How are lines used in the VHL?

[studentized 166]

It's been a longer term goal of mine to figure out what makes a line perform well. This is a pretty broad topic and one that I will probably have to do many write ups on to get a satisfactory answer, but to start I have to understand how lines are used in the VHL. 

 

Most teams have 2 lines worth of non-bot players. A few teams run with 3 lines like New York (who now are back to running 2 lines the past couple of games). Despite this, the STHS sim runs with 4 "normal" lineups and 2 lineups of each special teams type. This means some players will inevitably play on many lines and the VHL compensates for this by giving every player a 99 endurance rating. Or so I thought.

 

In my scrounging through STHS docs trying to figure out what all the little bells and whistles do in the sim, the first thing I needed to do was to understand the different versions of the simulation engine. There are four versions: 1.1, 1.5, 2.0, and 2.1. The gist of it is that 1.1 came first and is the O.G sim. 2.0 came next with a whole bunch of new features but the output was so random and gross that no one could use it. Then came 1.5 which took a few of the nice 2.0 features and added them to the 1.1 engine. 2.1 is essentially a bug fixed 2.0. As you can read here (and confirm here), one of the things added to 1.5 that is NOT present in 1.1 is the ability to control line strategies/time on ice. This is a feature that GM’s appear to be able to control and configure already due to the STHS client we use (check out our index) BUT in fact does not even impact the sim because we sim on the 1.1 engine (thank you to devise for confirming). If you continue reading that STHS forum post, some player stat categories from the 2.0 engine were added to 1.5 but not to 1.1, most notably EN (but also PS, which confirms the generally accepted idea around the forums that PS was a useless stat). All this being said, I’m not trying to say that it’s a bad thing to use the 1.1 engine. What it does mean is that we are at the mercy of STHS when it comes to how often each line gets used and when players become fatigued.

 

 

This begged the question: how are lines actually used in the VHL? The chart above shows the amount of "shifts" for each line across all teams in S69 so far (as of game 155). Here I've counted a "shift" as a line change parsed from the full play by play of the game logs where some event occurred in between. For example something like:
 

Normal Lineup #1 is on the ice for home team.
Normal Lineup #1 is on the ice for away team.
Player A shoots the puck.
...
Normal Lineup #4 is on the ice for home team.
Player B shoots the puck.
Normal Lineup #2 is on the ice for away team.
...

would be counted as one shift for home line #1 vs away line #1, one shift for home line #4 vs away line #1, and one shift for home line #4 vs away line #2. In particular, there is no time component being measured here (the first shift is almost certainly longer than the second shift in my example) which is a limitation, but my hope was that this type of thing would even-out in the long run.

 

As for the chart, the biggest thing that stands out is the higher amount of shifts for normal line 3 vs normal line 2. At first I thought there were bugs in my code, then I thought that this was one of those discrepancies between shifts and time played, but then I looked at SHTS which tracks minutes played for each player and separates out speciality teams minutes played (PP + PK). from even strength minutes. So I went to look for differences in even strength TOI between 3rd line players AND 2nd line players. This was not as easy as it sounds because almost every team has mixed their lines a bunch to this point. Fortunately for me, Toronto hasn’t. Legion LW Koda Adok has played all 26 games on the 3rd line and racked up 293 even strength minutes. Legion RW Matthew Kai has played all 26 games on the 2nd line and racked up 298 EV minutes. This might seem a bit cherry picked and anecdotal I guess, but the general pattern is pervasive; 3rd line players play pretty much the same amount of minutes as 2nd line players.

 

So the third line matters. I’d even go as far to say as it matters as much as the second line. Traditionally in hockey, this isn’t that unusual. But given the VHL’s limited amount of skilled players to put on lines I ask myself: “is it better to re-use your top guys on the third line or to play different, less skilled guys?”. I won’t attempt to answer that in this media spot because I don’t think I have the supporting data to answer it just yet. But it is something to keep in mind and would be curious if anyone else has any thoughts on the matter.

 

Ok, we now have a better idea about which lines are used and how often they are used on aggregate. What about their matchups? If you were trying to build the perfect “first line”, would you only have to worry about beating the other teams first line? Or would you have to account for matchups all over. The answer is that all lines eventually will play against all opposing lines, but some are much more frequent than others. Take a look at the charts below, the first showing even strength matchups and the second special teams matchups.

 

 

Some quick observations from these: 2nd lines do not regularly match up against other 2nd lines. I have no idea why this is, and perhaps most of these “off-line” matchups are short-shift ones but it still shows that a team will get many chances against your second line with their 1st and 3rd lines in a game, and it’s something worth accounting when building these lines. Even more interesting is how the third lines are used; just under two thirds of the time they are facing a higher line. If you're not going to stack the third line with your top guys, you better make sure they can at least defend well against them.

 

Looking at special teams matchups, the matchup differences become more subtle. Each #1 line plays a pretty fair share of time against the opposing #1 or #2 lines. One little insight I gained from this is that your 4v4 lines become your 4 on 3 power play lines when that situation arises (~7% of the time they play). Not sure if that warrants trying to make those lines a little more offensive or not.

 

I’ve well surpassed 1000 words now. Time to stop writing.

 

 

will claim for 12/1 and 12/8

Edited November 27, 2019 by studentized
typos

[Berocka 2,191]

3 hours ago, studentized said:

BUT in fact does not even impact the sim because we sim on the 1.1 engine (thank you to devise for confirming)

@GustavMattias

[Enorama 1,980]

5 hours ago, studentized said:

one of the things added to 1.5 that is NOT present in 1.1 is the ability to control line strategies/time on ice. This is a feature that GM’s appear to be able to control and configure already due to the STHS client we use (check out our index) BUT in fact does not even impact the sim because we sim on the 1.1 engine (thank you to devise for confirming).

 

what the fuck

[Beketov 8,642]

We are testing other engine versions but most of our testing has been focussed on 2 and 2.1 and trying to make them viable. 1.5 is something we never considered tbh, probably because we never looked into what it actually brought to the table. It is worth some testing to see if it doesn’t skew things for us just to make sure we get as well rounded an experience as we can.

 

Do you have a source from STHS regarding the 1.1 vs. 1.5 claims?

[studentized 166]

 

52 minutes ago, Beketov said:

We are testing other engine versions but most of our testing has been focussed on 2 and 2.1 and trying to make them viable. 1.5 is something we never considered tbh, probably because we never looked into what it actually brought to the table. It is worth some testing to see if it doesn’t skew things for us just to make sure we get as well rounded an experience as we can.

 

Do you have a source from STHS regarding the 1.1 vs. 1.5 claims?

 

Closest sources are those 2 links from STHS:

 

6 hours ago, studentized said:

read here (and confirm here)

 

Still a chance those are out of date I suppose (STHS not very transparent in general) but the data also supports it a bit so I believe it is true.

[studentized 166]

Looked into it even more and in regards to PS and EN not being used, it may not be the case. Those are both v2 features but they pop up around 2008/9. STHS docs say 1.1 stopped getting feature updates in 2010 so its possible those apply to all sim engine versions

[Beketov 8,642]

35 minutes ago, studentized said:

Looked into it even more and in regards to PS and EN not being used, it may not be the case. Those are both v2 features but they pop up around 2008/9. STHS docs say 1.1 stopped getting feature updates in 2010 so its possible those apply to all sim engine versions

We max EN for all players regardless so that one doesn’t matter however I do believe PS has an effect, however low, in our version.

[.sniffuM 1,741]

Back when Jardy was the simmer, PS wasn't an upgradable attribute in the VHL, so he'd give every player a PS rating based on some combination of their skating, scoring and puck handling. As far as I remember it did make some difference.

[Rayzor_7 741]

All I'm hearing is it's time to move to 1.5 testing

[Fire Tortorella 2,653]

25 minutes ago, .sniffuM said:

Back when Jardy was the simmer, PS wasn't an upgradable attribute in the VHL, so he'd give every player a PS rating based on some combination of their skating, scoring and puck handling. As far as I remember it did make some difference.

 

Tbh i dont even think 8t was that sophisticated...pretty sure he just made them up.

[.sniffuM 1,741]

1 hour ago, flyersfan1453 said:

 

Tbh i dont even think 8t was that sophisticated...pretty sure he just made them up.

 

He said he based them on those three attributes though. Are you calling Jardy a liar?!

[Fire Tortorella 2,653]

49 minutes ago, .sniffuM said:

 

He said he based them on those three attributes though. Are you calling Jardy a liar?!

 

"Based on them" meaning he looked at them and then shot gunned a value for PS...just meant that I don't think he had a strict formula or anything, since that was one of the original ideas when we added PS as an upgradable attribute in the first place.

[Josh 1,597]

6 hours ago, Beketov said:

We are testing other engine versions but most of our testing has been focussed on 2 and 2.1 and trying to make them viable. 1.5 is something we never considered tbh, probably because we never looked into what it actually brought to the table. It is worth some testing to see if it doesn’t skew things for us just to make sure we get as well rounded an experience as we can.

 

Do you have a source from STHS regarding the 1.1 vs. 1.5 claims?

 

• There is no referee and linesman.
• Line strategies do not apply.
• The percentage of ice time per lines does not apply.
• Team strategies do not apply correctly. Only the “If the score is equal then strategy is” applies all game long.
• The "Last Min" Lines does not apply.
• Player StarPower does not apply.
• 3 vs 3 Overtime
• Custom Overtime Lines

These are the features that are unavailable in 1.1. I was actually super surprised when I was reviewing this other day and assumed we were using 1.5 based off a note Quik made during some 2.0 testing.

 

Under league options:

http://sths.simont.info/ManualV2_En.php 

[Rayzor_7 741]

8 minutes ago, Josh said:

 

• There is no referee and linesman.
• Line strategies do not apply.
• The percentage of ice time per lines does not apply.
• Team strategies do not apply correctly. Only the “If the score is equal then strategy is” applies all game long.
• The "Last Min" Lines does not apply.
• Player StarPower does not apply.
• 3 vs 3 Overtime
• Custom Overtime Lines

These are the features that are unavailable in 1.1. I was actually super surprised when I was reviewing this other day and assumed we were using 1.5 based off a note Quik made during some 2.0 testing.

 

Under league options:

http://sths.simont.info/ManualV2_En.php 

I feel like as GM's we should be rioting

 

[Tagger 2,722]

Funnily enough I made a podcast a month and a half ago pointing out that the line %'s and team strategies didn't work in the version of the sim we were using. Think the title of "Things you might not know about" is somewhat fitting given some of the reactions to this article 

 

I remember bringing up 1.5 as an option in the BOG back when they were first considering a sim engine change and I believe Will had said he'd done a few tests and the results were fairly similar to that of the current sim 1.1. I think the SHL also uses 1.5 given that their index has the effects of line %'s and strategies displayed but no-one has any giveaways and takeaways (a V2 stat). 

 

As for the PS and EN stuff, yeah it's difficult to tell if it does have any effect or not in 1.1. Those attributes aren't in the Version 1 manual but like @studentized says whether they were added in after the fact I'm not sure. It doesn't help that there are two different STHS items with V's (the sim engine and the Program itself) so it's difficult to tell which he's talking about (unless he says V3 since I don't believe there's a V3 sim engine yet but there is of the Program). I guess we could always ask him on the forums see if he remembers. 

[Beketov 8,642]

30 minutes ago, Tagger said:

unless he says V3 since I don't believe there's a V3 sim engine yet but there is of the Program

There’s no V3 engine so you are correct; if something says V3.x it’s talking about the program itself not the engine.

[.sniffuM 1,741]

1 hour ago, flyersfan1453 said:

 

"Based on them" meaning he looked at them and then shot gunned a value for PS...just meant that I don't think he had a strict formula or anything, since that was one of the original ideas when we added PS as an upgradable attribute in the first place.

 

I didn't mean to imply that he had a strict formula if I did, more that I meant to imply that I didn't know exactly what basis he went on to determine what he set it at. But I do 100% remember saying he considered those three attributes in determining a guy's PS attribute.

[studentized 166]

Claiming for second week