Dice rolls

[thewood1]

"Where in-flight refuelling is concerned, before 'Tiny' (approximately) it used to be simple and worked fine.'

A lot of things used to be simple. And some very vocal voices wanted more detail and complexity. And thats what we got. The game has a steep learning curve and there isn't single person these forums that are or can be experts on all of the detailed features.

As to Patriots, did I miss where you posted a save showing the issue?

[fitzpatv]

I just tried to attach the save file, but got an invalid file extension error. Perhaps the file is too big?. It is 4,412kb.

[Dimitris]

Compress it (to zip or rar/7z etc.) and then upload it.

[fitzpatv]

OK, I hope this helps resolve the issue.

[thewood1]

I'm going to post a separate thread in tech support on the PAC-2. There seems to be an issue with it in the final end-game. Its not missing, its just plain losing lock in the terminal phase. I have checked PAC-3 and its ARH so its not PAC radar issue. I have checked several other SARH SAMs and they work. So its specific to PAC-2.

[Tcao]

Hi fitzpatv

I run your scenarios for dozen minutes but nothing happens. So I modified your save , delete most of the units on the map but keep the Patriot bty in Kuwait, then add a group of Mig-23 to Iran side.

Yes, there is something wrong here.

cmo bug 2.11.jpg (243.18 KiB) Viewed 982 times

change the DB to the latest version(504) will partially solve the issue but will leads to a different type of strange behavior.

Look at the spoofed rate

weapon: MIM-104C Patriot PAC-2 #11440). Final probability: 15%. Result: 54 - FAILURE

2/20/1994 4:40:31 AM - Decoy (Generic Chaff Salvo [4x Cartridges]; Tech: N/A) from MiG-23MLD Flogger K is attempting to seduce sensor: AN/MPQ-53 (Tech: Late 1980s)(Guiding weapon: MIM-104C Patriot PAC-2 #11440). Final probability: 15%. Result: 2 - SUCCESS

2/20/1994 4:40:31 AM - Radar providing guidance for TVM/SAGG-guided weapon was spoofed - cannot impact



2/20/1994 4:40:31 AM - Decoy (Generic Chaff Salvo [5x Cartridges]; Tech: N/A) from MiG-23MLD Flogger K is attempting to seduce sensor: AN/MPQ-53 (Tech: Late 1980s)(Guiding weapon: MIM-104C Patriot PAC-2 #11441). Final probability: 15%. Result: 79 - FAILURE

2/20/1994 4:40:31 AM - Decoy (Generic Chaff Salvo [4x Cartridges]; Tech: N/A) from MiG-23MLD Flogger K is attempting to seduce sensor: AN/MPQ-53 (Tech: Late 1980s)(Guiding weapon: MIM-104C Patriot PAC-2 #11441). Final probability: 15%. Result: 3 - SUCCESS

2/20/1994 4:40:31 AM - Radar providing guidance for TVM/SAGG-guided weapon was spoofed - cannot impact





2/20/1994 4:41:03 AM - Decoy (Generic Chaff Salvo [5x Cartridges]; Tech: N/A) from MiG-23MLD Flogger K is attempting to seduce sensor: AN/MPQ-53 (Tech: Late 1980s)(Guiding weapon: MIM-104C Patriot PAC-2 #11442). Final probability: 15%. Result: 53 - FAILURE

2/20/1994 4:41:03 AM - Decoy (Generic Chaff Salvo [4x Cartridges]; Tech: N/A) from MiG-23MLD Flogger K is attempting to seduce sensor: AN/MPQ-53 (Tech: Late 1980s)(Guiding weapon: MIM-104C Patriot PAC-2 #11442). Final probability: 15%. Result: 67 - FAILURE

2/20/1994 4:41:03 AM - Decoy (Generic Chaff Salvo [5x Cartridges]; Tech: N/A) from MiG-23MLD Flogger K is attempting to seduce sensor: AN/MPQ-53 (Tech: Late 1980s)(Guiding weapon: MIM-104C Patriot PAC-2 #11442). Final probability: 15%. Result: 91 - FAILURE

2/20/1994 4:41:03 AM - Decoy (Generic Chaff Salvo [5x Cartridges]; Tech: N/A) from MiG-23MLD Flogger K is attempting to seduce sensor: AN/MPQ-53 (Tech: Late 1980s)(Guiding weapon: MIM-104C Patriot PAC-2 #11442). Final probability: 15%. Result: 28 - FAILURE

2/20/1994 4:41:03 AM - Decoy (Generic Chaff Salvo [5x Cartridges]; Tech: N/A) from MiG-23MLD Flogger K is attempting to seduce sensor: AN/MPQ-53 (Tech: Late 1980s)(Guiding weapon: MIM-104C Patriot PAC-2 #11442). Final probability: 15%. Result: 43 - FAILURE

2/20/1994 4:41:03 AM - Decoy (Generic Chaff Salvo [5x Cartridges]; Tech: N/A) from MiG-23MLD Flogger K is attempting to seduce sensor: AN/MPQ-53 (Tech: Late 1980s)(Guiding weapon: MIM-104C Patriot PAC-2 #11442). Final probability: 15%. Result: 66 - FAILURE

2/20/1994 4:41:03 AM - Decoy (Generic Chaff Salvo [5x Cartridges]; Tech: N/A) from MiG-23MLD Flogger K is attempting to seduce sensor: AN/MPQ-53 (Tech: Late 1980s)(Guiding weapon: MIM-104C Patriot PAC-2 #11443). Final probability: 15%. Result: 5 - SUCCESS

2/20/1994 4:41:03 AM - Radar providing guidance for TVM/SAGG-guided weapon was spoofed - cannot impact







2/20/1994 4:41:33 AM - Decoy (Generic Chaff Salvo [5x Cartridges]; Tech: N/A) from MiG-23MLD Flogger K is attempting to seduce sensor: AN/MPQ-53 (Tech: Late 1980s)(Guiding weapon: MIM-104C Patriot PAC-2 #11444). Final probability: 15%. Result: 69 - FAILURE

2/20/1994 4:41:33 AM - Decoy (Generic Chaff Salvo [5x Cartridges]; Tech: N/A) from MiG-23MLD Flogger K is attempting to seduce sensor: AN/MPQ-53 (Tech: Late 1980s)(Guiding weapon: MIM-104C Patriot PAC-2 #11444). Final probability: 15%. Result: 4 - SUCCESS

2/20/1994 4:41:33 AM - Radar providing guidance for TVM/SAGG-guided weapon was spoofed - cannot impact

I will do more testing later this evening

[thewood1]

I think the countermeasure issue is a separate issue. I added a thread in tech support. A test scenario I built with 504 has the PAC-2 issue. Just remove the countermeasures and you'll see they aren't the main issue.

Can I also recommend creating a thread in tech support and let this thread move on.

[fitzpatv]

Great work. Looks like we're getting somewhere and I accept that it would have been sooner had I provided the files, but I really didn't know how to until recently. My previous attempts revolved around trying screenshots, which is what my interface couldn't do.

Do you want me to create the thread in Tech Support or will one of you guys do it?

[fitzpatv]

Ah, I see it's already been done. OK, I'll leave this thread alone now.

[Dimitris]

OK, I hope this helps resolve the issue.

Eeeerm..... where are we even supposed to look in this save?

[fitzpatv]

Judging from the posts above, it would seem that thewood and tcao (who I assume are on your team) are on the case. The relevant bits are the Patriot and Grumble engagements.

Regarding the original issue with the random numbers, can't help feeling that rmeckman may have been close to the answer when he said that it is likely that a given result will be followed by a similar one (if I understood him correctly). If this refers to RNG, rather than real-world probability, it would explain why the RNG in CMO often produces strings of similar results.

[thewood1]

We are not on the "team". I just try to help the devs sort through the clutter.

I will also point out your approach to trying to solve issues is challenging. No saves, poor descriptions, mixing issues in threads, etc. Its why I posted in the tech support thread. To give the devs a clean look at the issue without all the BS in this thread.

[thewood1]

"If this refers to RNG, rather than real-world probability, it would explain why the RNG in CMO often produces strings of similar results"

These statements about "dice rolls" and RNG are pretty strong for zero evidence. Its another thread where someone has a feeling about something and then people just start piling on. If its an issue, bring a save, data, etc. to the tech support thread for a real look at it.

[fitzpatv]

You're right, thewood1. I could have gone about this better than I did and, now that I understand the system a bit better, I will endeavour to do so in future and take your advice on board. Thanks for your assistance with this.

On dice rolls, I don't think that this issue is going to be easy to prove either way, so I'll just shut up about it, accept that extreme sequences of results can and do happen in the game and try to do the best I can regardless.

[rmeckman]

If this refers to RNG, rather than real-world probability, it would explain why the RNG in CMO often produces strings of similar results.

"Randomness" is basically a modeling assumption used to account for factors that can't be measured easily or to build a simpler model that can be run more quickly. Coin flips, dice rolls (the two cubes with little dots on them), and missile/bomb hits are all outcomes that result from a specific sequence of forces applied to the objects in question. For example, someone could take a limited set of measurements related to a coin flip (e.g., which side starts up, where the finger is located on the coin at start of flip) and develop a model that eliminates 5% of the "randomness" and therefore correctly predicts the outcome 55% of the time. Another person with even better measurements might be able to build a model that predicts the coin toss 60% of the time. Lacking these detailed measurements, the simplest model is to assume 50% heads and 50% tails.

CMO contains a set of models for missile, projectile, and bomb hits/misses that use randomness to simplify the calculations while still capturing essential results. There isn't a real-world probability sitting out there somewhere for missiles; instead, the developers chose to use a uniform distribution as part of their model. If they wanted to, they could build a different missile model based on draws from a normal distribution (and having different streak behavior). Likewise, the Circular Error Probable for munitions originally used a bivariate normal distribution as a modeling assumption. The key question is whether a model you design fits the observations well enough to be useful for its intended purposes.

[Tcao]

This is turning into quite a debate!. A few comments on recent posts:

- Thanks to BDukes for pointing-out where the logs are stored. I've checked mine and have found records of the repeated misses by Patriots vs Floggers (v1.06.1328.13) in 2024_1_25_19.4.25 and then the lack of hit calculation stats for same (covering several replays) in 2024_2_3_15.29.4.

- On this evidence, the logs are of somewhat limited use. For instance, they shed no light on aircraft heading to tankers when they don't need to. Having said this, the reason is pretty straightforward - the refuelling threshold is too high and can't be altered by the player.

- In answer to thewood1, which version were you playing when your Patriots were hitting things?. The problem was introduced in v1.06.1328.14.

- Having said this (and in answer to blu3), there's a recent Tech Support thread on Air-to-Air Missiles which suggests that the problem with Patriots (and Grumbles) not engaging targets MIGHT have been fixed pending the next release.

- In reply to rmeckman, I am no mathematician and cannot really argue with what you say. However, are you talking about the way RNGs work or probability in general. In the latter case, I do find this rather surprising - but what do I know?

Just want to make sure the file 2024_2_3_15.29.4 log was after .14 update? That is very similar with my finding yesterday. Chaff repeatedly seduce sensor until a success pass. This is (99% sure) a bug. I will try to post a bug report in Tech forum later if I have time available .

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

Now with the log file 2024_1_25_19.4.25. I checked the log and found out you have 13 records of calculation. I don't see anything wrong there.

As blu3s mentioned, the Decoy Endgame occurs twice, each with a 15% chance of success (I would argue this is not the best way to simulate chaff and the chance of success is too high, but I can live with that). Those two events will drag down the success hit rate to ( (1- 0.15)*(1-0.15)) = 72.25%, or, it means a success spoofed rate at 27.75%. From the log, there are 3 spoofed Patriot missiles, 3/13 = 23.08%.

If we go further and look at the final success hit rate, that part is a little bit of complicate. The final hit rate is decided by many different factors, missile speed, agility rating, altitude, proficiency, if the pilot is aware of the incoming missile or not, impact angle. With the latest update, the impact angle become the main factor that could halve the final PoH (which I have some complaint about that) , so in theory you should have the final PoH at 72.25% * 0.5 = 36.13%
from the log your final success rate is 5/13 = 38.46%

[fitzpatv]

I can confirm that file 2024_2_3_15.29.4 log was after the .14 update. If logging a bug report, please ensure that it doesn't duplicate the one created by thewood1.

Thanks for this detailed and thorough analysis, tcao.

[maverick3320]

I'm having the same issue as Fitzpatv with the Patriots (and the "random" die rolls). Here's my results of my first SAM vs enemy aircraft engagement in Northern Fury 31. This isn't cherry picked. Note that I generally set all my SAMs to fire at "no escape zone" range. The first number is the hit percentage according to the game, the second number is the actual roll.

Patriot 7695 - Overshoots, never locks on
Patriot 7696 - Overshoots, never locks on
IHAWK 7731 - 15/31 MISS
Patriot 7728 - Overshoots, never locks on
Patriot 7748 - 15/11, SPOOF
IHAWK 7749 - 43/77, MISS
IHAWK 7751 - 30/97, MISS
IHAWK 7750 - 15/11, SPOOF
IHAWK 7752 - 15/5, SPOOF
Patriot 7754 - 15/3, SPOOF
IHAWK 7755 - 15/11, SPOOF
IHAWK 7758 - 43/100, MISS
IHAWK 7759 - 35/39, MISS
Patriot 7761 - 15/12, SPOOF
IHAWK 7763 - 46/19, HIT
Patriot 7764 - 15/11, SPOOF
Patriot 7771 - 15/15, SPOOF
Patriot 7767 - Overshoots, never locks on
Patriot 7768 - Overshoots, never locks on
Patriot 7770 - Overshoots, never locks on

Now, I'm not a rocket scientist, but I've taken a few levels of calculus at the university level. Someone please tell me this is completely random. I mean...come on.

looking deeper, I actually fired 23 total Patriot missiles; I have no idea why the 12 not shown above didn't show up in the logs. They all overshot the target as well. 0/23 Patriots fired hit at "no escape zone" range.

I also noticed that many missiles simply kept flying by one enemy aircraft onto the next, which is likely why there are so many spoof "events" shown in the log. Just a guess.

At the very least this has to be a bug with the Patriots. As to the misses and spoofs...? Save attached.

[blu3s]

There's a known patriot issue reported on the Tech Support forum and it's fixed for the next update.

[thewood1]

And this is why sticking a technical issue inside another thread leads to people wasting their time.