Annoucement

[TorpX]

This is just a short post to tell everyone in this neck of the woods that I uploaded an exterior ballistics program to Subsim. It is explained briefly HERE. Anyone interested in artillery, ballistics and the like, should check it out.



************************************************** ************************************************** *******

AutoTarg_1.0

Also, I finished a small mod for Auto-Targeting play in SH4.

It uses the Torpedoes_US.sim file to incorporate a small error to the gyro angle. This simulates routine human errors involved with computing a firing solution.

As the game comes to us in stock form, we have essentially 3 ways to play:

1. manual targeting, no map-contacts
2. manual targeting, with map-contacts
3. auto targeting

None of these options are ideal. In the first case, we are likely to have errors aplenty, range and AoB, often being very uncertain. This is realistic, but also laborious. Many don't like the work involved, or are not especially good at math. In the second case, errors are likely to be almost eliminated, unless we make a mistake in our calculations/measurements. This would be realistic if Ubisoft had designed some human factors into the process, but they did not. In the third case, errors are zero, since Ubisoft didn't see fit to model any errors in the plotting, or observations. This makes the game highly playable, but also very unrealistic. So, what can be done?

With AutoTarg, you at least have the option to add in some errors for the last two cases. Ideally, the errors would be a function of the particular skills of your crewmen, the circumstances present, weather, and such, but we are limited as to what we can do here.

AutoTarg_1.0 adds some small errors to the gyro guidance factors. I was able to make the error change with the date, so your accuracy will improve as the war goes on.

[Sailor Steve]

Cool! I certainly appreciate this, as I fall into category 3.

[Rockin Robbins]

Originally Posted by TorpX

None of these options are ideal. In the first case, we are likely to have errors aplenty, range and AoB, often being very uncertain. This is realistic, but also laborious. Many don't like the work involved, or are not especially good at math. In the second case, errors are likely to be almost eliminated, unless we make a mistake in our calculations/measurements. This would be realistic if Ubisoft had designed some human factors into the process, but they did not. In the third case, errors are zero, since Ubisoft didn't see fit to model any errors in the plotting, or observations. This makes the game highly playable, but also very unrealistic. So, what can be done?

The ideal way to fix the problem would be to have a scalable error envelope for each observation method with your measured number modified by a random number within the error envelope. That way you could have radar accurate to within 30 yards, visual within a certain percentage, stadimeter the same, sonar within a certain percentage with error envelopes tailored for each data acquisition method.

Then there would be the random errors in torpedo performance and user input that you've dealt with.

Too bad we can't do anything about the first part, data acquisition and measurement errors.

[Barkerov]

Originally Posted by Rockin Robbins

[/INDENT]The ideal way to fix the problem would be to have a scalable error envelope for each observation method with your measured number modified by a random number within the error envelope. That way you could have radar accurate to within 30 yards, visual within a certain percentage, stadimeter the same, sonar within a certain percentage with error envelopes tailored for each data acquisition method.

Then there would be the random errors in torpedo performance and user input that you've dealt with.

Too bad we can't do anything about the first part, data acquisition and measurement errors.

It would not be a difficult calculation to have random errors different for each mode of detection. All you need is the variance of each method and a random number generator.

What I would like to see over and above what you said is the error reduce as the target is observed for longer. So if you shadow it from a distance for a few hours even with visuals only you can get a much better idea of the course and speed.

[TorpX]

Originally Posted by Rockin Robbins

[/INDENT]The ideal way to fix the problem would be to have a scalable error envelope for each observation method ....

Too bad we can't do anything about the first part, data acquisition and measurement errors.

Yes, this would be very desirable.

The errors could be incorporated before the contacts were plotted. We could have the choice as to whether to accept the plot at face value, or adjust it ourselves according to our own judgment.

Originally Posted by Barkerov

It would not be a difficult calculation to have random errors different for each mode of detection. All you need is the variance of each method and a random number generator.

What I would like to see over and above what you said is the error reduce as the target is observed for longer. So if you shadow it from a distance for a few hours even with visuals only you can get a much better idea of the course and speed.

The problem is less one of mathematics, than software design. We can do the math, but lack any way to fix the code. I was also thinking about the number of observations, and such. Believe me, if I had a way to put this in the game, I would.

I think the best we can do, with the game as it is, would be to craft a house rule, using a table to factor in the most important factors affecting our approach. These being RADAR, no. of observations, visibility (day or night), skill level (skipper or crew).

[Barkerov]

That is always going to be the limiting factor, we are stuck with the basic framework the software was designed in. Despite that there has been some amazing work done to improve the game as much as it has been and your efforts along with others have definitely not gone unnoticed

I was just daydreaming about what could have been.....

[Rockin Robbins]

But we really waste our time pining for what could or should have been when what we have is so amazing. Yes, SH4 is full of "problems" and yes, Silent Hunter 4 is the best diesel-electric submarine simulator on the planet. Now we can fuss over the first part and I think we should, just a little. But we should much more celebrate the grand achievement that Silent Hunter 4 is.

[TorpX]

Originally Posted by Barkerov

I was just daydreaming about what could have been.....

Yeah, I do that too.