Upgrade obsolete troops or build anew?

[drlake]

Every game I play, I seem to develop an army that contains a surprising large percentage of obsolete troops. When I look at the cost of upgrading them, I start to question under what circumstances that makes sense. I'm tending toward the following "rules", but want some input from you guys on this:

1) If the unit has more promotions than new production, it is worth upgrading.

2) If the unit has the same number or fewer promotions than new production, it can be replaced unless by upgrading you create a unit you can't build (like a MG with CG promotions).

3) If the unit is a critical defender for a city facing imminent attack, it is worth upgrading.

Under most circumstances, it seems like I can just replace and eventually disband obsolete troops, since the gold cost of upgrading them is far more than the hammer cost of building a new one. Am I missing something?

 

[mariogreymist]

It also depends on whether your kingdom is more adept at producing or . If you're playing a primarily economic game, it is often preferable to upgrade units.

 

[UncleJJ]

The only circumstance when I intentionally upgrade troops, other than those you mention, is to rapidly exploit a newly acquired technology. A good example is when you are about to research Steel. Then I'll position a stack of trebuchets together with other troops on an AI border and also set other cities to building new trebuchets. Once Steel is researched the city build orders are updated to cannons and I spend the gold to upgrade the existing trebuchets to cannons (costs 80 gold on normal). Next turn declare war.

Of course these are not really "obsolete" trebuchets, they are being build to use this upgrade trick, so it is not a case that your original question was about. I use a variation on that technique when I'll send obsolete troops like pikemen and archers as part an overseas expedition knowing that when I research Rifling I can upgrade them and have rifles on the new continent.

Upgrading troops is generally more cost effective on marathon than other speeds. I also play on larger maps with marathon so troops take longer to get places and troops take a lot longer to build (they cost double hammers). So upgrading has more tactical / strategic purpose as well as being cost effective.

 

[Riot]

Every game I play, I seem to develop an army that contains a surprising large percentage of obsolete troops. When I look at the cost of upgrading them, I start to question under what circumstances that makes sense. I'm tending toward the following "rules", but want some input from you guys on this:

1) If the unit has more promotions than new production, it is worth upgrading.

2) If the unit has the same number or fewer promotions than new production, it can be replaced unless by upgrading you create a unit you can't build (like a MG with CG promotions).

3) If the unit is a critical defender for a city facing imminent attack, it is worth upgrading.

Under most circumstances, it seems like I can just replace and eventually disband obsolete troops, since the gold cost of upgrading them is far more than the hammer cost of building a new one. Am I missing something?

Nope, you got it right. 3 is the most critical condition, upgrading a unit there is a good idea usually. But I think upgrading a unit cancels out any fortification bonus too? Not sure.

If a unit has a bunch of relevant promotions, it might be ok. There's no reason to upgrade say a Shock Chariot into a Shock Cuirassier (Pikes aren't really a threat, and there are no more melee units after gunpowder becomes relevant).

Having obsolete units around isn't a huge deal - you should be building military anyway because you need them for:

1)barb defense
2)AI defense

Maybe you're going a bit overboard on units? Having too many of them costs money after all.

p.s: Don't bother building units for power rating, diplomacy matters much much more.

 

[UncleJJ]

But I think upgrading a unit cancels out any fortification bonus too? Not sure.

It does cancel the fortification bonus, so upgrade your troops as soon as you see it's going to be needed.

p.s: Don't bother building units for power rating, diplomacy matters much much more.

I frequently build vast amies and navies specifically to make chain capitulation easier. Power rating does affect the game and unless you're playing on immortal or deity it is easy to overawe most of the AIs with sheer power. Then you need only bother with diplomacy with the other powerful AIs, no silly demands for techs or gold from the tiddlers. Besides, who cares about diplomacy if your army is big enough?

 

[Yxklyx]

If my economy can make 1000 gold a turn and I want to go to war and I have lots of obsolete units - I'll upgrade them using one to three turns at 100% gold income.

 

[drlake]

If my economy can make 1000 gold a turn and I want to go to war and I have lots of obsolete units - I'll upgrade them using one to three turns at 100% gold income.

Good point. I usually don't have that kind of income to play with, at least not in the games I've played recently. Depends in part on the leader I'm playing, I suspect.

 

[Zexx]

I will often upgrade my troops when I have gained a tech lead on an AI that I doubt I can keep for very long.
If I'm one tech ahead of an Immortal Boudica, and I get, say, Artillery before she does, I will slap my sliders to 0% and upgrade cannons to artillery as soon as I can, since an advantage can often determine a war before it even starts.
Now, at some point in this war, she will have Artillery. The AI just techs that fast, that they can be behind, spam soldiers, and still tech quickly. So the advantage won't last, but while it does, I intend to exploit it. (And that's another bone of contention. Seriously, a mediocre production city popping out marines in 1 turn? That's crap. AI bonuses are silly.)

 

[DigitalBoy]

Upgrading a highly promoted unit or upgrading a unit in a city that's about to be attacked are both good reasons to upgrade, but the one situation where I upgrade like crazy is

Spoiler :

knights to cuirrasiers; if I've built any significant numbers of knights, it's probably to wage war, and at 50g per unit, upgrading knights is a cheap way to keep the mayhem going.

 

[tempuraki]

If the units are CR melee units, then I upgrading to gunpowder units. otherwise they stay back and make cities happy.

 

[Riot]

It does cancel the fortification bonus, so upgrade your troops as soon as you see it's going to be needed.



I frequently build vast amies and navies specifically to make chain capitulation easier. Power rating does affect the game and unless you're playing on immortal or deity it is easy to overawe most of the AIs with sheer power. Then you need only bother with diplomacy with the other powerful AIs, no silly demands for techs or gold from the tiddlers. Besides, who cares about diplomacy if your army is big enough?

Haha. What I meant is, building units to try and build your PR up enough so that AI's won't attack you. Deterrence doesn't work at all.

 

[mariogreymist]

Upgrading troops is generally more cost effective on marathon than other speeds. I also play on larger maps with marathon so troops take longer to get places and troops take a lot longer to build (they cost double hammers). So upgrading has more tactical / strategic purpose as well as being cost effective.

I think this reasoning is fundamentally flawed. Upgrades on Marathon have a 3x cost multiplier, and building units only 2x. So for maximum efficiency, I'd say building has more effect.

Not to say that the idea of using upgrades to have a modern army on turn one with steel, rifling etc isn't particularly good at marathon (a speed which favors aggressive wars); but as to the idea that it's cost effective vs building new troops: it's not.

 

[UncleJJ]

I think this reasoning is fundamentally flawed. Upgrades on Marathon have a 3x cost multiplier, and building units only 2x. So for maximum efficiency, I'd say building has more effect.

Perhaps you could do some worked examples for us for both marathon and normal speed. I suggest the traditional upgrade of a mace to a rifle. Then we could see how flawed my reasoning is

 

[WeaselSlapper]

One of my favorite things to do is upgrade my CR3 macemen to Rifles or Infantry and send them out to dominate AI cities, I find this epically useful when I'm the first to Assembly line (which I usually am) CR3 Infantry against enemy rifles will usually have 90%+ odd of winning without sacrificing any siege units. I usually find myself building as many maces as my production can spit out just before I get rifling just to use this trick (usually this units are only CR2 but that's still plenty enough extra attack to overwhelm the AI)

 

[mariogreymist]

Perhaps you could do some worked examples for us for both marathon and normal speed. I suggest the traditional upgrade of a mace to a rifle. Then we could see how flawed my reasoning is

UJJ: Actually...I take it back. Working it out, the for upgrades is based on the cost of the units in question, so the discount in construction cost (2x) carries over from the maceman while the gold invested benefits from the 3x marathon multipliers. You're right. (not to mention the base 20 fee is less significant at marathon)

 

[Mudcrab]

Upgrade your CR units, the experience is hard to recreate. Probably no need to upgrade defenders, unless for rule 3. You can opt to upgrade horse units, but look at their promotions. It is sometimes useful to sacrifice old obsolete horse units with critical improvement raids on your enemey. Also upgrade your stack defender units, because otherwise you CR units will be cut down by enemy horses.
Also, the army upgrade trick is good, as mentioned above. If you are getting close to rifling or steel, consider you cash situation and prepare for war. Cash up by a choice of :
decrease the science slider to increase cash per turn, but be mindful of teching by your enemy;
get a few cities to produce wealth to increase cash;
attempt to generate great merchant, or save an earlier one for a trade mission;
trade resources, techs and maps for cash with neighbors but not the target - this also helps to create forced peace with your neighbors, so worry less over backstabbing.

And be very aware of experience when upgrading - it gets reset to 10, so don't upgade a unit 1 or 2 batttles away from the next level.

 

[drlake]

One of my favorite things to do is upgrade my CR3 macemen to Rifles or Infantry and send them out to dominate AI cities, I find this epically useful when I'm the first to Assembly line (which I usually am) CR3 Infantry against enemy rifles will usually have 90%+ odd of winning without sacrificing any siege units. I usually find myself building as many maces as my production can spit out just before I get rifling just to use this trick (usually this units are only CR2 but that's still plenty enough extra attack to overwhelm the AI)

Ooooh, I like this! I just need to figure out how run my economy to generate a larger surplus to pay for the upgrade costs...

 

[WeaselSlapper]

Ooooh, I like this! I just need to figure out how run my economy to generate a larger surplus to pay for the upgrade costs...

Normally I save my gold for a long time for this (don't spend to hurry or upgrade other units) so I don't have to dump my research slider down to 10-20% total (I use 10-20% for research because at 10-20% you still get a little research progress at usually a minimal cost of gold per turn). If I wasn't able to save my gold and I want to upgrade them all in a hurry I dump my slider and stockpile a bunch of gold or I sell some techs to the AI I'm not planning on killing. I usually won't turn my slider down until after I get Assembly Line because I love getting the Pentagon and I want my production city to get started on that right away.

 

[TheMeInTeam]

It does cancel the fortification bonus, so upgrade your troops as soon as you see it's going to be needed.



I frequently build vast amies and navies specifically to make chain capitulation easier. Power rating does affect the game and unless you're playing on immortal or deity it is easy to overawe most of the AIs with sheer power. Then you need only bother with diplomacy with the other powerful AIs, no silly demands for techs or gold from the tiddlers. Besides, who cares about diplomacy if your army is big enough?

Similar to your hammering me for refusing to acknowledge SE, CE, or GE, I'm hammering you for again implying something different than how the cap mechanics work.

Since the cap bottleneck is almost always the average of all civs, every unit you build is getting diluted by all civs alive. Every unit you kill is much more important.

And the whole "chain" thing is to abuse the land target rule, which further makes your power less relevant. Yes, you need to be over a certain amount, and you almost definitely are after clearing their SoD and taking a few cities if you have enough land to hold them...after that your power matters almost nothing except a painfully diluted factor in the average.

Games where power is sufficient to deter garbage AI DoWs are games that are over. I have yet to see even a single exception to this in a competent, high level player's game. Especially because if you're at war, dogpilers get to add the power of your enemies too. A worst case is an asp hat like monty where you need almost 200% of his power at peace, and now he's stacking the power of your war enemies into his for this check ---> if you're strong enough not to fail checks like that, you should be in autopilot if you're not playing for HoF. Just to make sure this is clear: If you wind up at war with a powerful AI (as in one that is about your strength, virtually nobody is too weak to declare on you. Many would consider themselves strong enough with one warrior/city

The real use of military is offensive and defensive protection.

As for upgrades being more efficient on marathon, yes and no. Upgrades are a function of hammer cost + some extra per upgrade, and as such they are discounted on marathon, but so is unit production. I'm not sure if mara upgrades come out ahead in the *relative* sense compared to upgrades on other speeds. If they do, it's probably not by a great deal (like some kind of bad scaling on the per upgrade cost). However, the point about the tactical advantages of it rings painfully true. Units are just so much faster on mara.

 

[Tomice]

TMIT, do they count my vassals and def pacts as well?

I had this situation:

For ages my neighbour Fred was way ahead in power, cause I expanded to another continent and crippled my economy. When the investment started to pay, I almost catched up to his power. During centuries of superiority, he never declared war.

Now I wanted to have a mediocre AI which was in war with 2 crap AI's as vassal, and I ended up in war with them. Suddenly, one turn later Fred and his vassal declared war.


So your theory seems perfectly true, but I didn't mention I had a def pact that broke away when I accepted the new vassal and got engaged in his wars.

Question: What made him declare, the 2 crap AI's he adds to his power or the end of my def pact? Or the combination of both? What role does the new Vassal play in the AI's considerations?