 16 NHL teams;
• Added options to enable/disable accented characters and upper case text;
• Added options to select either custom foreground and trim layers or a preset jersey image;
• Added text adjustment settings which allow the text to be repositioned and the size increased/decreased;
• Added two-tone effect to jerseys;
• Adjusted the positioning of the text; and
• Improved default jersey layer images.

 

[Doothpick]

Archi you're a god. Thank you very much for this!

 

[BKarchitect]

@archibalduk Love how you let the background and trims be layered - so not only do you have the 31 standard NHL jersey designs but really hundreds of additional possibilities!

I assume the batch generation will be able to read/link the team colors set in the database?

Just amazing stuff...

Test images from in-game...for me the idea of having a graphic jersey representation of every player in the database that can be updated is spectacular.

Let me ask this-will there be a way to customize the presets for any given club? So that you can "design" their jersey and then it will be stored somehow so that when you want to regenerate the jersey images, the utility will know the presets?

 

[archibalduk]

@archibalduk Love how you let the background and trims be layered - so not only do you have the 31 standard NHL jersey designs but really hundreds of additional possibilities!

I assume the batch generation will be able to read/link the team colors set in the database?

Thanks so much for your kind words!

Yes you're right - the colours will be pulled from the DB. It's going to involve two stages:

(1) Exporting a Generator-compatible spreadsheet from the Editor (I'll add this to the next Editor release) and this will export a colour scheme for every person. This will either be the home colours of the player's club contracted (or maybe club playing for???) or, for free agents, their nation's colours.

(2) Loading the spreadsheet into the Generator which will then generate all of the jerseys.

I did think about just loading the DB/saved game directly in the Generator but this would add a great deal of complication, so I'll leave them separate for now. Potentially I could incorporate the Generator into Editor v2 but I'm not sure whether it is helpful to merge them.

Let me ask this-will there be a way to customize the presets for any given club? So that you can "design" their jersey and then it will be stored somehow so that when you want to regenerate the jersey images, the utility will know the presets?

Yes! There will be a separate spreadsheet where you enter the team/nation name and either (1) the trim and foreground layers or (2) the preset svg file. The Generator will then use these for generating the jerseys. Where the club doesn't have any set layers or preset, the Generator will either use a selected default design or can generate a random design. The randomness will probably be calculated by adding up the numerical value of the letters in the club name (e.g. A = 1, B = 2, etc) in order to ensure that players from the same team/nation have the same random design.

EDIT: One thing to add: I think we could in the future expand the Generator to generate the front side of the jerseys in order to create jerseys for team news graphics. The Generator could probably pull the team logos from the EHM graphics folder and add them to the front of the jerseys.

I'm so glad you came up with your mock-ups and the idea for this, BK. It's such a simple yet effective idea.

 

[BKarchitect]

Thanks so much for your kind words!

Yes you're right - the colours will be pulled from the DB. It's going to involve two stages:

(1) Exporting a Generator-compatible spreadsheet from the Editor (I'll add this to the next Editor release) and this will export a colour scheme for every person. This will either be the home colours of the player's club contracted (or maybe club playing for???) or, for free agents, their nation's colours.

(2) Loading the spreadsheet into the Generator which will then generate all of the jerseys.

I did think about just loading the DB/saved game directly in the Generator but this would add a great deal of complication, so I'll leave them separate for now. Potentially I could incorporate the Generator into Editor v2 but I'm not sure whether it is helpful to merge them.



Yes! There will be a separate spreadsheet where you enter the team/nation name and either (1) the trim and foreground layers or (2) the preset svg file. The Generator will then use these for generating the jerseys. Where the club doesn't have any set layers or preset, the Generator will either use a selected default design or can generate a random design. The randomness will probably be calculated by adding up the numerical value of the letters in the club name (e.g. A = 1, B = 2, etc) in order to ensure that players from the same team/nation have the same random design.

EDIT: One thing to add: I think we could in the future expand the Generator to generate the front side of the jerseys in order to create jerseys for team news graphics. The Generator could probably pull the team logos from the EHM graphics folder and add them to the front of the jerseys.

I'm so glad you came up with your mock-ups and the idea for this, BK. It's such a simple yet effective idea.

I was trying to think of ways to utilize a front view with logo - never occurred to me to look at the news graphic. That would be neat.

I still can’t get over how quickly you took my mock up idea and made it into something much more than I was even imaging...already having loads of fun playing with the jersey editor...it could create a whole new field of EHM editing and tinkering!

 

[WhalerTurnedBruin55]

@archibalduk Love how you let the background and trims be layered - so not only do you have the 31 standard NHL jersey designs but really hundreds of additional possibilities!

I assume the batch generation will be able to read/link the team colors set in the database?

Just amazing stuff...

Test images from in-game...for me the idea of having a graphic jersey representation of every player in the database that can be updated is spectacular.

Let me ask this-will there be a way to customize the presets for any given club? So that you can "design" their jersey and then it will be stored somehow so that when you want to regenerate the jersey images, the utility will know the presets?

Those look sweet!

@archibalduk , any way to incorporate/integrate with the editor?

Would be nice to be able to do a single export from a league, with colors integrated from assigned teams in the db. OR have a joined spreadsheet export for JUST the jersey export. Basically joining tables with player data (first, last, dob, current team), and then the 3 team colors (may need to export hex for the colors of current db?)

 

[xECK29x]

@archibalduk Love how you let the background and trims be layered - so not only do you have the 31 standard NHL jersey designs but really hundreds of additional possibilities!

I assume the batch generation will be able to read/link the team colors set in the database?

Just amazing stuff...

Test images from in-game...for me the idea of having a graphic jersey representation of every player in the database that can be updated is spectacular.

Let me ask this-will there be a way to customize the presets for any given club? So that you can "design" their jersey and then it will be stored somehow so that when you want to regenerate the jersey images, the utility will know the presets?

Nobody makes EHM pop like you BK, this looks so fresh

 

[BKarchitect]

Lol...from my head to reality via the editing master himself.

Here's a few more screens in-game after playing around with the editor...there will be a few custom-cut jerseys that may be hard to duplicate (ie, the Coyotes throwback) but it seems like we could always add new templates. I've noticed in my database I'm going to have to swap some foreground and trim colors to produce more accurate jersey results.

The one drawback to this entire endeavor is that if you call up/send down a lot of players to a farm team with a different color/jersey combo...obviously that won't automatically update in-game and will require a batch update via the utility. Sounds like archi is going to make this process quite easy but it'll still require some manual labor. In my head if I'm playing an NHL team, I'm thinking - I'll probably generate the jersey profiles when everybody is on the NHL team that way my call-ups will be in the correct jersey. Since I don't "play" as the farm team as much...a jersey mismatch there in mid-season isn't a huge deal. And then update via the batch exporter after big trades or at critical junctures like the trade deadline or start of a new season.

 

[archibalduk]

Those look sweet!

@archibalduk , any way to incorporate/integrate with the editor?

Would be nice to be able to do a single export from a league, with colors integrated from assigned teams in the db. OR have a joined spreadsheet export for JUST the jersey export. Basically joining tables with player data (first, last, dob, current team), and then the 3 team colors (may need to export hex for the colors of current db?)

Long-term, it could potentially be integrated into the Editor with a bit of work. Initially, it's easier to code it as a separate tool but it will basically be two clicks anyway: one click to export the requisite spreadsheet from the Editor and one click to import it into the Generator. The spreadsheet will pull all of the colours from the database, so it'll just be a single spreadsheet. The only separate part will be the jersey designs spreadsheet as this data isn't available in the database.

Here's a few more screens in-game after playing around with the editor...there will be a few custom-cut jerseys that may be hard to duplicate (ie, the Coyotes throwback) but it seems like we could always add new templates.

Heh I was thinking of the old Coyotes jersey too! It should be an easy add if we can find a decent image of it.

I've noticed in my database I'm going to have to swap some foreground and trim colors to produce more accurate jersey results.

Hold that thought! I realised yesterday that there isn't any reason to separate out the various layer designs. So in the next release, you'll be able to use a trim image as the foreground layer and foreground image as a trim layer. So hopefully that avoids changing anything around in the database.

The one drawback to this entire endeavor is that if you call up/send down a lot of players to a farm team with a different color/jersey combo...obviously that won't automatically update in-game and will require a batch update via the utility. Sounds like archi is going to make this process quite easy but it'll still require some manual labor. In my head if I'm playing an NHL team, I'm thinking - I'll probably generate the jersey profiles when everybody is on the NHL team that way my call-ups will be in the correct jersey. Since I don't "play" as the farm team as much...a jersey mismatch there in mid-season isn't a huge deal. And then update via the batch exporter after big trades or at critical junctures like the trade deadline or start of a new season.

Yes I think going by the player's Club Contracted probably makes more sense than Club Playing but the Editor can export alternative spreadsheets anyway.

 

[Beef Cake]

Would it be plausible to use either the contracted or playing for column? Say we play a ton of juniors, it would be nice to have all WHL/OHL/Q jerseys at the time of playing. Or if we for some reason enjoy playing the AHL.

If both options arent plausible, contracted all the way!

 

[archibalduk]

Would it be plausible to use either the contracted or playing for column? Say we play a ton of juniors, it would be nice to have all WHL/OHL/Q jerseys at the time of playing. Or if we for some reason enjoy playing the AHL.

If both options arent plausible, contracted all the way!

Totally agree. It's easy to allow both, so it's no problem.

UPDATE: Basic batch generation is now up and running but with just a single design. The next task is to allow for the random generation of jersey designs as well as selected designs on a team-by-team basis.

 

[GloryDaze4877]

How do you figure?

Zacha had 8 goals in 16 atoi last year (65 GP)
Crouse had 15 goals in 13 atoi last year (66 GP)
Virtanen had 18 goals in 13 atoi last year (69 GP)

1) I look at a player's career numbers, and how they are trending, not just one year. Virtanen got a big bump after last year. Crouse too. How's that working out? (Crouse 2 assists in 18 games and -6, Virtanen 1 goal in 19 games and -3)

2) Does your hockey world not have assists? They actually count as points too.

 

[GloryDaze4877]

I mean it takes literally 2 seconds to make edits yourself, and a kindergarten aged kid could do it in the editor after doing a min of google searching. I don't get why this is such a big deal lol.

I'm in a league. We don't allow player edits and must abide by the file.

 

[Beef Cake]

1) I look at a player's career numbers, and how they are trending, not just one year. Virtanen got a big bump after last year. Crouse too. How's that working out? (Crouse 2 assists in 18 games and -6, Virtanen 1 goal in 19 games and -3)

2) Does your hockey world not have assists? They actually count as points too.

Virtanen had more points than Zacha including assists. You said Zacha was statistically better last year, but didn't have more goals or points.