Simulation engine versus 2d Engine

[mboucher90]

Congrats for the addition of the 2d engine, looks great !

i got one question about the "simulating" option choice in the "game settings" tab :

If i choose de classic instead of the 2d engine, does the game simulated will have the same result, but just quicker.

Like if i sim one game in the simulation game and the same game with the 2d engine, at the end it will be the same game with same penalty at X moment, goal at X moment etc...?


Thanks,

[Shadow4u]

Nope - nor should you expect it to be. Every time a sim is done - you get a different result. I understand the question and honestly do not know the answer. Given that, does it really matter ?

[Bones26]

It is indeed an interesting question.

As I would presume that the outcome of any match is ultimately resolved by the basic underlying logic as established by the team / match parameters contained within the hard coded game engine programming (o's and 1's), unlike the inherent unpredictability that would otherwise be seen in real life human matches, then would it not be reasonable to assume that replaying (i.e. repeating) the same identical match with exactly the same identical game parameters, should be expected to provide the same game outcome in the vast majority of cases, and that irrespective of the viewing format used?

Otherwise, how does one go about, with any degree of confidence, in addressing a teams "G.M. / coaching" options if in the end, the final match results are randomized to such an extent that under the conditions described above, one could conceivably win one match 5-2 and then upon reply (again all with the same game and team parameters as the set out in the initial 5-2 game) you lose that replayed match 2-5?

I can appreciate that there might be some randomization on the periphery built into the game engine calculations (i.e. a probability factor for the frequency of certain penalties for example) but I would think that those instances should not be such a determining factor so as to render game outcomes in the end the same as a random dice roll.

Then again, I'm not a programmer, so perhaps I'm missing something here, but it does give one something to ponder...

[tward13]

Originally Posted by Bones26

It is indeed an interesting question.

As I would presume that the outcome of any match is ultimately resolved by the basic underlying logic as established by the team / match parameters contained within the hard coded game engine programming (o's and 1's), unlike the inherent unpredictability that would otherwise be seen in real life human matches, then would it not be reasonable to assume that replaying (i.e. repeating) the same identical match with exactly the same identical game parameters, should be expected to provide the same game outcome in the vast majority of cases, and that irrespective of the viewing format used?

Otherwise, how does one go about, with any degree of confidence, in addressing a teams "G.M. / coaching" options if in the end, the final match results are randomized to such an extent that under the conditions described above, one could conceivably win one match 5-2 and then upon reply (again all with the same game and team parameters as the set out in the initial 5-2 game) you lose that replayed match 2-5?

I can appreciate that there might be some randomization on the periphery built into the game engine calculations (i.e. a probability factor for the frequency of certain penalties for example) but I would think that those instances should not be such a determining factor so as to render game outcomes in the end the same as a random dice roll.

Then again, I'm not a programmer, so perhaps I'm missing something here, but it does give one something to ponder...

No, there have been times in online leagues where something causes the commish to resim. Even though the change may be only for one team all the results for unaffected teams are not the same. Various randomizations coded into the game will cause cascading changes. As with real life, there are expectations, but also a level of randomness in all events.

[clamel]

I see it as every game, 2d or similated, actually is a massive amount of dicerolls.
If you play out a 2d game the dicerolls should come as you play along and then the players might move towards that.

If one do have a lot of dicerolls and each and everyone then directs the game to a next depending on the previous it's hard to see you almost never get the same results. That's with all those stats in the game. Shots on goals, faceoffs, shots blocked, shots the goalie saves. Will a penalty accur or not in a situation.

That's my thinking when you see similar games like the baseball or FM, EHL, or any sport simulation.
Then I'm not sure but on a "moving" sport like hockey, football or basket must have so many more incidents than a baseball game. So many factors and hopefully we have a diceroll for each ot them

[Bones26]

So if we accept that there are likely randomizations coded into the game and that those could cause cascading changes, then isn't it important to have some understanding how these randomizations are designed to play out.

Look I'll be the first to admit that this is no doubt a blind spot on my part that I'm having difficulty wrapping my head around. I'm stuck on the premise that if this cascading randomness is in fact an integral aspect of the game play, then wouldn't it follow that any player, line formations, or tactical changes and so on introduced by the player would inevitably be without any discernable consequence / feedback for that player, as they would be inevitably overshadowed by this same cascading randomness in determining the final outcome.

One of the more common shortcomings expressed regarding FHM6 was that there was a lack of any discernable feedback regarding ones management / coach inputs that one could use to make future adjustments. In short, the results often seemed random...

[tward13]

Originally Posted by Bones26

So if we accept that there are likely randomizations coded into the game and that those could cause cascading changes, then isn't it important to have some understanding how these randomizations are designed to play out.

Look I'll be the first to admit that this is no doubt a blind spot on my part that I'm having difficulty wrapping my head around. I'm stuck on the premise that if this cascading randomness is in fact an integral aspect of the game play, then wouldn't it follow that any player, line formations, or tactical changes and so on introduced by the player would inevitably be without any discernable consequence / feedback for that player, as they would be inevitably overshadowed by this same cascading randomness in determining the final outcome.

One of the more common shortcomings expressed regarding FHM6 was that there was a lack of any discernable feedback regarding ones management / coach inputs that one could use to make future adjustments. In short, the results often seemed random...

There is randomness, but they happen for the most part within the realm of expectations. Sure in any given replay McDavid may or may not score, but given enough replays of a particular game he will score a reasonable number of times.

[KevT90]

I believe the question was whether the simulator(s) would use the same parameters whether you are using the new 2D engine or simulating without it ? Not sure exactly how it's done in the upcoming FHM (I did not watch in full the videos released speaking about the new contents). Is there more than one engine now ?

For instance, Eastside Hockey Manager gives you the option to use a full sim engine or a quick sim engine. Parameters used to perform the simulation are not the same, therefore both engine generate stats that differ substantially from one another.

[JeffR]

Originally Posted by KevT90

I believe the question was whether the simulator(s) would use the same parameters whether you are using the new 2D engine or simulating without it ? Not sure exactly how it's done in the upcoming FHM (I did not watch in full the videos released speaking about the new contents). Is there more than one engine now ?

For instance, Eastside Hockey Manager gives you the option to use a full sim engine or a quick sim engine. Parameters used to perform the simulation are not the same, therefore both engine generate stats that differ substantially from one another.

There are three engines, really: the new 2D one, the old one that we're calling "Classic" now, and the one for nonplayable teams.

2D is the slowest, but prettiest to watch. I wasn't sure if we'd be able to get close to the old engine's stats generation in time for release, but after Sebastian and I spent a lot of time tuning it in the last couple of weeks, it's just as good in most categories and better in some (notably fighting.) As of this evening, shots and goals are running a bit under 10% too high, there's a slight tendency for the very best players to produce a little too much, and the assist:goal ratio has crept up a little too high. I'm hoping to get all of those knocked into shape overnight. So it'll be a little better than I had anticipated. Still, it sims noticeably slower than the old engine, hence the option to use the old one if you want.

I'll assume everyone's familiar with how the old engine works; we've changed very little there, since eventually, once we get to the point where the new one can run as fast as the old, it'll be going away.

The nonplayable system just generates realistic stats for teams that aren't playing, adding to the totals of players on those teams on a weekly basis. It's essentially the equivalent of EHM's quick sim, but is much faster at the cost of a lack of standings/game results in those leagues. EHM's, as you noticed, can generate some very different numbers; if I remember correctly, Graeme Kelly adapted it from Football Manager and I told him where the numbers needed adjusting to get to something like a semi-realistic state. When we built our nonplayable system, I tried to avoid some of the failings of the EHM one.