1.6 feedback

[Ahriman]

You are correct, the Mahdi Zealot has an iPower of 0.

It seems like there are a lot of bugs here, potentially. As you say, this is probably from a missionary unit being the core, and the Sardaukar one is probably from a copy paste from the senior one to create the junior one.

Seems like this parameter needs a general audit. This would explain a lot of the very strange AI power ratings.

 

[Deliverator]

Seems like this parameter needs a general audit.

I agree and maybe review unit costs at the same time. The unit list was never really designed it has just evolved over time, so it not surprising there are some parameters that need refinement.

 

[Ahriman]

Somethnig to consider: a strength 10 unit is much better than 2 strength 5 units. So the power curve for units should increase non-linearly somehow.

 

[davidlallen]

I think the unit costs were designed, and reviewed. However, the power values are certainly random due to the historical evolution. Is there an equation one should use, given the cost, combat strength and movement points, to estimate the power? Then I can build a spreadsheet which has the unit name, these three parameters, and an estimated power; somebody can fill in a reasonable power and check the cost; and I can push these stats back into the file.

 

[Ahriman]

How about, for each unit:

power value = (unit hammer cost)/10 + (unit strength)*(1+0.02*#experience points)

The hammer costs tend to increase nonlinearly, and they are a representation that not combat units like atreides heirs are still powerful.
Drop the XP if necessary, its not that important.

And then 0 for transports, workers, settlers, spies, scouts.

 

[Deliverator]

It would be useful to have a spreadsheet of the vanilla iPower values and then perhaps we can reverse engineer a rough formula from that.

 

[Ahriman]

Possibly, but I suspect that the vanilla values are directly related to unit strength. In vanilla, there aren't really any units who are stronger than they appear because of extra abilities, but we have some low-strength units who should still count towards power ratings.

That was why I suggested hammer cost, since hammer cost proxies for unit value, but doesn't quite increase fast enough (a hammer cost 120 is probably more than twice as good as 2 hamemr cost 60 units).

 

[davidlallen]

Here is the sheet for DW 1.6.3. I just used "(cost / 10) + strength" for NewPower. Any comments on other cost or strength adjustments also welcome. I will be cutting 1.6.4 in the next six hours or so, but there are always more releases.

 

[Ahriman]

I would recommend setting the power values for spies and transports = 0. Or at least very much lower than in the sheet.
Otherwise looks reasonable.

 

[Slvynn]

Agree on reccomendation and i would not reccomend to assign power ratio for Scout Thopters. I think we talked about this enough
They are scout/ transport units that not add to civ's military power in any means.

 

[idahopotato]

Speaking of which, I had a scout thopter get attacked by a sandworm and somehow managed to kill it, the sandworm that is. How is that even possible?

 

[davidlallen]

Speaking of which, I had a scout thopter get attacked by a sandworm and somehow managed to kill it, the sandworm that is. How is that even possible?

This must have been in the very early game, and not on the highest difficulty levels. At the difficulties I play at (Noble and below ... I am no expert!) each human player gets a small number of free wins against animals. So for the first one or two combats, no matter what the odds, the human unit will win. Try it -- with Fremen, for example, build an initial warrior (which has sandrider promotion and can walk on desert) and attack a worm.

 

[Slvynn]

Is it possible to negate this in puprose of Lore preservance?
seems it should be something in difficulty level definitions....

 

[Ahriman]

Agree on reccomendation and i would not reccomend to assign power ratio for Scout Thopters.

Scout thopters *are* transports.

Pretty funny about the low difficulty wins vs animals, I'd forgotten about that effect.

 

[Slvynn]

Scout thopters *are* transports.

I knew, again you trying to correct me, while i just supporting this opinion , %-}
Tell me please am i permitted to explain things i want to explain?

remebering of Thopters was a kind of joke, calling back yesterday tension with a smile, and i find amusing being corrected same way again. There is very main, valid point, which is not related to game. Can you deny it?

 

[Ahriman]

That wasn't meant to be an insult or attack here, just a clarification. Sorry if you took it as something else.

While not trying to get pick a fight, its worth noting that GodEmperor's explanation of some of the tag effects suggest that the change you were arguing for yesterday (making scout thopters strength zero) would not have had any effect in the manner you claimed it would, because the AI already valued scout thopters at zero ipower. And the thopters already had transport AI. They were sitting in the city as unused transports, not because the AI thought they were city defenders.

 

[God-Emperor]

It might be good to have some simple modifications for abilities of units that are not transports, workers, or settlers with each factor giving a minimum of +1 if applicable:

• If it has a base speed > 1 then +10%
• If it has a combat bonus that is +25% or less against something then +10%
• If it has a combat bonus that is > +25% against something then +20%
• If it has a defensive bonus that is +25% or less in some situations then +10%
• If it has a defensive bonus that is > +25% in some situations then +20%
• If it has some effect provided entirely via python (and so hidden from the AI) then perhaps a flat +2.

These are just guesses at adjustments that might be useful.

There is a long discussion over in the BBAI forum on "better unit iPowers" where it is suggested that using some non-linear adjustment from iPower is probably better. Perhaps (iPower)^1.5.

On the other hand, by including a factor proportionate to iCost as well as the iPower at least some of this is already included. Because of this, the "^1.5" is probably not necessary to get reasonable values.

Don't forget to use iAirCombat, instead of iCombat, for air units. (I haven't checks the spreadsheet to see if this was done.)

EDIT: I just checked - the spreadsheet uses iCombat for air units, which is 0 for all of them.

 

[God-Emperor]

After a quick look at the spreadsheet...

Because of the cost related component in the new iPower ratings, the adjustments I suggested for the various combat and defensive bonuses are probably not needed. The speed and python related bonuses might still be a good idea.

Looking at the air units, if you add in the iAirCombat they look a bit high. They would range from from 17 for the Wasp Interceptor (about the same as a Missile Trooper) to 42 for a Cielago Stealth Bomber (more than a Lasgun Trooper, similar to a Sardaukar Noukker). Perhaps adding in only half of the iAirCombat would be better (giving a range of 13 to 31, the range from Grenade Trooper to Bee Sting Trooper). Rating air units is relatively hard - they can have a large effect, but they count for exactly nothing in defending a city against attackers once they actually attack (other than other air units, in the case of the interceptors).

 

[Ahriman]

It might be good to have some simple modifications for abilities of units

My feeling was that utility issues not captured in strength would be captured in hammer cost.

EDIT: I just checked - the spreadsheet uses iCombat for air units, which is 0 for all of them.

Good catch.

Perhaps adding in only half of the iAirCombat would be better

This seems reasonable.

 

[davidlallen]

When you are ready, please update the new values into the sheet, and I will push them into the xml for 1.6.5.