It was something to do with how the RNG was seeded.
But that is a dim memory from long ago...
Dice rolls
Moderator: MOD_Command
Re: Dice rolls
I have seen, dimly remembered, proof in some game that RNGs produce "streakiness". The averages end up correct, but the rolls themselves (in that game) could run in streaks. Some of the streaks were low...then there'd be streaks of high rolls.
It was something to do with how the RNG was seeded.
But that is a dim memory from long ago...
It was something to do with how the RNG was seeded.
But that is a dim memory from long ago...
Re: Dice rolls
Dimitris, I don't disagree with your last post. I used to be a software tester and fully appreciate that it is very difficult for complex code to be bug-free. I also understand the commercial argument you advanced. I'm not getting at the developers or yourself and am grateful for what you have done to deliver a fine game, for all its faults. Nevertheless, the game does have a number of serious problems at the moment, of which I would prioritise fixing the chaotic state of in-flight refuelling (see my recent AARs) and the bug introduced by v1.06.1328.14, which renders SAMs like the Patriot and Grumble utterly useless (ditto). I'm sorry that I can't provide screenshots or logs but my laptop's interface seems incapable of taking them from CMO - no idea why. If anyone else would care to raise the necessary bug reports and provide the logs, I'd be grateful.
As for the RNG, I'm still convinced there's fire behind this particular smokescreen and c2k may be onto something in the post above. Worth looking into.
As for the RNG, I'm still convinced there's fire behind this particular smokescreen and c2k may be onto something in the post above. Worth looking into.
Re: Dice rolls
Every time I have seen people complain about dice rolls in any game, someone else has to do the work. Rarely do I see anyone bring any proof other than a comment that the AI is cheating. Its very simple to test at least directionally. As to the RNG streaking, RNG libraries had some issues maybe 10-15 years ago. They were rare and depended on how they were reseeded. Those development libraries were corrected. And even then the impact was minimal.
Just google "RNG streaks" and you'll see hundreds of threads complaining about streaks. Most of them are just sore losers...but not all.
Just google "RNG streaks" and you'll see hundreds of threads complaining about streaks. Most of them are just sore losers...but not all.
Re: Dice rolls
Could you point me to the thread on the tech support forum that talk about this issues? I might have skipped them,fitzpatv wrote: Fri Feb 09, 2024 10:38 am and the bug introduced by v1.06.1328.14, which renders SAMs like the Patriot and Grumble utterly useless (ditto)
Thank you.
Re: Dice rolls
I'd like to the scenario with the issue. I have Patriots firing and hitting with no issues so far in my scenarios.
Re: Dice rolls
All these guys might have this stuff captured in their logs and may not know it. Guys check your command Logs folder. They're listed by date.thewood1 wrote: Fri Feb 09, 2024 2:40 pm I'd like to the scenario with the issue. I have Patriots firing and hitting with no issues so far in my scenarios.
Mine are here:
C:\Program Files (x86)\Steam\steamapps\common\Command - Modern Operations\Logs
M
Don't call it a comeback...
Re: Dice rolls
Yeah, there are a number of issues that are game-breaking. We're used to relatively fast turnover, but it's much slower now. I think given that it's best to test internally further and maybe more before releasing to the general populace. Nobody wants to wait weeks to play a game unless there is some great reason. It really sucks when it's something kind of simple that was missed.fitzpatv wrote: Fri Feb 09, 2024 10:38 am Dimitris, I don't disagree with your last post. I used to be a software tester and fully appreciate that it is very difficult for complex code to be bug-free. I also understand the commercial argument you advanced. I'm not getting at the developers or yourself and am grateful for what you have done to deliver a fine game, for all its faults. Nevertheless, the game does have a number of serious problems at the moment, of which I would prioritise fixing the chaotic state of in-flight refuelling (see my recent AARs) and the bug introduced by v1.06.1328.14, which renders SAMs like the Patriot and Grumble utterly useless (ditto). I'm sorry that I can't provide screenshots or logs but my laptop's interface seems incapable of taking them from CMO - no idea why. If anyone else would care to raise the necessary bug reports and provide the logs, I'd be grateful.
As for the RNG, I'm still convinced there's fire behind this particular smokescreen and c2k may be onto something in the post above. Worth looking into.
Here are some more example issues.
Chinese combatants are now missing datalinks to SAMS. Chinese combatants are big in the CMO world.
https://github.com/PygmalionOfCyprus/cm ... ssues/4172
Drones and certain munitions disappearing.
https://www.matrixgames.com/forums/view ... 0&t=401356
https://www.matrixgames.com/forums/view ... 0&t=401303
https://www.matrixgames.com/forums/view ... 0&t=401245
https://www.matrixgames.com/forums/view ... 0&t=401445
M
Don't call it a comeback...
Re: Dice rolls
So, you are in the middle of an air battle and have just launched two missiles at a hostile aircraft. As the first missile homes in, you get a random draw of r1 = 81 in the Message Log for a miss. Surely, the second missile will do better. Up pops up another message saying the second missile got a draw of r2 = 81 for another miss. Those cheaters! Surely the game must be rigged to get 81 twice in a row!
Both r1 and r2 are supposed to be pseudo random draws from a uniform distribution spanning the range from 0 to 100. If everything is working as designed, what should the difference d = r2 - r1 look like? If you go through the math, d should follow a triangular distribution as shown in the hastily created image below. (For simplicity, I'm showing continuous instead of discrete distributions.) Rather than being some rare outlier, d = 0 (the same value twice in a row) is actually the most common outcome right in the center of the distribution. More generally, this distribution shows that r1 and r2 should be within 5 points of one another about 10% of the time and within 10 points about 19% of the time.
This is likely another case in which people's perceptions of what "random" should look like doesn't line up with the math. To show there was a dodgy RNG, one would have to do something along the lines of demonstrating that d does not converge on this triangular distribution after a large number of draws.
Both r1 and r2 are supposed to be pseudo random draws from a uniform distribution spanning the range from 0 to 100. If everything is working as designed, what should the difference d = r2 - r1 look like? If you go through the math, d should follow a triangular distribution as shown in the hastily created image below. (For simplicity, I'm showing continuous instead of discrete distributions.) Rather than being some rare outlier, d = 0 (the same value twice in a row) is actually the most common outcome right in the center of the distribution. More generally, this distribution shows that r1 and r2 should be within 5 points of one another about 10% of the time and within 10 points about 19% of the time.
This is likely another case in which people's perceptions of what "random" should look like doesn't line up with the math. To show there was a dodgy RNG, one would have to do something along the lines of demonstrating that d does not converge on this triangular distribution after a large number of draws.
- Attachments
-
- triangular.png (9.85 KiB) Viewed 778 times
-
caelunshun
- Posts: 130
- Joined: Sun Nov 19, 2023 12:44 am
Re: Dice rolls
I agree with everything that has been said about bugs in releases. The releases of the past six months have consistently introduced new bugs—not obscure bugs but rather bugs one is almost guaranteed to notice within their first hour of using the new version. Due to the slow release model, we then have to wait a month before they are patched.
To fix this, either the release pace should be made more frequent or internal pre-release testing needs to be significantly improved.
To fix this, either the release pace should be made more frequent or internal pre-release testing needs to be significantly improved.
Re: Dice rolls
I am really sorry you feel that way. We need issues properly reported on the Tech Support forum (see guidelines) in order to be able to efficiently investigate them.fitzpatv wrote: Fri Feb 09, 2024 10:38 am Nevertheless, the game does have a number of serious problems at the moment, of which I would prioritise fixing the chaotic state of in-flight refuelling (see my recent AARs) and the bug introduced by v1.06.1328.14, which renders SAMs like the Patriot and Grumble utterly useless (ditto). I'm sorry that I can't provide screenshots or logs but my laptop's interface seems incapable of taking them from CMO - no idea why. If anyone else would care to raise the necessary bug reports and provide the logs, I'd be grateful.
The dev team is not blind to the existence of various issues (see item #5 here) and already devotes a large part of its available time & effort resources in investigating and resolving them. Balancing between standing still to fix problems and moving forward in order to keep the Command series alive, relevant and evolving is always a delicate tightrope walk.
Re: Dice rolls
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?
- 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?
- Attachments
-
- 2024-2-3_15.29.4.txt
- (293.58 KiB) Downloaded 15 times
-
- 2024-1-25_19.4.25.txt
- (76.91 KiB) Downloaded 15 times
Re: Dice rolls
Thanks for providing the log. As you point out, a huge percent of the patriots were spoofed by the decoy.
First of all, I analyze the dice roll in the Decoy events, to see if there's a particular issue on this gameplay. As you can note below, the dice roll on the Decoys Endgame have no particular patterns and it's mean value are 51.25
What is happening is that for every Weapon Endgame, the Decoy Endgame occurs twice, one for the Generic Chaff Salvo [4x Cartridges] and another for Generic Chaff Salvo [5x Cartridges], both have a 15% of chances of success so the combined probability of the two events of success are 27.25%, quite high. (IMO only one of the Decoys should be in the unit)
Regarding all the missiles that are not spoofed, most of them are run out of energy/petered out. This could means that Patriot missiles are being launched too far away.
Hope it helps.
First of all, I analyze the dice roll in the Decoy events, to see if there's a particular issue on this gameplay. As you can note below, the dice roll on the Decoys Endgame have no particular patterns and it's mean value are 51.25
What is happening is that for every Weapon Endgame, the Decoy Endgame occurs twice, one for the Generic Chaff Salvo [4x Cartridges] and another for Generic Chaff Salvo [5x Cartridges], both have a 15% of chances of success so the combined probability of the two events of success are 27.25%, quite high. (IMO only one of the Decoys should be in the unit)
Regarding all the missiles that are not spoofed, most of them are run out of energy/petered out. This could means that Patriot missiles are being launched too far away.
Hope it helps.
Regarding this, there's no such an issue with air-to-air missiles. As explained in the post, AIM-120D works fine launched at NEZ against different set of high maneuverability targets.fitzpatv wrote: Sat Feb 10, 2024 1:32 pm 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.
Re: Dice rolls
Yep agree.What is happening is that for every Weapon Endgame, the Decoy Endgame occurs twice, one for the Generic Chaff Salvo [4x Cartridges] and another for Generic Chaff Salvo [5x Cartridges], both have a 15% of chances of success so the combined probability of the two events of success are 27.25%, quite high. (IMO only one of the Decoys should be in the unit)
Just looking...did the game ever start degrading the maneuverability adjustments based on successive SAM engagements? I don't remember. The assumption was aircraft agility is getting worse at dodging the SAMS the longer its burning energy dodging. I see some decreases, but there is no clear log entry as to why.
M
Don't call it a comeback...
Re: Dice rolls
Not sure if its aircraft directly, but G-loading is tracked on pilots. I think that degrades response capabilities.
Re: Dice rolls
Just found it in the manual:
● New air combat feature: Pilot G-tolerance (a.k.a. “how to resolve the "fighters pulling hard turns indefinitely" realism flaw”).
When an aircraft is considered to be performing "combat maneuvers", an extra UI element is shown on unit status panel: https://i.imgur.com/odY2E8O.png
This represents the crew's tolerance to hard maneuvering. The longer the aircraft is continuously pulling a hard turn, the more this buffer fills up. (Getting out of a hard maneuver, e.g. reversing a turn, reduces this strain. This is one of the reasons that scissors are a popular practice in RL dogfights.)
Once the tolerance is exhausted, the crew begin suffering G-LOC and have to significantly relax the turn (this is easier to notice in Tacview, but you should also observe a noticeably larger turn radius on the top-down view) while regaining their senses and control of the aircraft. Of course the aircraft is particularly more vulnerable during this recovery period.
● New air combat feature: Pilot G-tolerance (a.k.a. “how to resolve the "fighters pulling hard turns indefinitely" realism flaw”).
When an aircraft is considered to be performing "combat maneuvers", an extra UI element is shown on unit status panel: https://i.imgur.com/odY2E8O.png
This represents the crew's tolerance to hard maneuvering. The longer the aircraft is continuously pulling a hard turn, the more this buffer fills up. (Getting out of a hard maneuver, e.g. reversing a turn, reduces this strain. This is one of the reasons that scissors are a popular practice in RL dogfights.)
Once the tolerance is exhausted, the crew begin suffering G-LOC and have to significantly relax the turn (this is easier to notice in Tacview, but you should also observe a noticeably larger turn radius on the top-down view) while regaining their senses and control of the aircraft. Of course the aircraft is particularly more vulnerable during this recovery period.
Re: Dice rolls
Thanks for the analysis, blu3s. Possible issue with too many chaff attempts per combat, then. Regarding the Patriots that petered-out, the engagement range was 20nm, maximum, which should not have been too far away unless the Patriot really is a pile of junk. Later in the game, enemy Grumbles behaved exactly the same way, getting no hit calculations (let alone hits) on my HARMs from 58 attempts.
Here's another example of the AI's trick rolling from a dogfight I've just played, when I did very well to down two fully-armed Floggers with my 2nm ranged AMRAAMs but not before some crazy misses. The point I'm trying to make is that this is not unusual - it happens all the time and, judging from other comments here, not just to me.
Here's another example of the AI's trick rolling from a dogfight I've just played, when I did very well to down two fully-armed Floggers with my 2nm ranged AMRAAMs but not before some crazy misses. The point I'm trying to make is that this is not unusual - it happens all the time and, judging from other comments here, not just to me.
- Attachments
-
- 2024-2-10_13.38.42.txt
- (460.21 KiB) Downloaded 14 times
Re: Dice rolls
I did a quick search for "refuel" in a couple of your recent after-action reports. Some of these issues appear to to be refuelling settings that can be adjusted in the Tanker Planner under Mission Settings, such as "Tanker follows receiver's flight plan", "Use tankers assigned to specific missions", and "Receivers start looking for tanker when down to __% of mission". I have used a number of them without encountering serious issues. These settings can also be accessed in the Flightplan Editor at Refuel waypoints.fitzpatv wrote: Fri Feb 09, 2024 10:38 am Nevertheless, the game does have a number of serious problems at the moment, of which I would prioritise fixing the chaotic state of in-flight refuelling (see my recent AARs)
- Attachments
-
- Tanker Planner.png (134.83 KiB) Viewed 615 times
Re: Dice rolls
Some of the issues are players not understanding the full details of ROE/Doctrine. Not all, but some. Is it possible to know everything about them? Most likely not. Thats why before declaring bugs, might want to ask the question first.
Re: Dice rolls
Perhaps so. There's a lot to know about this game and the risk (especially with the rate of change lately) is that players aren't going to understand the full range of options available to them.
Where in-flight refuelling is concerned, before 'Tiny' (approximately) it used to be simple and worked fine. Now, judging by rmeckman's comment, it seems that players are being pushed down the Missions route. I personally prefer to run things manually, as it generally gives me a greater level of control over what is going on. Perhaps that has changed and a player can now achieve greater control via understanding the finer details of the Mission Editor in its current form. In that case, I'd probably be well-advised to take a step back and practice doing things this way for a while before returning to playing serious scenarios. Risk is, though, that everything will have changed again by the time I've done so. All a bit challenging for new players without my several years of experience.
A couple more points about the Patriots:
- The many that petered out did so after they had flown past the Floggers. They simply failed to lock onto the targets and sailed by, often well within the WRA-specified 20nm engagement range. It wasn't a matter of altitude. In all cases that I observed closely, the missile and target were in the same height envelope.
- I replayed this encounter five times out of sheer disbelief (as, I believe, the log shows). In no case did the Patriot battery score a single hit. It happened again a sixth time later in the scenario. There was then the incident with the enemy Grumbles. Finally, the same Patriot battery fired at a couple of MiG-29s when I played today and the same thing happened again. Is this really within normal parameters?.
Where in-flight refuelling is concerned, before 'Tiny' (approximately) it used to be simple and worked fine. Now, judging by rmeckman's comment, it seems that players are being pushed down the Missions route. I personally prefer to run things manually, as it generally gives me a greater level of control over what is going on. Perhaps that has changed and a player can now achieve greater control via understanding the finer details of the Mission Editor in its current form. In that case, I'd probably be well-advised to take a step back and practice doing things this way for a while before returning to playing serious scenarios. Risk is, though, that everything will have changed again by the time I've done so. All a bit challenging for new players without my several years of experience.
A couple more points about the Patriots:
- The many that petered out did so after they had flown past the Floggers. They simply failed to lock onto the targets and sailed by, often well within the WRA-specified 20nm engagement range. It wasn't a matter of altitude. In all cases that I observed closely, the missile and target were in the same height envelope.
- I replayed this encounter five times out of sheer disbelief (as, I believe, the log shows). In no case did the Patriot battery score a single hit. It happened again a sixth time later in the scenario. There was then the incident with the enemy Grumbles. Finally, the same Patriot battery fired at a couple of MiG-29s when I played today and the same thing happened again. Is this really within normal parameters?.