LARGE number of RADs not deploying

[Beethoven1]

The problem is in server game vs Bobo, on Axis turn 2. I am also attaching a similar PBEM save (separate game) for comparison.



First of all, I am aware of the fact that RADs do not deploy within 2 hexes of an FBD, and in NONE of these cases is that the explanation (I was careful where to put my FBDs). Other possible things such as the 3 counter per hex stacking limit or enemy units also do not explain the non-deployment.

I have a HUGE number of RADs that did not deploy (6 undeployed), and all should have been able to deploy somewhere). I am aware that on occasion maybe 1 or 2 RADs might not randomly deploy for unknown reasons, but this is way too big of a number, especially for turn 2 when it makes a big difference for how quickly Axis can get the Baltic rail functional.

This HQ (I corps) had 4 RADs attached to it.

radbug.PNG (287.54 KiB) Viewed 1287 times

It could have fixed all of the 4 circled hexes (and/or perhaps some others not shown in the screenshot), but it did not fix even a single one.

IN ADDITION to that, you can see that "K83 R.A.D. Labor Detachment" was assigned to another HQ and did not deploy, despite having multiple hexes it could have fixed.

And ALSO in addition to that, in a separate HQ, "XIV R.A.D. Labor Group" did not deploy. In particular, it could have fixed the hex (175, 149), or possibly some other ones which I am not mentioning, but that is one particular example of where it could have repaired. This HQ did have some other RADs which deployed, but not that one. The only ones that deployed at all were those ones, and that HQ did not even deploy all of them (!!!).

You can also see that clearly the problem could not be explained by something else like supply or weather. My depots got plenty of freight and sent out large amounts to units without a problem. I Corps for example has 112% supply.

The RADs themselves also got good supply without an issue (if that makes any difference). Just as one example, that "XIV R.A.D. Labor Group" which I mentioned has 129% supplies, 133% ammo, 122% vehicles, and 171% support, 0 admin failures, and 100% undamaged TOE.



By contrast, attached is a save of Axis turn 2 from another game (this is PBEM, not server) which I started. I renamed it to .zip so that it would be allowed to upload, but it is a .psv file, so you just have to change the filename extension back to open it.

The Axis turn 1 opening is close to identical, and the RAD deployment scheme is also reasonably close to identical. In that case, 100% of the RADs deployed as they are supposed to. However, this other game was started pre-OOB 2.0 though on the old .02.52 patch.

Also, in this other game, on subsequent turns (through turn 4), RADs have deployed properly without any issue.

[Beethoven1]

So since this is a server game, I unfortunately can't really go forward until you can look at this and hopefully find a solution, because I don't want the previous save(s) to be overwritten and don't want the game to get screwed up by this.

1) Is there any possibility to go back? If you re-run the logistics phase from the last previous save, do the RADs deploy any differently?

2) Most importantly, is this something that is going to continue being messed up for future turns? If so, it seems pretty game breaking and I do not really see how we can continue if so many RADs are going to just not deploy.

In particular it is in sharp contrast to my other game (PBEM) with a nearly identical opener started on the old patch, and in that case 0 issue on any of the first 4 turns.

I also tested multiple times on the previous patch when I was originally planning my turn 1, and didn't have an issue there (other than confirming that RADs do not deploy within 2 hexes of a FBD).

[Joel Billings]

I can't open the zip file you've uploaded. It says it is not a valid archive. The best way to figure out if something has changed in a later version than 1.02.52 is to send me a save from the end of the Soviet turn (so I can load and run the German logistics phase - if you don't have that, from the end of your prior turn is ok as I will just skip through the Soviet turn). I can run that using 1.02.52 and the latest exe, and see how the RAD behavior differs. If there is a difference that looks more than just due to normal random factors, I can try to have Gary look at this. I can tell you it likely won't be an easy thing for him to figure out unless there's a clear difference in behavior between different versions. I am almost certain that Gary is using the names of the units to determine which units can do rail repairs, not the OB numbers. So the OB change should have no impact. I looked at AI test games from 1.02.52 and the latest version, and through turn 4 the rail repair units going to the map look very similar in the 2 games. So at least in the AI games, there doesn't appear to be anything new that is blocking the rail repair units in the most recent versions. You can send the save directly to 2by3@2by3games.com.

BTW, only 1 rail repair unit is going to the map on the German turn 2 in these test games.

The server system seems to only save one save per player turn, so if you've gone past your turn to the Soviet turn, all you would have is the start of the German turn and then the start of the Soviet turn. We can roll back through saves, but you'll lose entire player-turns.

BTW, the RADs were never intended to repair mainline rail hexes. They were meant to fill in the gaps in the rear areas and complete some of the cross lines over time. The fact that players have managed to mini-max their use is not surprising, but if we were doing this again, I'd argue we should have code limiting their use to no closer than 10-15 hexes of the frontline. Adding their repair ability to the FBD's provides more repair capacity than we intended. That said, I understand the frustration of thinking they'll do something and then not seeing it happen.

[Nikel]

As Beethoven1 said, it is a .psav file, that was changed to .zip

I can change the format to .psav and load it in the game, requesting a password.

[M60A3TTS]

I tried to replicate this and could not. Starting with a game I also had under the .52 patch there were no issues with deployment. I then fired up a fresh game under 1.03.09 and created the same situation and once again got deployments. I even deliberately left hex A un-repaired in the event the routine had changed to where repairs could only happen where there was a connection to a functional rail network. But nothing suggests any change in the repair routine, the 2 RADs still deployed.

[Beethoven1]

I can't open the zip file you've uploaded. It says it is not a valid archive.

See this:

By contrast, attached is a save of Axis turn 2 from another game (this is PBEM, not server) which I started. I renamed it to .zip so that it would be allowed to upload, but it is a .psv file, so you just have to change the filename extension back to open it.

The best way to figure out if something has changed in a later version than 1.02.52 is to send me a save from the end of the Soviet turn (so I can load and run the German logistics phase - if you don't have that, from the end of your prior turn is ok as I will just skip through the Soviet turn). I can run that using 1.02.52 and the latest exe, and see how the RAD behavior differs.

There are 2 files in the new zip attached here.

I think the first one may be end of Soviet turn 1, but since I don't have the Soviet password I can't verify. If that does not work, try the other Soviet turn 1 one, which should be start of Soviet turn 1.

If there is a difference that looks more than just due to normal random factors, I can try to have Gary look at this. I can tell you it likely won't be an easy thing for him to figure out unless there's a clear difference in behavior between different versions. I am almost certain that Gary is using the names of the units to determine which units can do rail repairs, not the OB numbers. So the OB change should have no impact. I looked at AI test games from 1.02.52 and the latest version, and through turn 4 the rail repair units going to the map look very similar in the 2 games. So at least in the AI games, there doesn't appear to be anything new that is blocking the rail repair units in the most recent versions. You can send the save directly to 2by3@2by3games.com.

BTW, only 1 rail repair unit is going to the map on the German turn 2 in these test games.

BTW, the RADs were never intended to repair mainline rail hexes. They were meant to fill in the gaps in the rear areas and complete some of the cross lines over time. The fact that players have managed to mini-max their use is not surprising, but if we were doing this again, I'd argue we should have code limiting their use to no closer than 10-15 hexes of the frontline. Adding their repair ability to the FBD's provides more repair capacity than we intended. That said, I understand the frustration of thinking they'll do something and then not seeing it happen.

OK, understood, I appreciate it if you and Gary take a look. As far as repairing the front line or something like that though, in this case I was not repairing any main lines on the front line, just 100% repairing gaps in the Baltic rail, basically all the gaps I can fill between East Prussia and Daugavpils. So I am not trying to do anything unusual, I don't think.

The server system seems to only save one save per player turn, so if you've gone past your turn to the Soviet turn, all you would have is the start of the German turn and then the start of the Soviet turn. We can roll back through saves, but you'll lose entire player-turns.

To clarify, is the one single save per player turn the most RECENT save, or is it the save at the start of the turn? My understanding was it was the most recent save. If so, there should be a pretty recent save from near the end of Soviet turn 1 there.

I have deliberately not saved anything and won't save anything or go on with the game in order to make sure this doesn't get screwed up.

[Joel Billings]

I'll take a look at your saves when I have a chance. When you end your turn, and a save is created, I think that save is basically the start of the Soviet turn. If you had saved it in the middle of your turn, I would have expected that it would be the last save listed for you, but I'm not sure it keeps that one once you've ended your turn. So in this case only the start of your turn save is still on the server. I say that because it seems I always find start of turn saves on the server. The only time I find something in the middle of a turn is when a player specifically has saved mid turn, alerted me to get the save, and didn't continue on until after I got the save.

[Joel Billings]

I spent over an hour on your saves, trying to see if I could see anything. I did confirm that the RAD detachment (smaller units) can work under 1.03.10. I think it is possible that the fact that some of the hexes in the north are not connected back to rail in your server game might be making a difference, but I can't say for certain. In your server game there were a lot of the hexes near Daugvapils that had no rail path back that didn't hit enemy hexes (not true in the PBEM game). IIRC, when Gary looked at these rail units in the past, it seemed that he couldn't rule out that there is a possibility that a HQ will attempt to send out to a hex that it in the end decides is invalid, and that this might prevent the unit from going to a valid hex. I did notice that in the server game you had your RAD units in 3 HQs, instead of 5, and the 2 HQs up north didn't send anyone out while the 1 further south sent 4 of it's 5 units out. It's possible that something caused the 2 HQs to stop using units, but I don't know. It's possible that some bad logic in the routine is causing units to fail, but since I don't have a save that I can repeat for this, I don't see how Gary can easily debug it. Since some rail repair units are going out, it's not a global 1.02.52 vs 1.03.xx version problem. The only way I could hope to get at this would be to have a save at the end of the Soviet turn that when run, has the units not going out. For a server game, that would mean the Soviet player would have to save at the end of his turn, but not end his turn. They I'd have to get the save. At that point the Soviet player could end his turn and the German player could see what units went out. I could also run it, and if both the player and my run showed units not going out, Gary could use the save to try to track things. Proving a negative is the hardest thing to debug (i.e. debugging why a unit goes out is easy, finding out why it doesn't is much harder).

Recently, we started setting up two new people with access to the WitE2 source code. I will ask them if they are willing to take a look and see if they can identify what factors might be impacting the RAD units going out or not. I can't say whether they'll be interested or able to figure out the code for this. It's not worth Gary looking at this without a save that repeats, as the odds that he'll see something that is not WAD is almost nil. I realize this doesn't really help you in your game, but that's the situation. If you play your 2nd turn, if you can get your Soviet opponent to contact me to grab the save before he ends his 2nd turn, let me know. If that turn has similar problems that we can repeat, then perhaps we can figure out what's going on.

[821Bobo]

If you play your 2nd turn, if you can get your Soviet opponent to contact me to grab the save before he ends his 2nd turn, let me know. If that turn has similar problems that we can repeat, then perhaps we can figure out what's going on.

Joel no problem, I can save before ending my T2 and ping you if that helps.

[Joel Billings]

Yes, please ping me after you've saved at the end of your turn, but before you actually end your turn. Then wait to hear from me before going back in to end your turn. Thanks.

[Beethoven1]

I spent over an hour on your saves, trying to see if I could see anything. I did confirm that the RAD detachment (smaller units) can work under 1.03.10. I think it is possible that the fact that some of the hexes in the north are not connected back to rail in your server game might be making a difference, but I can't say for certain. In your server game there were a lot of the hexes near Daugvapils that had no rail path back that didn't hit enemy hexes (not true in the PBEM game).

Thanks for the effort. I am not sure what you mean in this part with regards to "some of the hexes in the north are not connected back to rail in your server game might be making a difference." Here are the hexes I had flipped in my PBEM game at the end of Axis turn 1:

pbemnorthhexes.PNG (1.23 MiB) Viewed 1129 times

And by comparison, these are the ones I had flipped in the server game at the end of Axis turn 1 (I fortunately took a screenshot without units):

servernorthhexes.PNG (1.65 MiB) Viewed 1129 times

These appear to be virtually identical from what I can see, and the RADs deployed fine in the PBEM game (in basically the same areas and very similar hexes, we are talking the same hexes or hexes 1 or 2 hexes away at most, on the same rail lines).



But you should be able to replicate the fact that they do deploy in the PBEM from the saves I sent you (assuming I sent you end of Axis turn 1, which I think I did?).

To make things a bit easier, here also is a save of a single player game where the RADs deploy in that same sort of area. Since this is single player, anyone can open it without a PBEM password if anyone else is interested. In the single player game, there is an even larger area of un-converted enemy territory:

singleplayerhexes.PNG (2.41 MiB) Viewed 1129 times

-- note -- the upload tool stopped working, so I will post the single player save in a post right after this

But if you just open the single player save, and click through to end the Axis turn and then end the Soviet turn, you should see that on Axis turn 2 the RAD deploys somewhere around Daugavpils. This is what I got in the single player test. 1 RAD that I assigned to XXXIX Mot Corps deployed properly (I only bothered to assign a single RAD to a corps HQ for the test):

singleplayerRAD.PNG (2.1 MiB) Viewed 1129 times

And also, let's not forget that regardless of this north area around Daugavpils, in the server game one of my RADs which was in the HQ right around East Prussia did not deploy, despite there being no enemy controlled hexes between it and East Prussia at all (and it also repairing a hex literally right next to it with one of the other RADs that did deploy).

[Beethoven1]

Here is the single player save described above.

[Joel Billings]

The difference is in the server game, you do not have a contiguous rail line (even counting those that are damaged) in your control that goes from Riga (or pre-invasion German territory) to Daugvapils, while in the other games you do. There are hexes in the server game that are Soviet controlled that block the rail lines, unlike in the other saves you've sent. Is this making a difference? I don't know, I'm just speculating that it might. I was able to take your PBEM save, and block the northern rail hexes off from being able to trace back to Riga or Germany (I turned a few hexes to Soviet control). When I did that and I ran the game, the HQs in the north did not send out any units, while when I didn't do that, they did send them out. So that's my current hypothesis on why they didn't go out. Now there were some hexes in the server game that those HQs could go to, but many more than they couldn't go to (if my hunch is correct), so that may have been enough to confuse the system to keep them from going out anywhere.

[Beethoven1]

The difference is in the server game, you do not have a contiguous rail line (even counting those that are damaged) in your control that goes from Riga (or pre-invasion German territory) to Daugvapils, while in the other games you do. There are hexes in the server game that are Soviet controlled that block the rail lines, unlike in the other saves you've sent. Is this making a difference?

From what I can tell, the area you are referring is this area of 11 hexes (in particular the blue hex), is that correct?:

daug.PNG (162.63 KiB) Viewed 1069 times

I did not flip those hexes, so the highlighted rail line between Daugavpils and Riga goes through one hex of enemy territory (but with 0 units inside it) at the blue hex, that is true.

But look again closely at the previous 2 screenshots I posted in my previous post.

End of Axis turn 1, PBEM game:



End of Axis turn 1, server game:



So you can see that I did not flip that same 11 hex area, including the rail hex I colored blue in my smaller screenshot, in BOTH games.

The only difference in terms of hexes I flipped on turn 1 around that rail line is around Riga itself, because since the NKVD in Riga retreated rather than routed in the server game, I had to regiment Totenkopf rather than keep it as a full division (which would have flipped the hexes NE and SW of Riga.

There are also some minor differences in other areas of the map, but nothing that seems to be related to any rail lines that I can see.

Thank you again for your responsiveness and looking into it.



ALSO, perhaps more importantly for immediate purposes - can you confirm that you have now downloaded/saved ALL of whatever saves are currently on the server?

If so, then I will try to continue on with the game, and we can see if hopefully the issue resolves itself. But just want to be 100% sure that you have any saves that we could (possibly) go back to later, if it turns out to be some sort of persistent and larger scale issue which turns out to be significantly messing up the game (i.e. more than just 1 or 2 RADs randomly not deploying on-and-off on rare occasions or something).

[Joel Billings]

The PBEM game saves that you uploaded show a path of Axis hexes all the way from Riga to Daugvapils, with no Soviet controlled break (unlike your screenshot in your last post). Which save shows the PBEM game without a linked line from Riga to Daugvapils. In the PBEM game save I have, you do control the blue circled hex, which gives you the linked line, so all those rail hexes north of the river have a connection back to Riga.

I have the 2 saves that are on the server, which is the start of Soviet turn 1 and start of Axis turn 2. The problem is the server doesn't keep many saves, and generally just keeps the two most recent end of turn saves. Once you move on, there won't be any way to go back to these saves, as they will drop out. But I don't see a way around this at this point, since we can't find any potential issue without have a save we can use to duplicate the problem. If I get the next Soviet end of turn save (Soviet turn 2), and I run it forward and see a problem, then you could hold up until it was looked at, but you won't be able to go back to the start of Soviet turn 1.

What might be best would be for you to complete your turn 2, then save off and send it to me. I can then run it through a Soviet turn and to the turn 3 German logistics phase to see if the RADs are working any better. In this case there is a chance the Server may still have the start of the Soviet turn 1 save in case you want to go all the way back there if a change is made. I think the chance of a change being made is very low given what I know about that code, but I won't say they are zero if we have save we can look at.