Intriguing idea: Hiding the rolls

[Trax]

Works for me.
Could the program determine every possible corps that could reinforce a battle and pre assign the die rolls for each?

[Marshall Ellis]

I am bundling a 100 d6 rolls for reinf for each combatant. I believe this should cover the possible number of corps without a leader that could roll for reinforcement.

[Jimmer]

ORIGINAL: Trax

Along with the die rolls the defender should not see the attackers chit pick before returning the battle file.

Agreed.

In fact, the attacker shouldn't even be told whether the defender has already chosen one or not. The reason is that if it is already chosen, it's likely the computer chose it, which changes the odds significantly.

[Jimmer]

ORIGINAL: Marshall Ellis

I am bundling a 100 d6 rolls for reinf for each combatant. I believe this should cover the possible number of corps without a leader that could roll for reinforcement.

With the rolls hidden, this becomes a lot more viable. However, you would want to always hide the current round's rolls in this case. In other words, allow the person to see only the rolls that occurred in the previous round.

The reason for this is that if the person knows whether a particular army reinforces or not, he could change the order of checking to get a better result. Example:

France attacking Prussia. Both sides get average first round results. Soult is running the combat for the French with 3 corps. von Blucher is running the Prussian side with 5 corps. Napoleon with 3 corps and Davout with 1 corps are able to reinforce in.

If the player can re-load the turn, ONLY for the purposes of finding out who gets a good reinforcement roll, that's valuable. If the rolls are pulled off a stack of rolls pre-generated, he can use this knowledge to choose which one should try first (if it matters). In the example, let's say Nappy rolls a 6 and Davout rolls a 2. Well, knowing this, he could go back and try again, this time having Davout attempt to reinforce first.

If the results are hidden, then he wouldn't be able to discern this knowledge.

By the way, holding a variable for "how many reloads have been done" won't help. If I were inclined to cheat, I would have two copies of the game on two different systems. I would load up the game on my test system first, and see where I got the best results. Then, I would throw them away and go back to the primary system and use the knowledge I gained elsewhere to better my results.

Thus, reinforcement results need to be kept hidden as well. Of course, with the model in place now, this is impossible. The current model has the leader roll, and if successful, the reinforcing player decides which corps join the battle. It should be "all or nothing". If Napoleon succeeds, then every corps joins the battle regardless of whether Nappy wants to leave some behind. So, you would have to get rid of the two-step reinforcement.

However, if that's too hard, you can do it this way:

Generate a table of all stacks that can reinforce. For stacks with no leader, separate out the corps in the table. Sort the table by area (and then by corps number in area, if no leader present). Finally, have a check box next to each corps. The player would check the box to indicate it will reinforce IF it succeeds on the roll.

This is different that the original method, but security demands in the computer version are quite different from having humans look over the die as it is rolled.

[Jimmer]

Another alternative to bundling "100 rolls" at the start is to bundle 100 individual rolls for every possible die roll that might be made during that one day of combat. This is one of the reasons I asked for a pseudocode. Here is a partial list (possibly full, but I am not sure):

- Chit choice for defender when picked by the computer
- Artillery roll(s)
- Guard commitment roll(s) (by round and player)
- Combat roll
- Reinforcement roll(s) for every stack with a leader and for every loose corps counter which has a legal reinforcement possibility. (NOTE: You could even roll for the illegal ones, if it is too hard to check that at this point in the code. You also can make the rolls even for things that may not occur.)
- Pursuit roll (the roll stays the same even if a cav leader reinforces into the combat -- it's only the modifier that changes).
- Outflank roll

All of the above should be pre-generated, and should be rolled for all three rounds of combat. This is true even if they don't occur or if that particular roll will not be used. Note that this latter statement means you can pre-roll for ANY combat, no matter what the force composition is, always using the same code to make the rolls (for land; naval would be different). The only exceptions would be reinforcing rolls, which is dependent upon the circumstances. However, you could even go so far as to roll to reinforce for every counter on the map (if need by technical strictures), just to make sure you get the ones that count (that would depend upon whether the roll gets attached to the counter or not).

[Marshall Ellis]

Baby steps ... :-0!

I have coded the following today for inclusion into 1.07.00:

All die rolls, casualties, morale losses, and the factors to eliminate are hidden until the attacker is choosing his causalties from the first round! I am writing the pre-rolled dice function as we speak.

SO, this is how it will work:

Attacker pick chit
<send to defender>
defender pick chit
<send to attacker>
Attacker gets file and sees 1st round results.
choose casualties
<send to defender>
Defender sees 1st round results
choose casualties
<send to attacker>
repeat until end.

I have also found several battle file cleaning issues that I am working on!

Question: Is there any reason that the pre-rolled dice could not be implemented across the board (Non PBEM games as well)?

Once the 1.06 general release is out then I will post a 1.07 BETA quickly (After my testers have played with it a bit) so that you guys can try this...

Thoughts?

[Marshall Ellis]

Actually, I think the prerolled dice for non-pbem is a non issue since anyone could backtrack to the land phase and restart BUT this would make it easier for me to only have to support one code set.

This is scary guys ... I am asking and answering my own questions in the forums! It's like I'm talking to myself and answering :-0!

[Dancing Bear]

ORIGINAL: Marshall Ellis

Baby steps ... :-0!

I have coded the following today for inclusion into 1.07.00:

All die rolls, casualties, morale losses, and the factors to eliminate are hidden until the attacker is choosing his causalties from the first round! I am writing the pre-rolled dice function as we speak.

SO, this is how it will work:

Attacker pick chit
<send to defender>
defender pick chit
<send to attacker>
Attacker gets file and sees 1st round results.
choose casualties
<send to defender>
Defender sees 1st round results
choose casualties
<send to attacker>
repeat until end.

I have also found several battle file cleaning issues that I am working on!

Question: Is there any reason that the pre-rolled dice could not be implemented across the board (Non PBEM games as well)?

Once the 1.06 general release is out then I will post a 1.07 BETA quickly (After my testers have played with it a bit) so that you guys can try this...

Thoughts?

This works for me.

[bresh]

Sounds good Marshall.

Wish it would come sooner !
Im at point where i would rather wait for this, than start any new games.

Btw, while looking at your battlecode, happened to figure out why leader-casulties always took "commanding leader" ?

Regards
Bresh

[Marshall Ellis]

Bresh:

I will take a look at the leader casualty selection as well!
Please start some more games Bresh OR get some different friends

[Jimmer]

ORIGINAL: Marshall Ellis

Baby steps ... :-0!

I have coded the following today for inclusion into 1.07.00:

All die rolls, casualties, morale losses, and the factors to eliminate are hidden until the attacker is choosing his causalties from the first round! I am writing the pre-rolled dice function as we speak.

SO, this is how it will work:

Attacker pick chit
<send to defender>
defender pick chit
<send to attacker>
Attacker gets file and sees 1st round results.
choose casualties
<send to defender>
Defender sees 1st round results
choose casualties
<send to attacker>
repeat until end.

I have also found several battle file cleaning issues that I am working on!

Question: Is there any reason that the pre-rolled dice could not be implemented across the board (Non PBEM games as well)?

Once the 1.06 general release is out then I will post a 1.07 BETA quickly (After my testers have played with it a bit) so that you guys can try this...

Thoughts?

Looks good to me!

[Mardonius]

Looks great in concept.

[Marshall Ellis]

UPDATE!

I have coded the pre-rolls and run them through some PBEM battle tests and it looks OK so far!
I have individually rolled attacks, artillery, pursuit, withdraws, flank attempts, guard committment and reinforcement. This isolates everything so that no roll changes based on the order something is done. It looks like this was a good idea!

[Dancing Bear]

Excellent! Any progress with the hiding the chit choices?

[Marshall Ellis]

They are hidden now (BIG EXHALE). That code was a little tougher than I thought LOL!
They are hidden until the attacker if first taking casualties. The simple way to think of this is that nothing is revealed until the attacker is removing casualties from the very first die roll.
&nbsp;
&nbsp;

[Mardonius]

This is good news. We have been using an outside server to roll big die rolls (outflank, reinforce, withdraw) but this can get very cumbersome when dealing with single-round multivariate situations. Hiding the rolls and chit choices while allowing for needed intelligence for things like guard commitment decisions (btw how do you incorporate this?) and reinforcement decisions seems to be the best answer.

[Jimmer]

ORIGINAL: Mardonius

This is good news. We have been using an outside server to roll big die rolls (outflank, reinforce, withdraw) but this can get very cumbersome when dealing with single-round multivariate situations. Hiding the rolls and chit choices while allowing for needed intelligence for things like guard commitment decisions (btw how do you incorporate this?) and reinforcement decisions seems to be the best answer.

Check back to some of the previous posts, but ...

Guard, artillery, and reinforcment all occur after one knows the PREVIOUS round's die rolls. What is being hidden is knowing the results while you are still looking at the combat screen for the same round in which they are rolled.

In other words, you see the results of round 1 at the beginning of your round 2, and round 2's results at the beginning of 3. (Although, Marshall may have only implemented them for the first round -- we deemed that the most critical.)

One other thing he was going to do was to pre-roll all of the rolls when the battle is generated. What's being hidden are the results of those rolls (until the appropriate time). Not sure if he did this as well or not.

[Mardonius]

Thanks Jimmer

[Marshall Ellis]

All rolls are pre-rolled individually so you reinf, commit the guards, artillery, etc. are all separately rolled so there should be no difference in any type of ordering!
&nbsp;
Make sense?
&nbsp;
&nbsp;

[Mardonius]

Sounds great!