Strategic warfare graphs

[dizzle3]

Looking back at the TFH strategic warfare graphs, I think it would be much more helpful to show estimates of enemy production and strength on the same graph, instead of just the losses (even if they are just estimates for numbers that you can't be certain of like enemy production). That way you can make a much better judgement of how your strategic warfare campaign is going.

As an example, this would be helpful as both the allied player and the german player, and shows the effect of the submarine war in 1941 on allied shipping:



It would also be helpful for both sides to see how the U-Boat fleet is holding up to attrition to just the effectiveness of wolfpack/ASW measures (this is the german submarine fleet in 1943)


Other items I think would be useful:
-Manpower gains/losses and manpower pools, for both your own country and the enemy
-Same thing for tanks
-Fighter aircraft production/losses and total fighter strength
-Bomber production/losses and total bomber strength
-Some measure to cope with strategic bombing, for example total production of the target country versus bomber production/losses

Thoughts/improvements/other interesting ways to look at this?

 

[scroggin]

I like it

 

[Mattias]

+1 IMHO a quite informative, yet simple way of modeling the guesswork of the intelligence-officers. The quality of their Estimates should of course improve with better techs! /M

 

[dizzle3]

+1 IMHO a quite informative, yet simple way of modeling the guesswork of the intelligence-officers. The quality of their Estimates should of course improve with better techs! /M

Yeah, I would tend to think (although nothing is revealed yet about espionage) that the reliability of these estimates would be improved by spies.

However, some things were not at all secret, for example total merchant marine size at the start of the war was practically a matter of public record. Once war starts, things should get a little fuzzy, but generally naval production should be pretty well estimated by a competent spy and reconnaissance network

And obviously no uncertainty should be needed when looking at your own figures, these should be certain.

 

[ingwe]

I think it is great! I would also like to see error bars on estimated items so that we know how accurate we think they are. Or something like that at least. If I have no clue how many subs Germany is producing so my intel is just a guess, I want to know. If I know that Russia produced exactly 2340 tanks last month, I also want to know that. Error bars could also be updated so that at time passes you have a better understanding of what was happening but in the moment you know less.

 

[dizzle3]

I guess on the strategic bombing side, something like this:

Yellow line is total production of the target country
Red line is how much damage was done to IC in the month
Blue line is how much IC was repaired in the month

Obviously, these would be highly uncertain for enemy industrial estimates. I worry that it might begin to look a little cluttered with error bars., but some smart way of showing uncertainty would be helpful

 

[Dalwin]

Yes I would like to see something along these lines.

 

[seattle]

I can only repeat myself: Since DH I fell in love again with the HoI-franchise. Finally I could wage a comprehensible Atlantic War. At any time due to the loss stats, I could see how effective my subs were in terms of sunk tonnage and loss ratio. With such analytical tools you can effectively plan and maintain a strategic warfare campaign. Graphs are even better. They have to work obviously, unlike the broken HoI3 graphs (which were way too tiny anyways).

 

[ingwe]

I guess on the strategic bombing side, something like this:

Yellow line is total production of the target country
Red line is how much damage was done to IC in the month
Blue line is how much IC was repaired in the month

Obviously, these would be highly uncertain for enemy industrial estimates. I worry that it might begin to look a little cluttered with error bars., but some smart way of showing uncertainty would be helpful

Agreed. Clutter would be something to be concerned about. I think if the error bars were a shaded yet translucent region so that they were more in the background but still visible it could work.

 

[dizzle3]

Like this probably is what we are both thinking.



The line is the best estimate, but the fan represents uncertainty.

I would probably restrict the fan to only be on the yellow lines, the incremental monthly error bounds aren't as interesting or useful and will tend to add clutter

 

[DocMorningstar]

God yes, I would love a little more attention put into the infographics about war progress. A picture is worth 1000 words. You can see, for instance, if your material usage is increasing or decreasing, and the rate at which it is changing.

 

[dizzle3]

Like this perhaps as a latest offering

 

[Dalwin]

You would not really even need to distinguish any line to show error range. As long as we know that enemy loss figures, for example, are an approximation and the system randomly alters the actual figure and posts that altered figure instead of the actual; we are fine.

There would be no need to clutter the graph.

 

[dizzle3]

Yeah, that's one possibility.

The other that I've shown above is to give a range around the most important number and to show the "average" of the others.

In effect what is shown is "we think the enemy has between 27.5 and 30m tons of shipping left, our best guess being 28.6m. We think that we sunk about 600k tons last month and that the enemy made 100k tons"

I guess I'm saying that the uncertainty around the incremental numbers is far less interesting than that around the cumulative, so to avoid clutter it is better just to show our best guess for the incrementals

 

[dizzle3]

You would not really even need to distinguish any line to show error range. As long as we know that enemy loss figures, for example, are an approximation and the system randomly alters the actual figure and posts that altered figure instead of the actual; we are fine.

There would be no need to clutter the graph.

The concern with this is that it gives no indication of uncertainty around the estimate.

In hoi3, when spies reported 800 manpower for the enemy, the problem was not that it wasn't exactly right, more that you had no idea as to how wrong it could be. More informative would be "there is a 90% chance that their manpower is between 600 and 1000, with our best guess being 800" is far more helpful. Even if the real number is 650, or even 550, it's good to have an indication of the potential range

 

[Axe99]

The concern with this is that it gives no indication of uncertainty around the estimate.

In hoi3, when spies reported 800 manpower for the enemy, the problem was not that it wasn't exactly right, more that you had no idea as to how wrong it could be. More informative would be "there is a 90% chance that their manpower is between 600 and 1000, with our best guess being 800" is far more helpful. Even if the real number is 650, or even 550, it's good to have an indication of the potential range

Aye, but unlike statistical estimation, there's no standard deviation/error on an intelligence estimate. It's not a sample in the statistical sense that those error bars are usually used (well, unless an awful lot of spies are deployed ). The other thing is once you've got three lines and three error bars, it will start getting a bit messy. I do like the idea overall .

 

[1alexey]

+1.
Error should probably be just noted under the graph, like:
"our bad spy network doesn`t allow us good estimates, expect error up to 80% of the value"
"our great spy network, code breaking and local resistance fighters provide pretty reliable intel, expect error to be less than 20% of value"
ex.

 

[PanH]

Yeah, there should be the error shown like e.g. 20% (real value can then be -20% to +20% of the displayed value).

 

[Jazumir]

In that sense, the brightness of the lines could indicate certainty of information, with exact numbers of incertainty given in a tooltip.

 

[Cybvep]

TFH had the right idea with strategic warfare graphs. Bad execution, though :C.