HFNHL 2013-14 Ratings Update

[Dr.Sens(e)]

I know that creating ratings is a tough job and you always get the heat from a lot of GM's, but as long as there's logic behind the ratings they can be defended. In this case I for one can't see any logic. The result is that we effectively prevent the addition of young players to the league this year.

There used to be a difference between first year players and the players that played a few games over the years. Why was this changed when we went from creating ratings from full year stats to this 164 game system.

The issue is rookies, not fringe players that score 4 goals in 10 games or play 48 games over several seasons ending up with high ratings. These are two different group of players that previously has been and should be treated in different ways.

You say that "If a rookie plays a full season, there is no penalty." Last year a full season was 48 games, which would mean no penalty, right? Then you say that "48 games still has a slight decrease, but only a slight one." Which one is it?

When you say the drop is pretty insignificant, what do you mean? I don't have the 164 game stats, but it seems like Yakupov rates about the same someone who scored about 16-19 goal scorer over 82 games instead of what his 29 goals/82 games average is? 29 goals/82 games would equal a SC rating of something like a 76 and not a 64. To me 12 points is far from a slight or insignificant drop.

Actually, I'm relooking at the formula, and believe I made a mistake for guys with between 40-80 games, where they were reduced, but mistakenly their goals per game was calculated based on 82 games, not the number they played. This impacts significantly guys like Huberdeau, Yakupov, Gallagher, Saad, Galch, Silfverberg, Conacher and a few other rookies who were calculated incorrectly.

As to rookies versus guys who have played 48 games across three seasons, these are typically rookies anyway who are just in a different situation. So you can't really draw a line here or differentiate, nor should we I don't think. But regardless, looks like mistakes were made here.

I'm rechecking the PA and other formulas to see the extent before redistributing.

 

[Dryden]

I agree with Ville. Took the words right out my mouth.

 

[Dr.Sens(e)]

So it turns out PA and SC for forwards with less than 82 games had been mistakenly calculated, as their goals per game (making up 60% of the SC rating) was calculated based on goals being over 82 games. This actually impacts players on every team, just about, although if they didn't score much, it is of little consequence. The goals per 60 minutes (the other part of the formula) was correct, so that is why it didn't stand out as much as it should have. Guys close to 82 games were only marginally impacted, while those with much lower games played had a bigger impact (although guys with less than 10 games see their ratings drop quite a bit already, so again, less noticeable). Seems like guys in the 40-60 range were the most impacted, basically.

The same happened for d-men with less than 82 games for PA. SC was done right for d-men.

Apologies for this gents. I had distributed the ratings and formulae to the admin team, but I guess this was missed. The appropriate formula was only copied over for D-men SC. For all players with 82 games played or more, this had no impact.

As a result, SC ratings for guys like Yakupov, Saad, Silverberg, Galch, Huberdeau, Gallagher and Brunner will get a good bump going from the 60-67 range, up to the 66-73 range. Same will happen to the PA ratings, at least on the same scale.

On D, the two more notable defenders to get a bump on PA were Schultz and Wiercioch, both of whom jump from 66 and 67, to 70 and 73 (Schultz had played less games, so more of an impact on his rating). There are many other players seeing a smaller bump.

The above rating adjustments might be lowered a hair, because the overall ratings were scaled down a bit during SIM testing. I'll reach out to the admin team to figure out how best to go about undoing the mess I've created here.

 

[Dryden]

Awesome. Was really odd having a goon rated higher than Saad

 

[Canuck09]

The above rating adjustments might be lowered a hair, because the overall ratings were scaled down a bit during SIM testing. I'll reach out to the admin team to figure out how best to go about undoing the mess I've created here.

This is one thing everyone should keep in mind when reviewing their ratings and comparing them to previous years.

After importing the initial ratings and testing with the most recent version of the Sim I had players scoring 150+ goals and 300+ points. The top teams were losing 5 games a season. It was nuts.

To correct this I had to go back to an older version of the Sim and reduce all PA/SC by 5% across the board. This will make comparing previous seasons ratings to this year difficult, so please focus more on comparison to other similar players than what they were in the past.

 

[Vagrant]

The "fake" Audy-Marchessault was released from my team but the new one wasn't added back. Halp!

Just wanted to make sure this didn't get lost in the shuffle. Hate to lose him like I lost John Mitchell. Still bitter.

 

[MatthewFlames]

Just wanted to make sure this didn't get lost in the shuffle. Hate to lose him like I lost John Mitchell. Still bitter.

Fixes are coming. Keep your eyes peeled.

 

[Vagrant]

Fixes are coming. Keep your eyes peeled.

For sure, the guys have been all over it this offseason. That was more a reminder for me than it was anyone. Hail sim minions!

 

[Dr.Sens(e)]

Gents,

I've uploaded the corrected ratings file for the players who have less than 80 games and who should have a bump. The increases range from 1-7 points for SC and PA. There are two tabs in the file, so check both for potential forwards and d-men you have on your roster with less than 80 games. Obviously don't bother for guys you aren't going to have on your roster.

For the changes required, please e-mail simgod and indicate what the bump should be and for which players. He has the file for reference.

Apologies, and good catch by those who noticed. I actually have four players impacted, although they are just my depth defenders.

 

[MatthewFlames]

Some of my players ratings went down

Only Wierdcock needs adjusting

 

[Dryden]

I thought that too at first till i looked at the roster and noticed the file doesn't factor in the -5% so they went up 1 or 2 points.

 

[Dr.Sens(e)]

Yeah, the original rating that was created for each player went down 5%, as per Jeff's note to reduce the overall scoring within the sim (so all the drops were relative). As you will see when comparing them to what is on the league page, there were bumps for all of these players relative to what is on the site right now.

 

[Hossa]

Better to catch the error now than later, certainly.

Just for my clarification though, the error we are resolving is still different than Ville's point about penalizing rookies in their PA and SC ratings, correct? Or will this bring the top rookies such as Yakupov to the same PA and SC ratings as their peers according to the league-wide formula, with no penalty or reduction of any kind?

 

[Ville Isopaa]

Better to catch the error now than later, certainly.

Just for my clarification though, the error we are resolving is still different than Ville's point about penalizing rookies in their PA and SC ratings, correct? Or will this bring the top rookies such as Yakupov to the same PA and SC ratings as their peers according to the league-wide formula, with no penalty or reduction of any kind?

Good to get atleast part of the problem fixed, but as Sean says, this doesn't resolve the problem of penalizing rookies that played the full season last year. Remember that this is not a question about if guys that play 10 games per season for the past 5 years should get full ratings, but about players that play a full season.

Now, what is a full season? The NHL has the cut for rookies at 25 games played during one season. After that the players is no longer a rookie. I guess it's a matter of opinion on when a player is considered a full time NHL player rather than a rookie or prospect.

I found the old SC/PA ratings calculation file from the 2009-10 season, where guys like Oshie 57 games played, Hansen 55 gp, Purcell 40 gp, Möller 40 gp, Pacioretty 34 gp, Wilson 27 gp, etc. got full ratings. The limit seems to have been 25 to get a couple of points off the SC/PA ratings and another cut at around 10 games for a bigger downrating to AHL type ratings.

Nick has used 60 games played, which equals 73% of a season. That's a pretty big difference to the old cut which was at 25 games.

Rookies in 2012-13 had a max games played of 48 games, meaning that they didn't even have a chance to get proper ratings as it was impossible to reach 60 games. Their goals and assists were worth less than other players during that year, even if they played the full year.

Even if we would keep the limit of a full rating at 60 games total as the main rule for this season, I think it would only be fair to have players that played atleast 73% in the 48 game schedule in 2012-13 to get a full rating. That would mean every player that played 35 games or more in 2012-13.

 

[Dr.Sens(e)]

Good to get atleast part of the problem fixed, but as Sean says, this doesn't resolve the problem of penalizing rookies that played the full season last year. Remember that this is not a question about if guys that play 10 games per season for the past 5 years should get full ratings, but about players that play a full season.

Now, what is a full season? The NHL has the cut for rookies at 25 games played during one season. After that the players is no longer a rookie. I guess it's a matter of opinion on when a player is considered a full time NHL player rather than a rookie or prospect.

I found the old SC/PA ratings calculation file from the 2009-10 season, where guys like Oshie 57 games played, Hansen 55 gp, Purcell 40 gp, Möller 40 gp, Pacioretty 34 gp, Wilson 27 gp, etc. got full ratings. The limit seems to have been 25 to get a couple of points off the SC/PA ratings and another cut at around 10 games for a bigger downrating to AHL type ratings.

Nick has used 60 games played, which equals 73% of a season. That's a pretty big difference to the old cut which was at 25 games.

Rookies in 2012-13 had a max games played of 48 games, meaning that they didn't even have a chance to get proper ratings as it was impossible to reach 60 games. Their goals and assists were worth less than other players during that year, even if they played the full year.

Even if we would keep the limit of a full rating at 60 games total as the main rule for this season, I think it would only be fair to have players that played atleast 73% in the 48 game schedule in 2012-13 to get a full rating. That would mean every player that played 35 games or more in 2012-13.

Some good points as I understand your perspective here, but some clarification.

Yes, back in 2009-2010 there was no reduction for lower games played, including those who only played 15 games. We had to go back and 'correct' some of the anomalies, but regardless, there were always fringe players that slipped through with above average SC and PA ratings because they scored 4 goals in 12 games, or 6 assists in 15 games. We changed the approach at the beginning of last year.

But I think the key question to ask here, is if a player scores 12 goals in 40 games, should they have the same SC rating as a player who had scored 50 goals over the last two seasons? My inclination when I introduced this rating formula last year was they should see a slight drop, but not a massive one. Turn their rating into that of a 20 goal scorer, instead of a 25 goal scorer, basically. After all, they haven't scored 20 goals in their career yet.

Now last year was unique obviously, given so many of the full time rookies never got to the 60 game mark because of factors outside of their control. Guessing they might have hit the rookie wall isn't really that fair, but giving them the same SC or PA rating as a player who has produced at their pace over two full seasons might not make sense either.

We can still adjust the rating approach if the vast majority are inclined to. But I think we should hear from everyone, versus those who just have high profile rookies are a lot of players impacted by this. I personally have 5 players in the 40-60 game bracket and stand to benefit more than anyone from this change, so I'm good either way. But none of the bumps will be that substantial, either. So I don't think it impacts things materially if we go one way or the other. There are some guys like Yakupov and Gallagher who see their SC rating jump to around 74 from 70, but those are by the far the biggest increases we'll see. The rest of the changes are going from 65 to 66, or from 60 to 62 kind of thing.

 

[Canuck09]

Just wanted to make sure this didn't get lost in the shuffle. Hate to lose him like I lost John Mitchell. Still bitter.

Not posted online yet, but this is fixed in the Sim.

 

[Dryden]

The Hawks have ratings for Steve Pinizzotto but should have the ratings for Steven Pinizzotto.

 

[TorontoGM]

Were the DU and EN ratings from an outside source? I noticed a few anomalies. ie. Erik Karlson who was injured for the majority of the season.

 

[Canuck09]

Were the DU and EN ratings from an outside source? I noticed a few anomalies. ie. Erik Karlson who was injured for the majority of the season.

Outside of PA/SC/PS the entire ratings set is from the SIMHL pack.

 

[Canuck09]

The Hawks have ratings for Steve Pinizzotto but should have the ratings for Steven Pinizzotto.

Just to close this off here as well, this is done.