Carrier Battles for Guadalcanal Forum Index

Carrier Battles for Guadalcanal


 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister 
 ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 

Search major improvement
Goto page: <  1, 2, 3, 4, 5, 6  >
 
Post new topic   Reply to topic    Carrier Battles for Guadalcanal Forum Index -> Topics -> New Features
Previous topic :: Next topic  
Author Message
larsenjp
Beta-testing

Online

Joined: 19 Aug 2017
Posts: 566

PostPosted: Fri 1 Dec - 21:04 (2017)    Post subject: Search major improvement Reply with quote

Bladerunner wrote:
However, as I wrote, events are not independent as each scout follow a defined path without overlapp.
So if one scout is a sector where the nothing to search, then other scouts have a greater chance to find


Yes you are right but you cannot know in advance which one has a better chance to find... Therefore, for the admiral, each scout has the same chance to find.
But it's obvious that wisely designing search patterns increase the probability to find.

Bladerunner wrote:
Very simple example: let's say you are looking for a TF which could be in 2 hexes.
You have 2 scouts and each oh them may scan 1 hex.
If you scouts were sent randomly, then the chance to find-d would be 1 - (0.5 x 0.5) = 75%
However, as a wise admiral, you will send each in a different hex, thus achieving 100% chance to find


Well let's say that you have 1000% chance of getting a plane inside the hex occupied by the TF.
But to me this is not enough; then the scout has to spot the ships which may not be automatic...

By the way, a visibility of 30 miles, that is almost 55 km seems really very high.


Back to top
Publicité






PostPosted: Fri 1 Dec - 21:04 (2017)    Post subject: Publicité

PublicitéSupprimer les publicités ?
Back to top
SBD


Offline

Joined: 28 Oct 2017
Posts: 55
Localisation: UK

PostPosted: Fri 1 Dec - 21:29 (2017)    Post subject: Search major improvement Reply with quote

Thank you both, whilst I am not disagreeing with the statistics, I wonder if they are applicable.
Is the solution is simpler?

The chances of spotting a TF decrease in direct proportion to the distance from the carrier,in this game context,
so at row 7, in our example, we have one plane in each of the hexes representing the arc of 210 miles out. 
We apply a pragmatic  % of one plane type spotting a TF for that battle period within any hex.
This can be any figure we like within reason, and I suggest it is user variable, from the super-trained eagle eyes to the conscript bespectacled.
This can help also to balance a game, if necessary.


This is the Pacific, pre radar on planes, it can be luck - nothing is predictable in the normal sense.
I suggest each plane is independent, assuming they are on course, the nearer one plane is to a TF, the further another one is away.
I agree with Bladerunner, the chances are independent.
I disagree with him when he says a wise Admiral sends a different plane to a different hex.
A number of scouts would be sent out to a set distance within a specified sector- no micromanaging please
If need be, you increase the number of planes or narrow the sector.
But scouts are  valuable, at least in the USN, the more scouts the less dive bombers to attack.


Where I think maths is applicable, is in row 6 for example where there are more planes than hexes. (6:8)
As we are constrained by the hex construct, we will have >1 plane in 2 hexes.
How do we know which hexes have more than two planes in?
Eight flight paths could be drawn across the hexes, and the two hexes with the flight paths nearer their centre have two planes in it?
A nightmare to program? But once established that algorithm would be applied to all rows from the carrier.
Eg In row 2 from the carrier we have eight planes and three hexes, sadly some maybe outside the 60 degree sector.
Simplicity outside, possible complexity hidden 😊


I hope we are getting there.......?


PS Another factor besides weather, and task force size, is training.
The IJN often ignored clouds and flew through them without going lower


Back to top
SBD


Offline

Joined: 28 Oct 2017
Posts: 55
Localisation: UK

PostPosted: Fri 1 Dec - 21:47 (2017)    Post subject: Search major improvement Reply with quote

At only 1000’ the horizon is about 33 nautical miles, in perfect conditions of course.
That’s a radius of 15 miles to port and starboard, at about 130mph, not too onerous even for one crew?


Archerfish, any comment please on modern visibility, particularly for the slower ASW, or is it all instruments now?


Bladerunner, if 8 planes are sent into a say a narrow 60 sector, some hexes have >1 plane,
how do you handle this at present?
Similarly if few planes are sent into a wide sector, where are there are less planes than hexes?


Back to top
larsenjp
Beta-testing

Online

Joined: 19 Aug 2017
Posts: 566

PostPosted: Sat 2 Dec - 00:21 (2017)    Post subject: Search major improvement Reply with quote

SBD wrote:
At only 1000’ the horizon is about 33 nautical miles, in perfect conditions of course.
That’s a radius of 15 miles to port and starboard, at about 130mph, not too onerous even for one crew?


This is the meteorological visibility. The question is: are you able to see the target within this maximum range.
I think it is much depending on the size of the target and possibly some others factors like contrast. I read somewhere that, at great distance from a plane, people do not really spot ships, they just spot wakes, just because the ships are not visible to human eye (i am not speaking about radar here).
Most probably you will be able to spot a liner or a carrier, most probably you won't be able to spot a PT boat or a sub...

I think it should be possible to estimate a value of the probability for an individual plane to spot a TF.
Then the problem will be to combine individual probabilities for all the planes belonging to the search pattern.


Back to top
Bladerunner
Administrateur

Offline

Joined: 25 Jan 2016
Posts: 1,278
Localisation: France

PostPosted: Sat 2 Dec - 10:02 (2017)    Post subject: Search major improvement Reply with quote

SBD wrote:
At only 1000’ the horizon is about 33 nautical miles, in perfect conditions of course.
That’s a radius of 15 miles to port and starboard, at about 130mph, not too onerous even for one crew?


Archerfish, any comment please on modern visibility, particularly for the slower ASW, or is it all instruments now?


Bladerunner, if 8 planes are sent into a say a narrow 60 sector, some hexes have >1 plane,
how do you handle this at present?
Similarly if few planes are sent into a wide sector, where are there are less planes than hexes?


30 miles visibility, does it mean 15 miles on each side or 30 miles on any side fo the aircraft ?


I does not matter how many scout are per hex.
I just compute the search mine length made by the distance and the angle (ex: 240 miles, 8 hex fo r60° at a range of 7) and divide by the visbility (30 miles x 2) to get the sporting chance of one airplane 


Back to top
USS Archerfish
Beta-testing

Offline

Joined: 16 Mar 2017
Posts: 823

PostPosted: Sat 2 Dec - 17:14 (2017)    Post subject: Search major improvement Reply with quote

SBD, my area of specialty is the flight deck since that is where I worked. Unfortunately, I didn't have an opportunity for fly in S-3s or P-3s so I can't really answer your question as to how easy it is to spot anything between 10,000 & 20,000 ft. (By the way, ASW is both visual and instruments but relies more on instruments)  On several occasion I did get to fly on Sea King helicopter.  At 30,000-35,000 feet you can about 220 miles but it's very difficult to see any ships at that altitude.  But even at 10,000 ft on a clear day can present challenges.  You cab see much better when the sun is behind your back and it's much more difficult when the sun is in front of you.  Also in the ocean there is often a light haze or fog.  It's more noticeable from the surface but this can reduce visibility too.  So there are a lot of weather and lighting variations that can affect visibility as well. 

Back to top
SBD


Offline

Joined: 28 Oct 2017
Posts: 55
Localisation: UK

PostPosted: Sat 2 Dec - 18:01 (2017)    Post subject: Search major improvement Reply with quote

Thank you Archerfish, I suspected that was the case, it’s infinitely better to hear from someone who has related experiences.
What a pity we can’t all go down to the pub and discuss our ideas and experiences....


Bladerunner, the theoretical horizon at 1000’ is >30 miles, so that ‘fits’ in nicely with hex size?
However there are many factors degrading the above, as discussed on this post.


No one can give a precise probility of a scout spotting a TF, 
so why not establish a % chance that feels right, and hopefully user variable, 
so if players want to change it they can do.
It’s a game Smile , not a simulation?


Bladerunner. 
Thank you for advising the specification of scouting with n scouts for x distance within a y degree arc.
I thought that was the pragmatic solution, before I emailed you, but I am not a programmer, though ex IT.
Please explain why the visibility (30) is multiplied by 2?


Are we close to agreeing/compromising on this?


Back to top
Bladerunner
Administrateur

Offline

Joined: 25 Jan 2016
Posts: 1,278
Localisation: France

PostPosted: Sun 3 Dec - 11:28 (2017)    Post subject: Search major improvement Reply with quote

Good point SBD


I am taking a pause here as the new system shall be implemented as a whole with shadowing and single plane
We will then see during testing what an acceptable search table 


Back to top
USS Archerfish
Beta-testing

Offline

Joined: 16 Mar 2017
Posts: 823

PostPosted: Sun 3 Dec - 19:26 (2017)    Post subject: Search major improvement Reply with quote

SBD, tell me the name of the pub and I'll be there!  Seriously.  I'll need some advance planning though.

Back to top
larsenjp
Beta-testing

Online

Joined: 19 Aug 2017
Posts: 566

PostPosted: Thu 7 Dec - 00:35 (2017)    Post subject: Search major improvement Reply with quote

Regarding probability of sighting ships from planes, i made some research and finally i found out this report:
http://www.dtic.mil/dtic/tr/fulltext/u2/214252.pdf

It was published in 1956 but was actually issued in 1946 and written by B. Koopman from the famous Operations Evaluation Group whose objective was to analyze scientifically naval operations and make recommendations to improve results. Theses guys, all high level mathematicians and physicians, just invented Operational Research.

The report is about searching and spotting ships, especially submarines, a major concern during WW2, and treats about search patterns (i.e. which one is the better regarding statistics) and about scanning, especially scanning through human eye; one full chapter (chapter 4) is dedicated to eye sighting.
Very informative but quite difficult to read.
Basically, the main things i understood are:
- human eye spots wakes, not ships since the contrast of the wake by respect to the sky in daylight is very high, about 0.5
- the maximum range of detection (noted Rm in the report) is much less than the meteorological visibility (noted V) and depends also of the height of the plane (noted h in the report, generally considered to be quite low by respect to V)
- the actual range of detection, noted R, may be even less because of some other influencing factors, especially the speed (or relative speed by respect to the target) of the plane and the angle of sighting.

The reason of this is just physiological: human eye can "see" only by periods of 0.25 second between two winks. To be sure, you need to scan the area more than 0.25 second. The higher the speed, the lowest the probability of detection.
Furthermore, the visual axis passes through the fovea, an area at center of the retina where retina cones are concentrated and where visual acuity is at its highest. It means that for targets near maximum range, the deviation between line of sight and the visual axis has to be almost nil i.e. the target must be "right in front". When the distance decreases, the angle increases and one can see targets "out of the corner of the eye" as it is said in the report.

The report concludes that, for eye sighting, the best solution is to scan only the 180° forward, leaving the central 90° to pilot and copilot and letting the other lookouts concentrate on the two remaining lateral 45° sectors... i.e. no scan behind! Otherwise, the scanning will be most probably concentrated just in the 90° or even the 45° central forward sector... i.e. just in front of the plane.

All in all, the probability of sighting decreases as the cube of the distance...
The report provides probability curves from which it is possible to calculate a probability of sighting...

I am trying to find some other reports to get more information.


Back to top
Bladerunner
Administrateur

Offline

Joined: 25 Jan 2016
Posts: 1,278
Localisation: France

PostPosted: Fri 8 Dec - 07:45 (2017)    Post subject: Search major improvement Reply with quote

Very good point jp !
Indeed I will implement this : 
  • Planes do not look back. so they only look on the side and in front

Besides being more realistic, it also solves me one puzzle. With the system of looking behind, a force at 7 hexes could be scanned twice (when the scout has a search radius of 6 and next phase when the radius is 8) but a force at 8 hexes could be scanned only once (when the search radius is 8). This was advantaging being at a even distance in hexes from search source


For the clue about the visibility, I would need to go over your document which I don't dare for the moment


Back to top
SBD


Offline

Joined: 28 Oct 2017
Posts: 55
Localisation: UK

PostPosted: Fri 8 Dec - 11:02 (2017)    Post subject: Search major improvement Reply with quote

Thank you jp
As mentioned, the only search radius information I had found is 50 miles, 25 starboard+25 port, deduced from Admiral Belliger’s comments pre Midway. 
Remembering 
He had limited planes so had to compromise on the thoroughness of the searches
The planes were slow moving PBYs.
They had dedicated three or four observers to the front and sides sitting in unrestricted bubble canopies.
They were looking for very large invasion TFs


Scouts, eg SBDs, will therefore have much lower chance of spotting than PBYs under ‘normal’ circumstances?
What is the search radius in the new version?
If it is still 30+30 may I suggest the chances of spotting are reduced accordingly?


I don’t understand the point of not spotting behind in the game, as this sort of detail cannot be reflected on this scale?
Though what was ‘behind’ on the outward flight, will be in ‘front’ on the return flight?


I do not want to appear ‘picky’ pedantic, as I have often said, it is a game, not a simulation;
but conclusions can be drawn from the real world.


PS A rear facing gunner in a SBD would find it difficult NOT to look behind, but again this is relative minutiae in our game scale?


Back to top
larsenjp
Beta-testing

Online

Joined: 19 Aug 2017
Posts: 566

PostPosted: Fri 8 Dec - 21:24 (2017)    Post subject: Search major improvement Reply with quote

SBD,

Actually, i think that the main concern of the rear gunners of your favorite plane was to scan the sky to spot enemy planes, especially fighters. I understand from the report and some others readings i had over the last few days that it is almost impossible to scan efficiently both the sky and the sea. You have to choose. So, on a SBD, i think that only the pilot was really scanning the sea. Furthermore, I think it is understandable that it is easier to spot a closing target rather than a fugitive one.

As a consequence, it is clear that some planes were most probably more efficient than others at searching. For instance, the Japanese planes or seaplanes had generally three men, one of them dedicated to observation. Hence, i guess they were more efficient than SBDs. Of course, planes with large crews, as bombers (B17 or G4M/G3M) or big seaplanes (PBY/H6K/H8K) were probably even more efficient even if gunners are not lookouts.

Speed is an interesting factor; actually low speed seems to increase the probability of detection in a significant way; but on the other hand, it reduces the search area. Most probably, there is an optimum somewhere...


Back to top
larsenjp
Beta-testing

Online

Joined: 19 Aug 2017
Posts: 566

PostPosted: Fri 8 Dec - 21:40 (2017)    Post subject: Search major improvement Reply with quote

SBD wrote:
As mentioned, the only search radius information I had found is 50 miles, 25 starboard+25 port, deduced from Admiral Belliger’s comments pre Midway.


Actually i did not really found information about visibility in the South West Pacific area.
I found some information about visibility in the Bay of Biscay (always the u-boote concern) and, statistically the minimum retained was about 5 nautical miles while the mode retained was about 15 nautical miles... To be confirmed. The mode as you may know is a statistical data: it means that 50% of actual observed visibility data were above 15, and 50% were below. This is not the average.

I guess it should be higher around the Slot or anywhere else in the SWP or South Pacific area...

However, in your example, there are two possibilities:
- either the meteorological visibility was 25 miles with a lower range of visual detection.
- or the range of visual detection was 25 miles with a higher meteorological visibility
I think it is most probably the first case since meteo services give the meteorological visibility, not the range of detection (they have no mean to measure it).


Back to top
SBD


Offline

Joined: 28 Oct 2017
Posts: 55
Localisation: UK

PostPosted: Sat 9 Dec - 22:12 (2017)    Post subject: Search major improvement Reply with quote

Thank you jp
We are wholly in agreement re the SBD gunner’s spotting.
I should have been more specific.
You mentioned the ineffectiveness of ‘rear’ spotting in the 1946 report, the SBD gunner would be a prime example of this, by many criteria.


I believe the Bellinger 50 miles was an approximate pragmatic figure for the various parties who would be reading, from Admiral King to a PBY commander. It’s the ‘have to be seen to be doing something’ syndrome? Though it’s not too far off.


The smaller IJN planes may have been more conducive to spotting, but they seem to be less thorough than the USN - reading Lundstrom. eg flying through cloud rather than dropping down below them.


I understand the optics you mention re rear spotting, but I think this is far too low level of detail for our game?
In any event the planes are spotting what was the rear on the outward flight on the return within the game, though maybe not literally.


Back to top
Contenu Sponsorisé






PostPosted: Today at 16:36 (2019)    Post subject: Search major improvement

Back to top
Display posts from previous:   
Post new topic   Reply to topic    Carrier Battles for Guadalcanal Forum Index -> Topics -> New Features All times are GMT + 1 Hour
Goto page: <  1, 2, 3, 4, 5, 6  >
Page 4 of 6

 
Jump to:  

Index | Administration Panel | Create own nforum | Free support forum | Free forums directory | Report a violation | Conditions générales d'utilisation
Powered by phpBB © 2001, 2005 phpBB Group