BUG: Defender Siege Value same for all terrains

[DSYoungEsq]

Defender Siege Value same for all terrains

I've been running a series of trials to determine the likelihood of fortress reduction during successful sieges and assaults. In the process, I found the excellent post by VKV describing how sieges work here at "How do sieges work?"

In that primer, VKV indicates that when besieging a city in plains (terrain type "0"), the Defender Siege Value does not get modified for terrain; in all other terrains, the DSV is increased by 2. Some posts I read while trying to see if my issue is known and reported seem to take this as an accurate statement.

I am running plain vanilla 1.08. In my trials, and in my most recent games (I went back and reloaded examples), the DSV is not different for cities in plains. As an example, in my trials, the DSV is the same in my siege of Tchuma as it is for my siege of Krasnoyarsk. Similarly, in a recent game, my siege of both Flandern and Zeeland started with the same DSV.

Is this the way it is supposed to be? Did it change from prior versions of the game? Or is it a bug?

 

[Owen]

There is a known bug where sieges on plains are not reduced in the way they should be when the game is reloaded with the siege already in progress.

 

[unmerged(6159)]

Isn't this a display bug?

 

[Owen]

Isn't this a display bug?

I thought it was a real effect, but given that the completion of sieges also depends on random chance, it is difficult to tell without direct debugging.

 

[DSYoungEsq]

I just loaded the Grand Campaign as England (because they start in a war with France). Walked Henry V into Picardie; walked the Duke of Bedford into Guyenne. Picardie is a plain with a minimal fortress; the DSV should be 1; Guyenne is a forest with a minimal fortress, theoretical DSV of 3.

Both sieges initiated with a DSV of 3.

Obviously, plains cities are not getting being treated differently by the game in computing the DSV. I'd ask VKV about this, but s/he seems to have stopped being involved here.

 

[ForzaA]

Picardie ONLY starts at +3 if you move there right from the start, it seems, so it is more likely a bug with starting, rather than a plains bug.

 

[DSYoungEsq]

God, I love testing out things like this (NOT!)

It is pretty clear this is another aspect of the "EU II can't handle the first month of a game after load" issue. Any siege in progress, or initiated before the end of the calendar month current when the game is loaded, will not differentiate the terrain in a plains province. This bug is clearly not cosmetic; I tried 48 sieges in plains provinces initiated with this error and none of them managed the P = P + 3 result that would have occurred in 10% of the siege results the first month had the actual DSV been +1, not +3 as shown.

Interestingly, EU I didn't have this error; at least not in version 1.07c, which I reloaded into the computer to see what happened. My 1492 French troops in Roussilon faced a DSV of 1, the same troops invading a forested province in the Spanish Burgundian inheritance faced the correct DSV of 3.

The error in EU II came with the game. Version 1.00 gives the same result to the English when they invade Picardie in January 1419 as does 1.08. However, if Henry takes a siesta first, then arrives in Picardie in February, he faces only a DSV of 1. (In an interesting turn I had forgotten, besieging troops in 1.00 get of a siege cycle within a couple days of arriving, rather than waiting until a month has passed.)

This is not a trivial issue. At the beginning of the game, without any leaders with positive siege values or artillery, against minimal fortresses, you can only expect a 30% chance each month at the start of increasing the progress of the siege if DSV starts at 3. There won't be an automatic reduction in DSV for D (duration) until after the 4th month, and then not if the city has a port you haven't blockaded. In the same situation, but with a starting DSV of 1, you have a 20% chance the first month of getting a P = P + 3 result, plus a 30% chance of the P = P + 1 result. This is a 50% chance of siege progress. This decreases the expected duration of the sieges dramatically.

This needs to be fixed.

 

[AndrewT]

Thanks for the great work. This was logged under Bug 180, but I've replaced that with a new Bug 367 which explains the situation much better. Let me know if it needs modification.

 

[DSYoungEsq]

That appears to concisely summarize the situation, AndrewT. Thanks

 

[unmerged(6147)]

That appears to concisely summarize the situation,

DSYoungEsq!

Why not continue your testing?

We have the problems with some sieges just going on for ever - reported here in the bug forum once a year at least . IIRC the FAQ says that in some situations (like a big fortress in a swamp province with free access to the Sea and no artillery present) the siege will not succeed. I have always wondered if this is true. I have had my share of these neverending sieges, perhaps thrice in 3 years gaming. One I remember was infantry at CRT level 1 sieging Shanghai (swamp) and another was Brunei (don't remember details) - both coastal provinces with a port.

Since they have been so few I believe them to be bugs rather than a general feature. Perhaps you can sort this out?

 

[DSYoungEsq]

Actually, this appears pretty clear from the FAQ of VKV that it is not a bug.

To succede in a siege, you eventually have to be able to reach a modified siege progress roll of 14 or higher. The most the raw random number contributes is 10. The added score results from a combination of increased "P" (progress) and increased "D" (duration). To get increased "P", you have to attain modified rolls of at least 5. To get increased "D", you simply wait around, unless the city is supplied by a port you aren't blockading.

Thus, if the city has a port, and you don't blockade it, and the city's DSV minus your ASV isn't 5 or lower (e.g.: fortress of level 4 in a swamp province, you don't have at least 40 artillery pieces or a leader with a positive siege capability), you will NEVER increase either P or D, only attrit yourself away slowly but surely.

I'm sorry if that is going over what everyone already knows; I appear to hack some off here by stating what they think is obvious. But if you have a siege where you aren't sure why the siege isn't resulting in progress, I suggest saving the game, then opening up the save file and looking to see what the values for P and D are. All sieges are easy to find in the save file; they are found at the very end of the file, ordered by province number. Each siege will have the recorded P, D and breach values (breaches only of importance if you intend to assault the fortress).

If any one has an example of a non-progressing siege, and the save file shows that P should be increasing, because of the combination of ASV and DSV, let me know all the parameters, and I'll try to set up some tests.

Also, don't forget that the damn random number generator in this game can get stuck. If the siege isn't progressing, try a save and reload after quitting the game and restarting it; then see if it proceedes. You may be generating the same random roll over and over...

 

[unmerged(6147)]

Actually, this appears pretty clear from the FAQ of VKV that it is not a bug.

To succede in a siege, you eventually have to be able to reach a modified siege progress roll of 14 or higher. The most the raw random number contributes is 10. The added score results from a combination of increased "P" (progress) and increased "D" (duration). To get increased "P", you have to attain modified rolls of at least 5. To get increased "D", you simply wait around, unless the city is supplied by a port you aren't blockading.

Thus, if the city has a port, and you don't blockade it, and the city's DSV minus your ASV isn't 5 or lower (e.g.: fortress of level 4 in a swamp province, you don't have at least 40 artillery pieces or a leader with a positive siege capability), you will NEVER increase either P or D, only attrit yourself away slowly but surely.

There are two problems with this

1. As I tried to say this is not always happening. In some very rare cases only. AFAIK it has only been three times in 3 years of intense playing for me I believe it is. BTW, in my Shanghai case it was quite early, say 1480, and it was at most a level 2 fort (small).

2. "only attrit yourself away slowly but surely..."

This is not correct. The besieger (in contrast to the besieged) will never suffer attrition as long as he respects the actual support limit for the province. Every player knows this by heart after having played a few GCs, so do you DSyoungEsq. This bug in VKV's description has always made me doubt his whole exposition. I somehow believe that he got information of some kind of elaborate scheme that was never implemented. But perhaps some of it was, I don't know.

As a matter of fact, I remember when he VKV first wrote about it, it was before it became public. "This is how it is going to be", he wrote (perhaps these words still can be found in the FAQ).

EDIT Just checked it. Actually it was more limited in scope than I remembered. He was discussing this "extra attrition" of both parts. He says that the "extra attrition" will not be implemented until release 1.6. We now know it was only implemented for the besieged. ENDEDIT

 

[unmerged(6147)]

I thought a little more about this “attrition” thing.

Perhaps I misinterpreted what VKV said.

This is what he says

“What is extra attrition? Each time the extra addition is increased, an amount of the side troops die. For instance, if the defender extra attrition was 7%, and is increased by 1%, then 8% of the defender troops die. It is applied immediately and not at the beginning of the month. Important note: Extra attrition will only be applied correctly in the next 1.06 patch.”

If we instead say: “whether attrition or not will take place for the besieging part depends on normal rules for the actual supply limit in the province and the size of the armies present in the province. If the size of the armies exceeds the actual supply limit for the besieger, then he will suffer attrition. But when we calculate the exact amount of attrition he will suffer, then an "extra" amount of attrition will be added according to what has been said before.”

This makes more sense but it means that the second sentence in the quotation above, i.e. “Each time the extra addition is increased, an amount of the side troops die.” is wrong.

But this is perhaps beside the point, which is the neverending siege bug, that perhaps is not a bug at all.