Skilled Swordsmen

[jlopez]

In order to reduce the effectiveness of Roman legionaries vs barbarian foot without having to remove the skilled swordsmen ability alltogether, how about this for a solution:

Skilled Swordsmen no longer gives a POA but instead gives a -1 to enemy infantry's death roll making winning combats harder but inflicting casualties more likely.

[martindneiluk]

In order to reduce the effectiveness of Roman legionaries vs barbarian foot without having to remove the skilled swordsmen ability alltogether, how about this for a solution:

Skilled Swordsmen no longer gives a POA but instead gives a -1 to enemy infantry's death roll making winning combats harder but inflicting casualties more likely.

See my other posts on simulation for barbarians Vs romans here:

viewtopic.php?t=19109&start=120 (I’m not sure how to quote other threads correctly)

Calculating the results for impact and one round of melee. I’m also assuming the barbarians benefit from an overlap on one side in combat (its difficult to get it on both sides in a battle line).

Results without your change are:

Romans routed <1%
Romans fragged 4%
Romans disrupted 18%
Barbarians routed 14%
Barbarians fragged 23%
Barbarians disrupted 35%

Base loses:

5% chance romans lose 2
35% chance romans lose 1
28% chance barbarians lose 2
52% chance barbarians lose 1

Results with your change are:

Romans routed <1%
Romans fragged 5%
Romans disrupted 22%
Barbarians routed 10%
Barbarians fragged 20%
Barbarians disrupted 36%

Base loses:

7% chance romans lose 2
37% chance romans lose 1
21% chance barbarians lose 2
50% chance barbarians lose 1

This change doesn’t help the barbarians significantly. As I said in the other post the Romans would have 50% chance of winning at impact given they are Superior. This is where the issues appear to lie: once the barbarians are down at impact, being average, I don’t think they cannot easily recover. I suggest examining the impact interaction not SSw or CTs and this *may* imply changing barbarian quality or POAs at impact (though there are other possibilities of course).

[lawrenceg]

In order to reduce the effectiveness of Roman legionaries vs barbarian foot without having to remove the skilled swordsmen ability alltogether, how about this for a solution:

Skilled Swordsmen no longer gives a POA but instead gives a -1 to enemy infantry's death roll making winning combats harder but inflicting casualties more likely.

See my other posts on simulation for barbarians Vs romans here:

viewtopic.php?t=19109&start=120 (I’m not sure how to quote other threads correctly)

Calculating the results for impact and one round of melee. I’m also assuming the barbarians benefit from an overlap on one side in combat (its difficult to get it on both sides in a battle line).

Results without your change are:

Romans routed <1%
Romans fragged 4%
Romans disrupted 18%
Barbarians routed 14%
Barbarians fragged 23%
Barbarians disrupted 35%

Base loses:

5% chance romans lose 2
35% chance romans lose 1
28% chance barbarians lose 2
52% chance barbarians lose 1

Results with your change are:

Romans routed <1%
Romans fragged 5%
Romans disrupted 22%
Barbarians routed 10%
Barbarians fragged 20%
Barbarians disrupted 36%

Base loses:

7% chance romans lose 2
37% chance romans lose 1
21% chance barbarians lose 2
50% chance barbarians lose 1

This change doesn’t help the barbarians significantly. As I said in the other post the Romans would have 50% chance of winning at impact given they are Superior. This is where the issues appear to lie: once the barbarians are down at impact, being average, I don’t think they cannot easily recover. I suggest examining the impact interaction not SSw or CTs and this *may* imply changing barbarian quality or POAs at impact (though there are other possibilities of course).

As Simon Hall pointed out, the important thing is to determine which side breaks first, for which you need to continue the simulation for more rounds of melee. If you do that, you will see the effect of the SSw extra +POA making a much bigger difference.

Even if the barbarians win the impact and the Romans are disrupted, with ++ in melee they still have a good chance of ultimately breaking the barbarians. Stopping the analysis after 1 round of melee does not capture this.

[shadowdragon]

In order to reduce the effectiveness of Roman legionaries vs barbarian foot without having to remove the skilled swordsmen ability alltogether, how about this for a solution:

Skilled Swordsmen no longer gives a POA but instead gives a -1 to enemy infantry's death roll making winning combats harder but inflicting casualties more likely.

I thought the issue was increasing the effectiveness of "barbarian foot" in general and not that the Romans were too powerful. After all the Roman's (the ones with skilled swordsmen capability) aren't exactly ripping apart the opposition in tournaments. Reducing Roman effectiveness would, of course, improve things for Romans versus barbarians but it would do nothing for barbarians against everything else. If there is an issue then it appears to be with barbarian foot effectiveness and not the Romans.

[martindneiluk]

In order to reduce the effectiveness of Roman legionaries vs barbarian foot without having to remove the skilled swordsmen ability alltogether, how about this for a solution:

Skilled Swordsmen no longer gives a POA but instead gives a -1 to enemy infantry's death roll making winning combats harder but inflicting casualties more likely.

See my other posts on simulation for barbarians Vs romans here:

viewtopic.php?t=19109&start=120 (I’m not sure how to quote other threads correctly)

Calculating the results for impact and one round of melee. I’m also assuming the barbarians benefit from an overlap on one side in combat (its difficult to get it on both sides in a battle line).

Results without your change are:

Romans routed <1%
Romans fragged 4%
Romans disrupted 18%
Barbarians routed 14%
Barbarians fragged 23%
Barbarians disrupted 35%

Base loses:

5% chance romans lose 2
35% chance romans lose 1
28% chance barbarians lose 2
52% chance barbarians lose 1

Results with your change are:

Romans routed <1%
Romans fragged 5%
Romans disrupted 22%
Barbarians routed 10%
Barbarians fragged 20%
Barbarians disrupted 36%

Base loses:

7% chance romans lose 2
37% chance romans lose 1
21% chance barbarians lose 2
50% chance barbarians lose 1

This change doesn’t help the barbarians significantly. As I said in the other post the Romans would have 50% chance of winning at impact given they are Superior. This is where the issues appear to lie: once the barbarians are down at impact, being average, I don’t think they cannot easily recover. I suggest examining the impact interaction not SSw or CTs and this *may* imply changing barbarian quality or POAs at impact (though there are other possibilities of course).

As Simon Hall pointed out, the important thing is to determine which side breaks first, for which you need to continue the simulation for more rounds of melee. If you do that, you will see the effect of the SSw extra +POA making a much bigger difference.

Even if the barbarians win the impact and the Romans are disrupted, with ++ in melee they still have a good chance of ultimately breaking the barbarians. Stopping the analysis after 1 round of melee does not capture this.

Clearly the analysis shows, as the others do, that the barbarians are dead ducks after one round of melee and impact under the proposed change and the current rules. Continuing the simulation will only make the results worse for the barbarians. Doing n rounds of -- POA melee when they are at 66% chance of being at least disrupted at the start is just a good way of eating up CPU cycles. In some circumstances the "number of rounds to break" is an an interesting metric and useful but not one needed to test the hypothetical impact of rule changes desigend to address atomic interacts per melee or per impact.

Remember - I'm not advocating the change just testing it out to see if it has the intended effect as propose d by the message poster (I'm not pro or anti barbarians or romans here either).

[Strategos69]

I agree with martindneiluk: once you have the first two rounds, the rest is just a matter of mad calculations that will not add much. (I would recall here that it is statistically possible to have some combinations of infinite rounds of combats and I assure you wouldn't like to have your computer doing that.) In fact, the idea of the Barbarians holding the Romans would only be interesting if the Barbarians had an advantage in the flanks, which does not seem either the case nor historically accurate. When Barbarians won, they did so by breaking through the Roman formation more than performing outflanking manouvers.

Therefore, as shadowdragon says, I don't think the point is makinf Romans less powerful, but Barbarians more worth playing. There are several options:
- PoA for Barbarians at impact for the 4th rank
- Modification to the CT (which we have seen alter the result by a small margin)
- Dividing impact foot into two categories, the one relying on a volley of javelins (Romans and Spanish) or the fierce of the charge (Germans, Gauls, Samnites and others) and you can give an extra PoA to the second ones fighting the first group

As I see combats between Barbarians and Romans, the most likely output should be a draw at impact (mabybe Barbarians winning impact but Romans succeding in their CT), then a Barbarian victory and finally a Roman victory. Afterwards, that's if Barbarians did not manage to damage much the Romans, then the legionaries shouldtake over them quite easily (the ++ PoA seems very accurate to what happened historically). And indeed that would be fun to play and really give them some chance to win.

I only have martindneiluk's calculations for impact but they seem that, if we change to the +1 PoA at impact, they point in the rght direction in my opinion.
In order
Romans disrupted 28.123% an increase from 13.659%
Draw 21.5%
Barbarians disrupted 13.08% a decrease from 26.823%
Barbarians fragged 3.329% a decrease from 6.708%
Romans fragged 2.525% an increase from 0.982%

I can only guess about what would happen with a round of melee but I guess that on the overall Romans would win but at a costly price and if the Barbarians break the line at a certain point, they would be able to use their superior numbers. Something interesting to give a try, at least...

[lawrenceg]

Clearly the analysis shows, as the others do, that the barbarians are dead ducks after one round of melee and impact under the proposed change and the current rules. Continuing the simulation will only make the results worse for the barbarians. Doing n rounds of -- POA melee when they are at 66% chance of being at least disrupted at the start is just a good way of eating up CPU cycles. In some circumstances the "number of rounds to break" is an an interesting metric and useful but not one needed to test the hypothetical impact of rule changes desigend to address atomic interacts per melee or per impact.

Remember - I'm not advocating the change just testing it out to see if it has the intended effect as propose d by the message poster (I'm not pro or anti barbarians or romans here either).

It's not clear unless you have the additional information that with a ++ in melee and superior, the Romans will ultimately win unless they are fragged after the first round of melee.

With the proposed change (from ++ to + in melee), this is changed from "will win" to "will probably win" (or, as Tim Porter has often commented in the past, if your troops (well, his troops anyway) have a single + POA they nearly always lose the combat). IMO you haven't quantified "probably" by stopping the analysis at this point.

Based on history, Romans should probably win. The complaint is that at present the probability is too high (looks like >95% according to your model).

FWIW if the ++ in melee stays in then I agree with your suggestion that increasing the probability of the barbarians winning the impact is probably the best way to even up the combat. If the Romans don't get fragged at impact then the ++ will nearly always ensure their ultimate victory.

Perhaps introduce "Barbarian Impact Foot (BIF)" which gets a net + POA vs normal impact foot (similar to SSwd vs Swd)?

[martindneiluk]

Clearly the analysis shows, as the others do, that the barbarians are dead ducks after one round of melee and impact under the proposed change and the current rules. Continuing the simulation will only make the results worse for the barbarians. Doing n rounds of -- POA melee when they are at 66% chance of being at least disrupted at the start is just a good way of eating up CPU cycles. In some circumstances the "number of rounds to break" is an an interesting metric and useful but not one needed to test the hypothetical impact of rule changes desigend to address atomic interacts per melee or per impact.

Remember - I'm not advocating the change just testing it out to see if it has the intended effect as propose d by the message poster (I'm not pro or anti barbarians or romans here either).

It's not clear unless you have the additional information that with a ++ in melee and superior, the Romans will ultimately win unless they are fragged after the first round of melee.

With the proposed change (from ++ to + in melee), this is changed from "will win" to "will probably win" (or, as Tim Porter has often commented in the past, if your troops (well, his troops anyway) have a single + POA they nearly always lose the combat). IMO you haven't quantified "probably" by stopping the analysis at this point.

Based on history, Romans should probably win. The complaint is that at present the probability is too high (looks like >95% according to your model).

FWIW if the ++ in melee stays in then I agree with your suggestion that increasing the probability of the barbarians winning the impact is probably the best way to even up the combat. If the Romans don't get fragged at impact then the ++ will nearly always ensure their ultimate victory.

Perhaps introduce "Barbarian Impact Foot (BIF)" which gets a net + POA vs normal impact foot (similar to SSwd vs Swd)?

Addressing this point and running the same analysis for 1 impact round and 4 rounds of melee (with the suggested rule change to make the Romans + in combat rather than ++ and -2 for losing against barbarians):

IMO you haven't quantified "probably" by stopping the analysis at this point.

Romans routed 11%
Romans fragged 13%
Romans disrupted 28%
Barbarians routed 52%
Barbarians fragged 20%
Barbarians disrupted 18%
By continuing the analysis we see the barbarians have a 5 times higher likelihood of routing than the romans in 3 turns (52%/11%). Just as definitive as the last result I’d say and quantifies “probably” pretty well I think.
The results are exact (not monte carlo) but come at the cost of needing about 1GB of RAM to run.

[Strategos69]

martindneiluk,sorry to ask again (I do because I think your application is great to really test alternative proposals), could you please do the same calculations as you did in the last post but with an extra PoA for the Barbarians at impact? They will be + at impact and -- at melee. In my opinion the -2 to CT should be only at impact, if necessary.I think that it would be interesting to compare both proposals in similar conditions.

[lawrenceg]

martindneiluk,sorry to ask again (I do because I think your application is great to really test alternative proposals), could you please do the same calculations as you did in the last post but with an extra PoA for the Barbarians at impact? They will be + at impact and -- at melee. In my opinion the -2 to CT should be only at impact, if necessary.I think that it would be interesting to compare both proposals in similar conditions.

Also the baseline situation of evens at impact, -- in melee.

I suspect that the barbarians will be routed in substantially more than 52% of cases.

[martindneiluk]

martindneiluk,sorry to ask again (I do because I think your application is great to really test alternative proposals), could you please do the same calculations as you did in the last post but with an extra PoA for the Barbarians at impact? They will be + at impact and -- at melee. In my opinion the -2 to CT should be only at impact, if necessary.I think that it would be interesting to compare both proposals in similar conditions.

Throwing user friendliness to the wind and using probabilities rather than percentages. Also using the convention “-3” for rout, “-2” for fragged, “-1” for disrupted and “0” for steady..... the results for one impact and 4 rounds of combat, for the above scenario using the BGs I specified in the original post, are:

Romans
-3: 0.098973
-2: 0.16492
-1: 0.30399
0: 0.43212

Barbarians
-3: 0.68172
-2: 0.16494
-1: 0.11162
0: 0.041713

After impact only (if interested):

Romans
-3: 0.0
-2: 0.12734
-1: 0.33507
0: 0.53759

Barbarians
-3: 0.0
-2: 0.056841
-1: 0.13668
0: 0.80647

[martindneiluk]

martindneiluk,sorry to ask again (I do because I think your application is great to really test alternative proposals), could you please do the same calculations as you did in the last post but with an extra PoA for the Barbarians at impact? They will be + at impact and -- at melee. In my opinion the -2 to CT should be only at impact, if necessary.I think that it would be interesting to compare both proposals in similar conditions.

Also the baseline situation of evens at impact, -- in melee.

I suspect that the barbarians will be routed in substantially more than 52% of cases.

Using the same scenario as the post above but with the current rules.

Romans

-3: 0.04585
-2: 0.079392
-1: 0.23035
0: 0.64441

Barbarians

-3: 0.75386
-2: 0.1337
-1: 0.083394
0: 0.029048

After impact only (if interested):

Romans
-3: 0.0
-2: 0.0285
-1: 0.16568
0: 0.80582

Barbarians
-3: 0.0
-2: 0.11738
-1: 0.27947
0: 0.60315

[Strategos69]

Thank you, Martin for the data! It is very useful for the debate. I copy all the results here for comparison purposes.


Null hypothesis
Romans
Rout: 0.05
Fragmented: 0.08
Disrupted: 0.23
Steady: 0.64

Barbarians
Rout: 0.75
Fragmented: 0.13
Disrupted: 0.08
Steady: 0.02

1) Proposed change (+PoA at impact for Barbarians)
Romans
Rout: 0.10
Fragmented: 0.16
Disrupted: 0.30
Steady: 0.43

Barbarians
Rout: 0.68
Fragmented: 0.16
Disrupted: 0.11
Steady: 0.041

2) Proposed change (Romans + in melee rather than ++ and -2 for losing against barbarians)
Romans
Routed 11%
Fragged 13%
Disrupted 28%

Barbarians
Routed 52%
Fragged 20%
Disrupted 18%

If we look on the side of the Barbarians, the second change would reduce more the probabilities of the Barbarians being routed. From the actual 0.75 to 0.68 in change 1 and 0.52 in change 2. In the other hand, they would be more likely fragmented or disrupted, whereas with the change 1 they either win or die (the probabilities of being fragmented or disordered are inferior). Basicaly the story with the proposed change 1 is that, if Barbarians win impact (and they are more likely to do so) they will be in a best position to win the overall combat. If Romans manage to win or draw, then the Barbarians are likely to be destroyed.

In my opinion, this second outcome should be the most desirable from a historical point of view. Barbarians would be this way a troop type very good at impact (the best one against other foot), but fragile if the Barbarians did not manage to break through the enemy quickly. I think this way the confrontations between Romans and Barbarians can be more interesting and would explain why it is a good idea to deploy your best troops on front and the worser in reserve just in case they are needed (as Caesar usually did, by the way).

[shall]

You won't be surprised to haer that we have been running similar multi-phase simulations wince the beginning with sampes of several 1000 at a time. Probably a reason why much of the combat system feels well balanced.

Martin, could you remind me what that base scenario is on which you are basing all of these numbers please. A little different to my runs, but I think you are running a different base case and the answer varies quite a lot depending on what you consider it the starting position to be.

Keep the debate and numbers going as very interesting reading. Am tracking it with interest .... We have a couple of pet solutions that we are playing with at present.

Si

[martindneiluk]

You won't be surprised to haer that we have been running similar multi-phase simulations wince the beginning with sampes of several 1000 at a time. Probably a reason why much of the combat system feels well balanced.

Martin, could you remind me what that base scenario is on which you are basing all of these numbers please. A little different to my runs, but I think you are running a different base case and the answer varies quite a lot depending on what you consider it the starting position to be.

Keep the debate and numbers going as very interesting reading. Am tracking it with interest .... We have a couple of pet solutions that we are playing with at present.

Si

Si

The original scenario was:

"...with equal points (balance, is very important IMHO): Roman BG of 6 superior drilled HF armoured SSw without general Vs Barbarian BG of 12 average undrilled HF protected Sw without general. Barbarians get benefit of overlap on one side only".

I don't take samples, the approach adopted is exhaustive.

It might be an interesting cross validation exercise to run the exact same scenario and then compare results. If results significantly diverge it might flush out any errors in either of our simulators (though I suspect yours is relatively bug free since it has been bedded in for longer; but I'm sure you will agree that it is relatively easy to make mistakes when doing on the fly changes to check on rule proposals).

Cheers

Martin

[shall]

Ok thanks Martin.

Will run it at the weekend and we can compare notes.

Si