first small test with 2.1

This temporary section in English only contains idea / feedback / suggestion / problem related to the simulation engine beta.
Post Reply
jwr38
The Accomplished One / L'Accompli
Posts: 82
Joined: Mon Feb 01, 2010 4:10 pm

first small test with 2.1

Post by jwr38 »

Here is the first test with all the same teams ( same player, skill, names (with tag) lines and setup). still a large range in points the same type of player gets in a season but there are trends there. I will let you guys look through before sharing my opinion.

http://thenutworks.com/sthsv21test/Outp ... Index.html

the second test

this is where the aryan racers have a star center. Take a look at him. You would sell your let nutt to get him on your team (including a good size of you cap). All the rest of the players are the same for each team again. The test he couldn't come out as the top center let alone the top forward.

http://thenutworks.com/sthsv21test/Outp ... Index.html


The third test is with the d not being so good as in the last test. Again you can see the same trends.
http://thenutworks.com/sthsv21test/Outp ... Index.html

Of course these are one of test but I don't have the real time to make a lot of tests like I would like to. So I offer this below.

Here are the files so you can run these same tests yourself. Just remember to set the sim engine to 2.1 by adding in the command. Right click and save as.
The first three are the test I have above.

This is my base. All the other stats generate from here for easy factor. This was the first test.
http://thenutworks.com/sthsv21test/save ... /team1.sth

the next two tests. First on with an elite center and strong d as before while the lessD file is where the D is not as strong overall.
http://thenutworks.com/sthsv21test/save ... rsStar.sth
http://thenutworks.com/sthsv21test/save ... rLessD.sth

These are some others that I test if you want to try them out.
http://thenutworks.com/sthsv21test/save ... ers110.sth
http://thenutworks.com/sthsv21test/save ... 10mitG.sth


I also created the player creator for version 1 and 2 of the sim. If want to make your own rosters and test them go ahead.

I have two versions.
for version 1
http://thenutworks.com/STHS/playercreate1.php
http://thenutworks.com/STHS/playercreategoal1.php

for version 2. It works but I have not set it up as easy as version 1. :)
http://thenutworks.com/STHS/playercreategoal.php
http://thenutworks.com/STHS/playercreate.php

James
jwr38
The Accomplished One / L'Accompli
Posts: 82
Joined: Mon Feb 01, 2010 4:10 pm

Re: first small test with 2.1

Post by jwr38 »

Okay I will try this again. Lost my last message.

I have done a few more quick tests to see what could cause some strange things I am seeing. In the end I created 7 tests in total that each had a little goal in mind.

In the first test I put mass edit to work. I changed all players stats to 60. This meant from FG to PO every player in the league has the same skills. I did do a little screw up during the test and not make the goalies all 60 stats too. it doesn't help but should be a heavy over riding factor as all the starting goals still have the same stats. My goal was to see if the line that you were on improved your +/- on a winning team and generally played better. Of course you will see that players who played on the first PK and with first line minutes were the top scorers. They don't score that many more goals on the PK but they get a boost in scoring during even strength.

http://thenutworks.com/sthsv21test/Outp ... Index.html

In the next test I looked to see if I made one player better would he preform better than before. Again I went for my favourite center on the Aryan Racers, Darryl KellyA. This time I changed all his stats to 70. He was going to be a machine as he is 10 point better than everyone. Time to lay down some kick ass I say! and indeed he did! with a massive 018 pim minutes in 50 games. 560 minutes of those were for fighting. Okay really it should of read 1018 minutes but ran out of space. Nice going Mr KellyA. He played like crap too. Not near any scoring ladder apart from pim minutes. Those skills were going to a huge waste. Coach tell him not to fight!!!

http://thenutworks.com/sthsv21test/Outp ... Index.html

So of course the next step would be to stop all that damn fighting and lower the scoring a tad to close things up a little. So scoring went down to 25 and fighting down to 1. At last Mr KellyA was scoring! Although not at a great rate as he still had 485 pim 5 minutes.

http://thenutworks.com/sthsv21test/Outp ... Index.html

The next logical step would be to lower fighting in the stats. I made fighting skill for everyone on 10. I still had the random fights but it cured the mad fight nights that were going on before. Since he was no longer off the ice he had more time on the ice were he could score. funny that. but of course he wasn't the best player in the league. He wasn't getting the first PK time. It seemed very important to get those minutes even if you don't score on them. So while he scored the most on his team he wasn't the best player.

http://thenutworks.com/sthsv21test/Outp ... Index.html

Now my goal was to make him clearly the best player in the league. So I increased his stats to 80 but kept the fighting at 10. I want him on the ice not fighting. This time he was the top scoring center and third best player in the league. His problem was that he needed to be on the first pk at all times. He needed the boost of playing there. here are two test showing the difference.

First when he is on the second pk unit ( check out his low ice time AMG compared to first line pkers). That is a good 4 minutes per game.
http://thenutworks.com/sthsv21test/Outp ... Index.html

Now I put him on the first pk unit. Watch him fly. :) He didn't get that many more points than other first line pkers so his skill didn't help much there but damn his 5 on 5 play takes off! Now we have the best player in the league.

http://thenutworks.com/sthsv21test/Outp ... Index.html


Of course that made me think. Was the DF scoring helping out? So I made another test. this time the same as the second test above but with 40DF skill. So it comes out that this defense skill will hurt his scoring. He needs to be on the top pk to get ice time. but because of the low df he could generate scoring chances the same. He still scored the same amount of goals on the pk but his 5 on 5 play went down hill and once again was not the top scorer in the league but only just. A 60 point winger was better than him cause he got good pk time and had a defense of 60. If that makes sense?

http://thenutworks.com/sthsv21test/Outp ... Index.html

So what could the problem be. I don't think the problem is how they players work but ice time and line match ups. Somewhere there seems to be a away of getting the first line pkers more minutes against tired players. The can't bet better or worse players as they all have the same skill so it must be the in game fatigue or something along those lines. These are my shots n the dark while I try to understand the results above.

I will post later when I upload all the game files for you to download and test out.
jwr38
The Accomplished One / L'Accompli
Posts: 82
Joined: Mon Feb 01, 2010 4:10 pm

Re: first small test with 2.1

Post by jwr38 »

Here are my game files for everyone to try out.
all stats 60
http://thenutworks.com/sthsv21test/save ... tats60.sth

As above but with KellyA at 70 stats
http://thenutworks.com/sthsv21test/save ... KellyA.sth

Same as above without the heavy scoring
http://thenutworks.com/sthsv21test/save ... -score.sth

Fighting at 10.
http://thenutworks.com/sthsv21test/save ... yAfg10.sth

Kelly now at 80
http://thenutworks.com/sthsv21test/save ... llyA80.sth

kelly now on the first pk unit and all stats at 80
http://thenutworks.com/sthsv21test/save ... A80all.sth

as above but kelly has 40 in DF
http://thenutworks.com/sthsv21test/save ... 0all40.sth

I would love to ask for two new stats to prove what I am thinking. Each time a player scores or gets scored on his fatigue level is recorded and averaged out over the season. So during the season you can see if the players that get first line pk minutes are fresher when scoring than those that don't get those minutes. Also it would let you see if some players are getting over used and let in scoring chances or something.

edit: these are very easy to use. download, open and then change turn on the sim engine 2.1 and start the season. all line ups and set to give the best results. i.e. same lines for each team unless the tweaked KellyA.
jwr38
The Accomplished One / L'Accompli
Posts: 82
Joined: Mon Feb 01, 2010 4:10 pm

Re: first small test with 2.1

Post by jwr38 »

I have no idea if anyone is interested because I have had no replies but anyway these are some other findings. I am heavily using +/- for this information. What I have noticed is that lines tend to either do well (good +/-) while other lines on the team do very poorly (neg +/-). Of course not all teams seem to have this problem. some teams every line has neg +/- and others have pos +/-. Many though have a very large range between lines. Often the third line has the largest neg movement. Now with every player being the same that really shouldn't be happening to every team's third or second line. While I understand that third line players shouldn't score as many points as first line minutes because of ice time, they shouldn't be scored against more often because they are not on a scoring line. Either there is a fatigue issue when playing on the third line or there is some hard code that makes third liners play a different style than the first two lines by default. I have no firm idea because I do not know what is in the code.

Does anyone else have any ideas from these tests or are they just being ignored?
SimonT
STHS Owner / Propriétaire du STHS
Posts: 14756
Joined: Sat Oct 08, 2005 4:18 pm
Location: Montreal, Canada
Contact:

Re: first small test with 2.1

Post by SimonT »

Hi.

I actually waiting to see if anyone else could command but it seem a topic between you and I now. 

Before I go over your stuff, did you check this topic (http://sths.simont.info/Forum/viewtopic.php?f=29&t=8145), especially the SameRating package? As interesting point is that with v2.0, the difference between the best and worse goalie is 0.08062 versus 0.057 for v2.1. That is a 41.4% improvement.

Now, I have an issue with your result and it’s your methodology. When you try something like boosting a player stat to 70 or 80 when all the others stats are at 60, you need to simulate numerous season of this change with the Automatic Test command. When you simulate 1 season only, you don’t eliminate the random aspect of it enough to get valid result of you test. That why in my test, I have simulated 100 seasons with the same rating to get a good average result.

For your last command, there is no difference in fatigue for third line players. So there is no fatigue issue when playing third-line or anything hard-code for third-line.
-SimonT
Forum Administrator / Administrateur du Forum
STHS Owner / Propriètaire du STHS
English V2 & V3 Manual - Manuel V2 & V3 Français
amn22492
New in Town / Le Ptit Nouveau
Posts: 44
Joined: Wed May 20, 2009 9:20 pm

Re: first small test with 2.1

Post by amn22492 »

I'm loving the player creator, it's interesting to play around with although a little complicated haha.
jwr38
The Accomplished One / L'Accompli
Posts: 82
Joined: Mon Feb 01, 2010 4:10 pm

Re: first small test with 2.1

Post by jwr38 »

version 2 is hard. it was going to be developed a lot like version one but I couldn't create any players that really worked well. mid range players ended up getting more points than high skill players. I haven't worked on it since
Post Reply