Sondage engin du sim/Sim engine poll

[PNHL]

Je voulais savoir quel version de l'engin de simulation vous utilisiez?

Si se n'est pas la 2.1, pourquoi?

Je fais ce sondage puisque il semblerait que se soit le meilleur, mais qu'il y a trop de buts. Je me disais que si les résultats du sondage sont concluent, peut-être que SimonT pourrait réviser la version 2.1 de l'engin pour diminuer le nombres de buts. Je crois que se sondage donnera un peu le l'opinion général à Simon.

--------------------------------------------------------------------------------------------------------------------
I wanted to know which version of the simulation engine do you use?

If is is not 2.1, why?

I do this survey because it seems to be the best, but there are too many goals. I thought that if the survey results are concluded, perhaps SimonT could revise the 2.1 version to reduce the numbers of goals. I think the survey gonna give a little general awareness of Simon.

[Marc the Habs Fan]

2.0

On a débuté la saison avec 2.1 et il y avait en effet trop de buts - surtout chez les bons joueurs qui scorait à des rythmes digne des années 1980s. On est retourné au 2.0 et le tout s'est stabilisé.

[36Henry]

2.1

There aren't too many goals if ratings are made in a way that is suitable for the engine.

[Foo]

1.5 est de loin le mieux balancé et réalistique dans ses résultats mais il faut utiliser les rosters créées excusisvement pour cet engin sinon les défenseurs ont beaucoup trop de points. Il est d'ailleurs facile d'ajouter un peu de random (beaucoup trop présent dans le 2.0 à mon goût) en joueant avec le slider morale.

Le 2.1 est à mon avis le pire engin de simulation que Simon a créé. Impossible d'ajuster le résultats avec plus d'une centaine de saisons tests.

[PNHL]

2.1

There aren't too many goals if ratings are made in a way that is suitable for the engine.

Can you post me a link of your league? I would like to see those ratings !

Thank you for your answer

[PNHL]

J'utilise le 2.0, mais effectivement, il a un peu trop de ''random''. Il y a trop de joueurs random qui connaisse des saisons folle. Les équipes mauvaises finissent parfois dans le top 5 ligue et les très bonne parfois top 5 pires. j'ai fait aussi des centaines de saisons avec le 2.1 en test et je n'arrivais jamais a rien de concluent!

Le positif du 2.1 = Il est très réel au niveau des tendances des meilleurs joueurs. Un Stamkos - Crosby - Ovechkin termineront dans les meneurs, ce qui est parfait, sauf quand ils ont 356pts en 82 parties
Le négatif du 2.1 = Impossible d'ajuster le niveau de but de façon a être réel sans diminuer de façon vertigineuse les cotes.

[Str_QNHL]

1.5 pour les memes raisons que FOO

[Kramden23]

1.5
After a bunch of testing it gives the best results for our league ratings.

[QSHL]

2.0 for us at the QSHL
We're very happy to use most of the new features, and are very relevent to us, but unfortunately, the 2.1 makes absolutely no sense despite a lot of hours of testing to fix incredibly high scoring games. Some GMs wish we'd go back to 1.0 (or 1.5) because their teams were built to win in these settings and have failed to adjust to new ratings/system. Yeah, once in a while, we see a mediocre team make top-5 and a powerhouse miss the playoffs and reasons to justify this is definitely a head scratcher, but in the end, this is also how NHL works to me : sometimes we see teams we wouldn't expect in the playoffs and good teams get bad seasons (and coach gets fired, etc)

Nous utilisons la version 2.0
Nous sommes très heureux d'utiliser la plupart des nouvelles options de la version 2, mais malheureusement, la 2.1 ne fait absolument aucun sens et il a été impossible de trouver un arrangement pour les hauts scoring malgré des heures de testing. Quelques DGs aimeraient qu'on retourne à la 1.0 (ou la 1.5) parce qu'ils ont connu beaucoup de succès à l'époque avec ces settings, et ne connaissent plus ce genre de succès dans la V2.0. Oui, parfois une équipe pourrie se faufile dans le top-5, et vice-versa, et c'est impossible pour moi de le justifier à mes DGs, mais selon moi, la NHL ne vit-elle pas ce problème aussi?? Parfois, une équipe gagnent sa division et fait son bout de chemin en série alors qu'on l'avait écarté en septembre, et une équipe que l'on voyait loin en séries ne se qualifie même pas!

[36Henry]

2.1

There aren't too many goals if ratings are made in a way that is suitable for the engine.

Can you post me a link of your league? I would like to see those ratings !

Thank you for your answer

Private Message sent

[36Henry]

For those saying 2.1 doesn't make sense, could you link to your leagues? Would love to take a look.

Here's a list of the top scorers in one of many test seasons I ran back in August using 2.1:


* - all stats prorated from a 332 game season.

For the season above the following applied:

• 5.25 goals scored per game compared to 5.31 in the NHL.
• 6.63 powerplays per game compared to 6.65 in the NHL.
• An 18.5% conversion rate on the powerplay compared to 18.2% in the NHL.
• 58.25 shots per game compared to 58.28 in the NHL.
• An average save percentage of .910 compared to .909 in the NHL.
• 47.48 hits per game compared to 47.59 in the NHL.
• Defensemen accounted for 15.9% of all goals, compared to 15.5% in the NHL.
• Defensemen accounted for 62.9% of all blocked shots, compared to 63.0% in the NHL.
• Defensemen accounted for 35.9% of all hits, compared to 35.5% in the NHL.
• Defensemen accounted for 30.8% of all assists, compared to 30.0% in the NHL.

Kris Letang led all defensemen in scoring with 72 points ahead of Karlsson (69),Subban (66), Paul Martin (62) and Mike Green (53).

Goaltenders ranged from .926 to .894 for save percentage and from 2.13 to 3.05 for goals against average.

Pittsburgh won the President's Trophy with a 54-21-7 record and 115 points.

In the East, seven of the eight teams that made the NHL playoffs made the playoffs in the Sim too, with only Toronto missing out in favor of Carolina.

In the West, six of eight playoff-teams made the playoffs. Missing out were St. Louis and Anaheim in favor of Edmonton and Columbus.

I don't think such a season is indicative of an engine that can't produce good results. I do agree though that it takes work to understand it as it treats ratings very differently compared to the 1.5 for example.

[PNHL]

Take a look at one of my test with Goal slider set at 1




And these are my ratings:

http://pnhl10.com/PNHL-ProTeamRoster.html#AnaheimDucks

[ynohtna]

Can I see your league ratings too? My test sim was somewhat ok but now people are getting super lines which score like crazy.

I've tried to adjust the ratings but ran out of time and got flack from my GMs for too much adjustments

I would like to see 2.1 work better

[36Henry]

Not sure I want to get into too much detail as some people have approached me in the past about selling a set of ratings in the future (as I've seen some other dudes do around here). Wouldn't want to give away too many trade secrets

But I've looked at the ratings for both your leagues (BFHL I looked at the 2012/13 season) and early on I too had instances of individual players or lines going absolutely mental like that. I believe I once had Sidney Crosby scoring 600 points in 82 games for example.

A few things struck me though. Both your leagues appear to have very similar ratings with the majority of forwards somewhere in the 50-70 range for the offensive categories like PA and SC. Likewise your goalies are also in the same kind of range with the top guys somewhere in the mid to low 80's as far as I can tell. Given the many similarities it may not be surprising that you both experienced problems of some kind. I certainly know how the cycle of testing and adjusting can lead things in all kinds of directions, but I believe there are other ways of making the adjustments that produce better results.

As for GMs reacting to changes I do know about that too You spend countless hours producing a set of ratings that clearly produce acceptable results and the reaction is always the same. "How can player X be 77 for PA. He should be at least 79!" or the classic "I don't believe that is how the Sim works! (often based on absolutely nothing)

Another factor that I suppose makes a difference is how you build your ratings. What stats impact each rating and how far back in time do you go when collecting the stats? I'm sure everyone who builds their own ratings have their own method for this and I'm also convinced that there really isn't a right or wrong way of doing it. But it is my personal belief that the way the Sim works is crucial when understanding what to base certain categories on.

For me, the 2.1 engine is vastly superior to the others. The first experience I had with it wasn't too great though as I went through the same phase of getting ridiculous results and no matter what adjustments I made it didn't really improve much. So I too went back to 2.0, which isn't bad as such, but there is a thing in it that renders it unusable for me once I discovered it. That's when I decided to give 2.1 another shot and once I figured out its secrets it turned out to be the best by a mile.

[PNHL]

Well, if 2.1 have ''secrets'' to have good results it's #$?#? bad. Why would Simon do that? Seriously I really really really really don't know how can someone can use these ratings unless having players like Stamkos at 70 PA and 75 SC....

My ratings are baised on 2 NHL season and one sim season

33% NHL 2012-2013 / 33% NHL 2013-2014 / 33% PNHL

Simon should lower the number of goals because, in my opinion, it's not normal that people still used an really old engine.