Realistic Ratings?

[Skyward Hockey]

This is meant as no slight to those that put in the tireless work that goes into making ratings packs, but most of the ones I have come across are not the most realistic when it comes down to it. I'm talking realistic to each player in the NHL. Now, I know that takes a crapload of time (tell me about it...haha), but a couple months ago, I set out to create every single one of the players (pro and farm) in my new sim league called "Skyward Hockey".

I still have 11 teams left to go, but I was wondering if any of the experts, as well as the other sim commissioners, wouldn't mind taking a look and letting me know what they think of the players.

I will say off the top - I have never been a fan of extravagent OV. So, the highest rated player in our sim (Alex Ovechkin) is a 76 OV'er. You can mess with the sim sliders, I suppose, if you wanted it to be higher. I also did this so that you wouldn't have the absolute slew of 54 OV'ers and lower that seem to plague a lot of ratings packs. Those guys are pure useless, and let's face it - if you're playing pro hockey in the NHL (in particular) or the AHL, you shouldn't have players rated that low. Our lowest rated player is Ian McKenzie at 55 OV, and we only have three currently rated lower than 56 OV. In my opinion, this gives your GM's a lot more options to manage his team, just like a real GM would. It also puts the onus on the GM to make sure he's paying attention and developing his team. We will have in-season rerating (+/- 1 on pro team per attribute, +/- 2 per attribute for farm players) on and also will have offseason rerates (+/- 3 OV) - so nearly every player in our sim could be of some use.

Oh, and finally - ignore the player named "Aaaaabigail Aaaassshat" - I just wanted to have something at the top to denote what a player would look like with all 99 ratings in our setup.

Here is the link to our Unassigned List: http://skywardhockey.com/Season1/SHL1-Unassigned.html (remember, 11 of the 30 teams are not done yet)

Appreciate any comments. Thanks!


Thomas Gidlow
Skyward Hockey Commissioner
http://www.skywardhockey.com
[email protected]

[ynohtna]

Are your ratings based on statistics or what you think of the player overall? Because Ovechkin was far from the best player last year.

We did lose a rater who did 3 years worth of stats to produce ratings but and we have ones now that go from year to year. I don't mind the year to year as it's easier to relate how good a player is relative to your season as opposed to trying to remember 3 years back.

Just blabbering.

[Skyward Hockey]

Yes, and no.

The ratings I'm creating are based on a players whole career, basically. Ovechkin rates highest because he's the most complete player in the game. You can't rate based on one season when you are starting from scratch - unless you are talking injuries, specifically serious ones (such as Crosby's or Savard's concussions).

There is always going to be players that play better than others in a given year. I'm trying to mold each player into basically a sum of their career to this point - the highs and lows. Now, these are for a new league and we will be 100% sim rerating, so I won't have to manually or otherwise rerate these guys again. We do have prospect creation and Group Two RFA creation in-season, but otherwise, once I lay the ground work, that's it.

I'm using a combination of things - TSN.ca is a great site for career stats. CapGeek.com provides detailed, in-depth contract information. And, as much as I can't stand HockeysFuture.com for their sometimes laughable prospect ratings (even though we use them as a template for our prospect creations...ugh), I use them for placing anyone not under contract with a team in their prospect list.

The "no" part? Well, I'd be lying if I didn't say I've put a little personal touch on the players. For example, I cannot see how anyone that's an NHLer would ever have a DU rating south of 70 - so, most all of our players have DU's above 80. The last ratings pack I used from somebody had 40% of the goalies with DU's under 60 - it was ridiculous. Another example is that I'm not creating useless players for our sim (ie: 52 OV'ers, or the like). If I come to a player that's basically a career minor leaguer (by the way, almost any player under a contract with an NHL team has played at least one NHL game - so someone must think they are decent!), he still is rated decent enough so that if one of my GM's wishes to develop him on the farm, they can. The lowest rated guy so far for us is a 55 OV'er. Most of the "career" minor leaguers I've done range from 57 to 63 OV - which makes them solid to great on your farm, or, at the very least, serviceable on your pro club (as it is in real life). Try playing some random 53 OV'er on your 4th line for a few games and let me know how that turns out.

[Kramden23]

What are your average numbers for a season? Your scoring has to be through the roof.

Ovechkin rates highest because he's the most complete player in the game

And then ya lost me!

[Skyward Hockey]

We haven't run our tests yet since we are still working on the rosters. However, the aim is to be at or maybe slightly lower than NHL standards for scoring. I don't think, though, that we have a super abundance of scoring players. The NHL is a scoring league now, anyway.

You disagree about Ovechkin? That's fair I suppose. But despite a 'down' season last year, no one provides the all-around ability quite like #8. Crosby is a more raw offensive talent (and now a further injury risk after a severe concussion), either Sedin, great all-around players, aren't quite as prolific in the physicality department, and Stamkos, Toews, and Staal, though stars for sure, are all less experienced. Individual tastes are just that, without a doubt. I am not saying I'd necessarily take Ovechkin first if I was starting a team - but for my money, and I would think most unbiased fans would agree, Ovechkin is the best overall player in the NHL.

[Shevchenko]

I think the more you rely on stats to do your ratings, the better it is. Your opinion can be the opposite of another person. People here disagree about a player everyone knows well like Ovechkin, imagine what it could be if we talk about a lesser known player like Mark Dekanich (for example). Stats are facts that can't be argued. Well this is my opinion, good luck with your ratings!

[Kramden23]

but for my money, and I would think most unbiased fans would agree, Ovechkin is the best overall player in the NHL.

When you're talking about a player who doesn't block shots, was second on his team in giveaways, and played a whopping total of 3 minutes 37 seconds of PK time all of last season (0:02 seconds per game) I would hardly call them the "best overall player". Obviously he's a talented player, but "best all around" or "best overall" is pretty naive.

It's funny you say "most unbiased fans would agree", yet your ratings admittedly are based on your own bias

Not looking to start an Ovechkin debate, just thought your assessment was kind of foolish. If you're talking about just the combination of goal scoring ability and physicality then possibly, but that's from from "best all around" or "best overall".

[Skyward Hockey]

Well, regardless of the "best player in the league" debate, what I meant by the ratings bias was that I am biased against creating players that would have no use in the sim, or creating players that are either far above or far below their NHL counterpart. I've taken great measures to make sure the players are as authentic to their career stats as I can. Too bad the sim doesn't have blocked shots or giveaway ratings in Ovechkin's case. Either way, I think the players I've created are just as good if not better looking than what you'll find in most ratings pack. Not being boastful or anything, but I didn't "phone it in" on any of the players attributes (not rating PO, for example), and I've used the best tools at my disposal to do it (TSN, CapGeek, HockeysFuture...meh..well, HF has its positives I guess).

[SDHL_Commish]

Let me just say that I like your idea, and applaud the apparent time and effort you've put into this project.

However...

Your DF ratings look like they're based on +/-, and that is just a no no.

Henrik Sedin - 92df, Pavel Datsyuk - 83df = realism fail.

Also, shouldn't Kovalchuk have a higher SC rating? These ratings are career based, no? Kovalchuk has a career 0.526 goals per game (= 44 goal season), while Perry's is 0.374 (= 31 goal season). Yet Perry has 88sc with Kovalchuk at 85sc. This kind of personal touch does not belong in a rating pack that is being touted as more realistic than the status quo. Otherwise by real (-ism, -istic) you actually mean your own perception.

Again, I really like what you're attempting to do here, and I hope you continue with it. That being said, I would not join a league using these ratings as they are currently constituted.

[Skyward Hockey]

I appreciate the comments! Constructive feedback is always key to getting better at this type of thing.

Something I haven't discussed in a ton of detail is another factor that I guess you could call a personal touch - Kovalchuk hasn't had a good run, for the most part, with his current team. What I try to do is gauge whether or not that will have an effect of the player's performance. The example between Perry and Kovalchuk, for instance. Doubtless Kovalchuk has had a better career (also a longer one), but will he continue to score at his average pace? I've basically set them so that a GM who surrounds his team with good players around a guy like Kovalchuk can get that SC rating up after a good season - thereby, leaving it to the GM. Not saying Kovie can't get his SC rating up on his own, but I felt that, looking at the stats and the trends over the last few years, Perry (the reigning goals champion, by the way) stands to be a better scorer than Kovalchuk.

Obviously, Perry could score 20 goals and go down in SC in our sim year. High SC doesn't guarantee 50-60 goals per year. But due to his age and his generally increasing abilities, it stands to reason his SC should be at least similar to Kovalchuk's. Perhaps I did go a bit too high at 88 SC - once I finish up all the rosters I'll do another pass over the players and I might make some adjustments.

As for the DF - well, other than knowing and watching the players, defensive ability (as well as CK - checking isn't all about hits) is tough to quantify. Datsyuk won the Selke two out of the last three years, so you're right, his DF rating is probably too low. Sedin's 92 was generally based on his plus/minus, and might be a bit high. Again, once I see what everyone looks like together, changes might be a good idea. But again - age is a factor too. Sedin is in the prime of his career, while Datsyuk, despite only 3 years difference, looks to be starting to turn the corner (a few injuries in the last couple seasons, although he was on pace points wise to exceed last year's totals).

There will always be some subjectiveness to doing ratings, especially when you are doing every single player and manually inputting every single attribute. I'd still say though, that the approach I've used has resulted in more complete looking players than what you'd find in most ratings packs - which is realistic. I doubt too many guys would be playing in the NHL with 50's ratings in key categories such as SC, DF, etc. A lot of leagues I've seen have players who might not be prolific on D or in scoring have these ridiculously low ratings. Unless they haven't scored a goal in a decade (and even then, SC isnt all goals - it's also the ability to score, from what I gather), they shouldn't be that low.

To summarize, I just want to give my GM's the best possible opportunities to build as realistic a team as possible given that we are a simulation league. I never really found that with rosters from ratings pack, hence why I embarked on this crazy "do it yourself" binge. It's a TON of work for anyone thinking about giving it a try - and I can also see where people that do it wouldn't mind cutting a corner or two to get through nearly 1,600 players or so. I just didn't want to cut any corners, despite some 'misses' here and there with a couple players having a slightly lower or higher attribute in spots.

Thanks, though, SDHL_Commish, and everyone else, for their comments. It's greatly appreciated!

[ynohtna]

I too like what you're trying to do. Whether I agree with the result is inconsequential really. You will have to go through your own trials and errors to be happy with the results.

I had the same comment about DF being seemingly based on +/- only but was going to save that for later.

I also think that you shouldn't focus on OV. It doesn't matter. Just because the OV is less than 50 or a particular attribute is below 50 doesn't make said player 'useless'. I could have a 30 OV guy but perhaps there are key stats that he's really high FO, DF and SK but crap at everything else. This player is potentially a valuable penalty killer. You don't pick a penalty killer based on OV, you pick a penalty killer based on what stats you think are important for a penalty killer. ie: who cares if his SC is 20. Who cares if EN is 50 (he's only doing PK shifts) who cares if FG is 10. But you do care that maybe he also has good SK and good DI.

What the range of 0-99 gives is the ability to provide a relationship of a players certain attribute versus other players and provide enough gap to distinguish these players.

50 can mean average. 0 could mean useless

Anyways, looking forward to how your ratings develop.

[Skyward Hockey]

While I agree with your basic approach to the "OV doesn't matter" debate, ynohtna, I do believe that creating players with 50 or below ratings is unrealistic. True, there are specialists in the NHL, whether it is D, PK, PP, etc. But everyone in the league has basic skills enough to "get by" in most situations. No one is "crap" at one thing - they may be less talented than most others, but these are NHLers (for the most part) that we are talking about, not beer leaguers.

Not everyone, for example, is a heavyweight title contender - but the ability does exist for players to get in a scrap maybe once in a blue moon. Hence why we default our lowest FG rating to 40 if the player rarely or never fights. Usually does the trick.

I like to think (and I very well may be in the minority) that the ratings in the STHS sim have baselines - usually 50 or 55 was the norm for baseline low ratings in FHL if I remember correctly. A rating of 0, IMO, should never occur in any player unless you are creating beer leaguers. Besides, I like to have a range of OV players, as we do in our new league (55 OV is the lowest and no one will be created lower, while 76 is our highest at the moment and will grow as the years go by). And lastly, having a range of OV that isn't too wide also gives you a lot easier control of things like having a farm OV max, or using OV to determine prospect creation. OV may mean nothing in terms of value in the sim, but it does have uses, at least.

*EDIT* -- Also, when you are a full sim rerate league as we are, it makes it so that performance corrects the ratings. So if I guy is rated well and continually under-produces, he will rerate down. It balances out, and it's a more even balance when players have realistic ratings. Again, though, IMO.