The Concentrated Fire Problem

[crackaces]

In thiis forum there was a posting about concentrating fire and wondering about would not more ships concntrating fire even if there is a penalty be better since it would at least insure a hit?

I propose that the problem might be reduced into a "NOT" equation. That is what is the probability that my ships will miss instead of calculalting hits. Bullethead has posted that ships at range have a general base probability of 5%. I do not have access to the code so I am making up numbers, but you can see the trend. If the model is posted someday you can fill in the numbers.. Let us say there is a 50% penality for more than 2 ships so the probability to hit goes from 5% to 2.5%. That means that every shot has a 97.5% chance of missing. So the probaility of missing is (97.5% * 97.5 * 97.5) since every shell must miss else there is at least one hit. Thus in this model 50% penality for multiple ships means three ships firing have a about a 93% chance of missing or 7.3% chance of hitting. In a way we have reduced the effectiveness of each round to about 2.4% chance of hitting although we have increased the probability of any of the shells hitting by a mere 2%. Four ships brings this number to an overall 90% chance of missing or 10% chance of hitting at the cost of 2.5% for each shell. That does not consider the extraneous variables such as the penality for being under fire vs. not.

Maximzing the effect of each round is particularly important with the light ships since the loadout is so limited. If only 5% of the shells are hitting and the ammo load for each gun starts at 100 rounds it is clear that about 5 rounds per gun under the best of conditions on average are going to possibly cause damage.

One additonal thought is that the long campaign game is a war of attrition. Maximizing the effect of each shell, each ship, each division, each task force so you inflict the maximum gain for the war effort at the minimum cost. Like any long project where things go wrong one day at a time until one day disaster .. the long campaign game is a culmative summation of decisions. Spread the fire out

Thoughts?

[Bullethead]

Originally Posted by crackaces

Thoughts?

First off, let me state that in our games there is no "base accuracy". Fire control systems of the day were highly dependent on human judgment which varied much in quality both intrinsically (as shown by BCF vs. GF shooting) and over time by the same ships during the same battle. The result is that in any single battle or part of one, an individual ship might shoot very well or very badly, or (more likely) have periods of both. Thus, "base accuracy" is only a long-term average at best, more of a rule of thumb. It's not something to rely on at any given time.

That said, back to the topic....

There was much argument on this subject during WW1, but the adherents of over-concentration (which I define for purposes of this discussion as too many shooters at one target to do good spotting) failed to realize 1 thing. That thing is that you can't reduce the problem to a "NOT" equation because each firing ship wasn't an "independent trial". IOW, the underlying assumption that each ship would have the same base accuracy and suffer the same accuracy degradation, from which odds could be figured, simply wasn't the case.

The main problem was that real ships didn't really know the true range, course, and speed of the target. IOW, unlike the game where you can find this out with the cursor, real ships had to estimate. They periodically took observations from multiple instruments, each of which had a significant level of irreducible mechanical error before you consider potential operator error, and averaged the results. And they had to do this multiple times and draw a best-fit line through the results to see any trends developing.

The result of all this was that each ship had its own idea of where the target would be when the shells came down. Sometimes they were dead on, sometimes they were way off. And the only real feedback they had to correct their errors was observing their shell splashes, because the range they were fired at was the ONLY thing the firing ship really knew for certain. But then, of course, there was also rangefinder and operator error in seeing how far off the splashes were.

As such, it was of vital importance that a WW1 ship be able to tell which splashes were its own. If it couldn't do that, then it really had no way of telling how much to correct, or even whether to correct at all. This is why accurate shooting was pretty much out of the question if more than 2 ships were firing on the same target at ranges were spotting was the key to the problem. Due to the time that splashes lasted, the 2 ships had to stagger their fire so that one's splashes would be gone before the other's appeared. Thus, if more than 2 ships were firing at the same target, none of them could tell whose splashes were whose, and none of them could shoot well. So if a ship was way off the range, it really couldn't tell, and would continue to miss by a wide margin, meaning it would contribute nothing to the overall hit rate, regardless of mathematics.

The Russian Black Sea Fleet from the get-go, and the RN after Jutland, both developed ways of concentrating 3-4 ships on the same target during WW1. These methods worked by removing the need for each ship to spot her own splashes. Instead, there was a designated master ship that did all the calculations and the other ships fired at ranges a bit above and below that of the master ship. The idea was to shotgun the target with what was in effect a single huge salvo, spotted as group, and hopefully wide enough to straddle at some point. It was hoped that, because of the known range offsets of individual ships, the true range would be quickly determined and subsequent group salvos tightened up around that point. Neither system was really tested in combat, however.

[crackaces]

Thanks for the great feedback Bullethead!

I do not want to get into a disscussion where you inadvertently disclose algorithums or game mechanics. But, what happened in 1916 has to be modeled into 1's and 0's in 2011 for it to work on my PC.

This modeling could be anything between generating a random number that deduces hits (I highly doubt that is what is going on in the code ) to various methods like generating a base accuracy (which could be random number based so it is not fixed) and modifying that number depending on factors like crew training, weather, wind, distance, damage, waves ..etc. Clearly a lot of thought has gone into the design and my suspicions are that the game has taken a lot of work to get more toward the latter but I obviously do not have access to code.

BTW) My thoughts on some base accuracy came from:

As has been stated above, gunnery at Jutland ranges was pretty bad, even with the best fire control systems of the day. It averaged 3-5% for well-trained ships and was considerably lower for British attlecruisers. Using the default slider value (20) for gunnery accuracy should give you results of a similar nature (http://jutland.wikia.com/wiki/Gunnery, para. 1)

which I believe I incorrectly attributed to you (Bullethead) my appoligies -- and I think VanDer Tann made that quote?

Besides that I assumed independent trials... there is one other failing in my math, and that is if the penality increases with each ship that targets. I hint at this because I sense now that this is a depedent (on the other ships) calculation.

Anyway -- thank you again for the feedback!

[Bullethead]

Originally Posted by crackaces

BTW) My thoughts on some base accuracy came from:
...
which I believe I incorrectly attributed to you (Bullethead) my appoligies -- and I think VanDer Tann made that quote?

No, I said that. But those "base accuracies" are, as I've said, long-term averages. It's quite possible that any given ship, even an RN BC, could shoot rather better than that for a while, or for even the historically best ship to shoot considerably worse for a while. That happened in real life. Do you have Campbell's Jutland: An Analysis of the Fighting? This books tracks nearly every shell, at least between capital ships. You can see this sort of time-based streakiness in there.