Retreat Bug?

[WanderingHead]

ORIGINAL: Tom Grosv
Ah ha! Now I know that supplies are important I did another experimental attack and this time packed Moscow with supplies - it worked! My retreating Kursk units went to Moscow! Well, it's better than dismantling the Gorki factory.

In that sense, the original rules could be better. It is clearer what to do to force the situation (just stack supplies). But that still left lots of bad decisions I found (in part perhaps because few people bothered to intentionally stack supplies).

ORIGINAL: Tom Grosv

ORIGINAL: WanderingHead
The start of game (Global Glory) values are Moscow: 41, Gorki: 46

Could I humbly ask you for a bit more detail - I've tried following your algorithm but I can't get my sums to add up. Should have studied maths harder at school.

To be fair, I didn't count these myself. I just slightly modified the program to have it easily print it out.

Also, I made a mistake in my description above . Above, I included +5 points just for being a friendly region, which is not actually implemented. It is actually

Code: Select all

foreach candidate_region (neighboring lost region)
 {
   metric(candidate_region) = supplies  + 5* (factories*production_multiplier + population + rail_units)
   foreach neigb_region (same player region neigboring candidate_region)
   {
     metric(candidate_region) += (supplies  + 5* ( factories*production_multiplier + population + rail_units))/2
   }
 }
 retreat to region with largest metric(candidate_region)

The numbers:

Moscow: 5 (1 rail) + 10 (2 pop) + 5 (1 factory) = 20
Kursk: 5 (1 rail) + 5 (1 population) = 10
Smolensk: 5 (1 rail) + 5 (1 pop) + 5 (1 factory) = 15
Yaroslavl: 5 ( 1 rail) = 5
Gorki: 5 (1 rail) + 5 (1 pop) + 5 (1 factory) = 15

Moscow total: 20 + 10/2 + 15/2 + 5/2 + 15/2 = 20+5+7+2+7 = 41

(all the fractions are rounded down)

I modified my code to show it like in the image. Such visibility could be included, seems fair to me. The only risk is that it is a little misleading ... if some neighboring regions are lost then the retreat value changes.

[Tom Grosv]

Thanks WH, that all makes sense now.

[Tom Grosv]

ORIGINAL: Tom Grosv

So, if I was able to go back in time and redo my move, what should I have done to ensure my units retreated to Moscow? I haven't got a clue - can anyone help me?

I've done some further testing using the info from WH and I now I think I understand how retreat works. Going back to my original game I realise that Moscow had 84 points and Gorki 90 points (the x3 factory multiplier is kicking in). All I would have needed to do is change the balance of supplies (remembering every supply in Moscow is worth 1/2 to Gorki). Perhaps not strictly ideal so maybe the tweaks suggested by WH and Lucky would be a good thing but the most important thing to me is I understand the process. I'm fairly sure the set up with the current algorithm suits the vast majority of retreats - players just may want to take a little care in a few critical locations.

[WanderingHead]

What I'm thinking is to change the formula as follows

Code: Select all

foreach candidate_region (neighboring lost region)
 {
   metric(candidate_region) = supplies  + factories*production_multiplier + population + rail_units
 
   metric_neighbors = 0
   foreach neigb_region (same player region neigboring candidate_region)
   {
     metric_neighbors += factories*production_multiplier + population + rail_units
   }
 
   metric(candidate_region) += metric_neighbors/3
 }
 retreat to region with largest metric(candidate_region)

The changes involved in the equations:
1) eliminate the 5* for all the non-supply terms, to make supply relatively more important. This allows the player to more easily influence it directly if desired.
2) remove the impact from supplies in neighboring regions (again making it easier to influence with supply placement).
3) decrease the effect of neighbors to 1/3 from 1/2.

In addition:
4) show the retreat value in the region popup for friendly regions, so the player can see it.
5) maybe include some randomization (?)
6) maybe include strategic points (?)

As I said, the retreat value in the region popup can be a little misleading because it can change from supply production and lost neighboring regions, but with that information and this rule spelled out in the release notes we could have the best of all worlds. Most retreats should make sense by default, the player can see it and reasonably predict it, and the player can influence it with supply positioning.

[Lucky1]

Thanks WanderingHead. Your assistance (as always) has been quite instructive and I think the proposed changes will be an improvement.
 
Since you are explaining the mysteries of retreats, I wonder whether it might be possible to explain how damaged items trace to factories. I have always been puzzled about this and wonder (for example) why my damaged Chinese fighter (and often, other units) queues up in Sinkiang....
 
Cheers,
 
Sean

[GKar]

ORIGINAL: WanderingHead

The changes involved in the equations:
1) eliminate the 5* for all the non-supply terms, to make supply relatively more important. This allows the player to more easily influence it directly if desired.
2) remove the impact from supplies in neighboring regions (again making it easier to influence with supply placement).
3) decrease the effect of neighbors to 1/3 from 1/2.

In addition:
4) show the retreat value in the region popup for friendly regions, so the player can see it.
5) maybe include some randomization (?)
6) maybe include strategic points (?)

I especially like no. 1). [:)]

I'd vote for no. 6) but would only introduce a small random factor, if at all.