SGOTM 13 - Gypsy Kings

[bcool]

There are 5 reasons I think we should whip the workboat in FH on T108.
-It gets the workboat out faster (more time to explore)
-it allows whip anger in FH to dissipate faster.
-It allows FH to grow to 4 pop next turn without the Corn Farm.
-I don't think it slows down the pyramids if we whip it earlier since the hammers that would have gone to the workboat start to go to the pyramids earlier.
-I need the Corn Farm in CC for my plan so that 4th pop in FH really has nothing great to work.

 

[adrianj]

I haven't posted in a while, so just thought I'd mention that I'm still eagerly reading everything and I'm pleased with how things are going. A lot of the discussion seems to be MM, which I think is good, however in this respect I can't really contribute.

 

[mabraham]

editted this post http://forums.civfanatics.com/showpost.php?p=10321701&postcount=692 to include tile and build plan

T142 your plan double-uses the PM

 

[bcool]

hmm, I must have not recorded it accurately. I'm assuming that I used the Grassland forest instead of the the PMine in CC.

edit: I just checked my autosaves and yes I used the GF instead of the PM in CC on T142

 

[mabraham]

hmm, I must have not recorded it accurately. I'm assuming that I used the Grassland forest instead of the the PMine in CC.

Yeah, and perhaps also on T141 instead of the engineer. My spreadsheet thinks the extra 2 is needed to grow to 6 on T142, but I haven't play-tested yet.

 

[Ronnie1]

After looking more closely....I don't like the fact that 2 Clams has to build its own workboats. This really slows down its development. I think we need to get at least 1 if not both from somewhere else.

At 1st glance, the extra city(over my last test) seemed to heavily outweigh my play through, but after closer inspection, I do have significantly more infrastructure and only 1 less pop point which gets made up on the very next turn when I found Piggery.

I think maybe a combination of the 2 paths is a better way to go.

You are right about the early WB whipping, there is no difference between the 2 options at T112.

More testing soon.

 

[bcool]

Yeah, and perhaps also on T141 instead of the engineer. My spreadsheet thinks the extra 2 is needed to grow to 6 on T142, but I haven't play-tested yet.

I have it growing to 6 IBT T141-T142 with the engineer (with 6 OF food)

 

[mabraham]

After looking more closely....I don't like the fact that 2 Clams has to build its own workboats. This really slows down its development. I think we need to get at least 1 if not both from somewhere else.

One relevant point here is that the fishes at BF and SM are worth +1 over those clams, so they're somewhat preferable.

Moai in PC is nice, but not a game-breaker. It can't get built before about T180, and only then can it start to contribute to the empire - and by then most of the growth has taken place. The 45 for 9 turns working the PM for one WB costs 9 turns on the Moai arrival, which is something like 45-63 working 5-7 water tiles over that time. The other WB is probably an extra whip, for some smaller actual cost, I guess. PC is also getting built pretty early, which we don't have to transfer into an earlier Moai time if that compromises some other city - the earlier settling gets an earlier PM in PC to build another WB...

Meanwhile the two WB produced elsewhere have actually gone onto working at least one other fish earlier, which gets (for example) BF up and growing much faster than it can otherwise do. It now becomes a serious candidate for whipping further workboats, galleys or triremes according to need. If we could do it all, then that'd be nice, but the best WB-building plan I can get out of SM is T126, 135, 137(w), 148, 153(w). By now, MC has built its own WB.

Two compromises are:

• T126->PC (which builds its own second WB), T135->SM, T137->BF, T148->Crabby (maybe later if it has to pop a border), T153->somewhere (or maybe a galley or granary instead)
• T126->BF, T135->SM, T137->PC (built its own by now), T148->Crabby, T153->somewhere

I can live with either, but I think an early PC needs to pull some of its own weight.

At 1st glance, the extra city(over my last test) seemed to heavily outweigh my play through, but after closer inspection, I do have significantly more infrastructure and only 1 less pop point which gets made up on the very next turn when I found Piggery.

Some of that infrastructure difference is the two lighthouses you mentioned a few posts back.

I think maybe a combination of the 2 paths is a better way to go.

You are right about the early WB whipping, there is no difference between the 2 options at T112.

 

[mabraham]

I have it growing to 6 IBT T141-T142 with the engineer (with 6 OF food)

Hmm, an 8- discrepancy suggests a significant error somewhere, either in my spreadsheet, my transcription, or the records you kept. I can see quite a few points over time where not all the 8 good tiles are worked, so I'm going to take my transcription of your plan as a starting point, and see what I can optimize.

For example, T131-132 FH is not working its GM while CC is working a GF trying to sneak under the worker-whip overflow cap. Slightly better is to shift FH to the GM for a turn, so that CC can work one turn on the PM and one turn on a coast. CC now makes 24*1.252, rather than 42, while now actually maxing the overflow off the worker. We trade a hammer on the Pyramids for a hammer on the GLib (eventually) and pick up 2.

 

[bcool]

I transferred my notes to the spreadsheet correctly, so I don't know where the error is. I appreciate the error checking/optimization. I apologize if it turns outmy notes are incorrect. I did my best to be accurate.

Good call on the example T131-132.

A closer accounting of the total outcomes of ron's test game and mine...

If my numbers are right, he is equal in pop (essentially) up 41 hammers up in infrastructure and current hammers 21 food up in bins
but I'm up 1 settler. (149 hammers/food) (this takes into account what I would make on T150)

If you guys can further improve on my plan with tweaks or better small city whipping, great.

 

[bcool]

I'm actually thinking the parthenon in either Bronzed Fish or Paired Clams will be more profitable than Moai 1st. Plus there is a chance we lose the race to Parthenon where as Moai can be built later (or milked for failure gold as we see fit.)

--failure gold from wonders that are built with production bonuses
If we partially build Moai in several cities then when we finally complete it in a city we get all those hammers invested in other cities converted to gold (basically allowing us to build wealth with a free +100% modifier). The limitation is we have to take turns building it because only 1 city at a time can build it.

This is less desirable to do with wonders that we actually want and can take advantage of their benefits. The Parthenon could give us 11 a turn in the capital which amounts to ~33 per turn. Not bad considering its cost 600 hammers before the marble discount. Lifetime benefit is a minimum of 80 turns *33 or ~2640 (The benefit is obviously greater if we have more than 1 city producing GS for us.

 

[Ronnie1]

The more I look at bc's last test, the less I actually like it. One of the things I thought we were trying to accomplish was to get side cities up and working the best available tiles ASAP. I'm not sure this was necessarily accomplished very well in that test.

Another other thing I don't really care for is how much pop was whipped off. I thought we discussed actually carrying some into the Mids because as soon as we switch to Representation they immediately become useful.

I see the point about the fish at Copper being +1 over the clams at 2 Clams, but none of the other tiles at copper are not workable before IW. If we want that food sooner, then we could just settle that site sooner and settle 2 Clams as city #5-6. By that reasoning, even the crab site is more valuable the 2C because the crabs are a health resource we don't already possess and there are forests there that can get chopped into something/anything useful.

I think we also talked about NOT planting cities just to plant them; ie, what will they add to the empire and what is the cost of doing so?

The more I look at what my test produced by T150, the more I like it.

1) Cities that are capable of producing , through the whip at least in the case of Marble.
2) 5 base at size 3, 7 at size 4, in 2 Clams with a LH + Granary in place for straight building or whipping there. Moai is stuck in the que at the end of my test, but that could easily be warrior MP's or Triremes or whatever we need.
3) Copper was settled and had Fish available on its second turn. When 2 Clams was settled it had 1 clam available on its 8th turn and both nets available when it hit size 2!
4) SM had its fish netted within 3 turns of being available BUT within 1 turn of the pop to work it.
5) FH is size 5 instead of size 2 when the Mids are built.
6) I know we talked about Marble + Copper starting with Lighthouses so they could whip those ASAP, that did not happen in bc's test.

I am going to run through another test, but I really think we should look at the results of this http://forums.civfanatics.com/showpost.php?p=10322030&postcount=693 test on its own merits.

EDIT: I am sure we will be able to trade for some techs somewhere in the next 40 turns, so a lot of this is likely to change also, another reason IMO to not necessarily get in a hurry to put cities down "just because". By building cities that actually have workable tiles ASAP, it gives us much greater flexibility to adapt to the changes that are sure to surface in that time frame.

 

[mabraham]

I'm going to take my transcription of your plan as a starting point, and see what I can optimize.

I can't improve much early on. I like the way he is switching settler builds to suit the happiness situations. (That also found a rare bug in how I was modeling the granary.) The T114 settler means we fail to work one of the GM tiles for several turns, but that settles PC faster to get more turns on its PM. Thus it is a net profit.

T119 FH is building a settler and CC a forge. bc has FH working a non-lighthouse clam and CC the PM. If we swap these, the settler build gets the same value. CC gets more food, which leads to growth one turn earlier at T127 when there was a GM being wasted. The forge builds one turn later on T128, but since it is working no mines at that time, there is no loss there. Net at T128, CC has lost 4-3=1 to gain 5+1-2=4. FH is the same. The is irrelevant.

T120-4 CC is building a pre-forge settler. To sneak under the max overflow cap, bc was working GM+GF on T122, and GF+coast T123, to get 752. Slightly better is to work GM+coast on both turns, to get 664.

I still don't see how bc had time to work an engineer at T141. With the above modifications, I stopped growth just as the box filled (fixed a bug here too) on T140 to avoid growing + before we wanted to switch to a settler. I think there is merit in CC switching to a worker on T142, actually.

BF is good for a workboat whip about T143, too.

Attached updated spreadsheet and PDF for my several improvements. From a play-test, it's accurate to at least T144.

Importantly, I don't think there is much to improve or change in about the T108-T120 window, so perhaps we should lock something in soon. Thorn's PPP also needs to include the opening few turns on Paired Clams (e.g. build WB working the PH for 5 turns to get to 15 while waiting for it to become a PM to build the WB in 6 more turns), and the fact that the galley is coming straight back to the mainland without the worker.

 

[mabraham]

The discussion seems to have gone a bit adversarial between these two test game approaches. That's OK - a hot forge makes good steel - but I thought I'd make a friendly reminder that we're all trying to get the same objective

The more I look at bc's last test, the less I actually like it. One of the things I thought we were trying to accomplish was to get side cities up and working the best available tiles ASAP. I'm not sure this was necessarily accomplished very well in that test.

We do need to be careful to make fair comparisons. A city settled earlier than it can have good tiles to work is not necessarily a waste. Up to about 6-7 cities they all pay for themselves in trade routes and such. In the absence of data recording when good tiles were available and worked, we can look at the total outputs at T150. Assuming bc was looking at your latest game, they sound pretty even, perhaps a slight edge to bc's game overall.

Another other thing I don't really care for is how much pop was whipped off. I thought we discussed actually carrying some into the Mids because as soon as we switch to Representation they immediately become useful.

Sorry, I don't see the point you are making here. In all the game reports - including your latest and bc's latest - at about T150 CC was at size 5 with about 35 turns for whip- to decay. It's unrealistic to hope to be around size 8 by T150 while Duckweeding stuff - and we've still got at least one more settler to whip onto GLib and/or NE. We're also short of until we get trades. We want to hover around sizes 3-8 while doing 3-whips onto settlers and 2-whips onto workers until about T170. Representation just eases our constraints. We should start to grow the population by about the time we start the NE, so that we get to whatever is our maximum sustainable size fairly soon after it is most profitable to be there. Our last whip- isn't going to decay until T185 on the above numbers, plus one or two more whips yet...

I see the point about the fish at Copper and Marble being +1 over the clams at 2 Clams, but none of the other tiles at copper are not workable before IW. If we want that food sooner, then we could just settle those sites sooner and settle 2 Clams as city #6. By that reasoning, even the crab site is more valuable the 2C because the crabs are a health resource we don't already possess and there are forests there that can get chopped into something/anything useful.

Paired Clams is good to settle fourth because it has three good tiles to work eventually. However it's going to get one of those pretty quickly (PM), and its next best two are not as good as the only one at BF. The question is whether improving PC is going to return better value to the empire than the extra 1 from improving BF.

If PC were going to build something of strategic value soon, then I'm all ears. In the fullness of time, PC under Caste can be at size 6 working PM, two Pworkshops, two clams and growing working a single coast putting out 11/turn. That's the best -pump we're going to see from our current land - though when BF and SM get all their workshops, they'll nearly equal it. By T190, PC can definitely finish the Parthenon by the time we really want it.

On the other hand, BF is just fine working its fish and coasts to whip lighthouse, granary and boats until we get worker and IW and border pop organized. Both cities are capable of doing useful stuff - but if prioritising Parthenon or MS is what we want to do, then we should be helping PC before BF.

I think we also talked about NOT planting cities just to plant them; ie, what will they add to the empire and what is the cost of doing so?

I don't see anything in any test game reports that are wasteful... We might choose to plan Crabby differently after we scout.

The more I look at what my test produced by T150, the more I like it.

1) Cities that are capable of producing , through the whip at least in the case of Marble.

I don't see a material difference. The state of tile improvements looks about the same. You have one more granary, three more pop in FH, but one fewer city founded, one fewer workboat and one fewer total pop.

2) 5 base at size 3, 7 at size 4, in 2 Clams with a LH + Granary in place for straight building or whipping there. Moai is stuck in the que at the end of my test, but that could easily be warrior MP's or Triremes or whatever we need.

I reckon bc's Paired Clams site is about 10 turns behind yours, by the time it grows to 4, whips the granary and gets back to a half-full food box. To offset that, he's had three turns on Crabby and one on Piggery ahead of you. Your Fish Hills could scamper out a settler in about 6 turns, but you'd need a workboat for it also.

3) Copper was settled and had Fish available on its second turn.

bc's game had it available on its first turn - in 750BC, not 425BC.

When 2 Clams was settled it had 1 clam available on its 8th turn and both nets available when it hit size 2!

Yes - you've started working a different seafood tile early.

4) SM had its fish netted within 3 turns of being available and the pop to work it.
5) FH is size 5 instead of size 2 when the Mids are built.
6) I know we talked about Marble + Copper starting with Lighthouses so they could whip those ASAP, that did not happen in bc's test.

SM got its fish at the same time yours did, I expect. FH has probably built one fewer settler in your game. bc's game could have been better with more lighthouses - but at T150 you each have four.

EDIT: I am sure we will be able to trade for some techs somewhere in the next 40 turns, so a lot of this is likely to change also, another reason IMO to not necessarily get in a hurry to put cities down "just because". By building cities that actually have workable tiles ASAP, it gives us much greater flexibility to adapt to the changes that are sure to surface in that time frame.

Sure, exactly what we plan to do will change within and between turn sets.

The only material advantage I see in your test game is that the concentration of empire resources on PC means it is in good shape by T150 to make a strong contribution to something of strategic value. bc's fast-settle approach can still do that. The only change is what happens to the first three workboats out of SM. The first is due T126 and can go to PC just after it has built its own lighthouse. SM keeps the WB due on T135 for itself, and sends the one (maybe) whipped on T136 to PC to plant T142. The PC granary can then be whipped T144 and by T149 PC is at size 3 working its good tiles and can start paying back to the empire. Meanwhile, MC and BF have been sucking hind tit, and SM hasn't got to work its Gmine much at all. Probably MC should slow-build its own WB and BF whip its own WB.

If we don't whip SM on T136, PC can still slow-build the granary by about T150 until the second WB arrives. This could well be best - we're swapping working the SM Gmine for one of the PC clams for a bunch of turns.

 

[mabraham]

I think the critical difference between the two plans is that bc is trying to get new cities out ASAP, whereas R1 wants to do so only with workboat support. Hence, the difference in FH of whipping a settler and a workboat (respectively) onto the Pyramids around T120.

As far as FH is concerned, it seems to make more sense to whip a settler or worker onto the Pyramids, because we make use of a whip opportunity, and we get to whip more of our into scarce when using multiple population to get a settler or worker, and we get to convert overflow into on the wonder. This offsets the greater time we have to work the non-wonder to get into whipping position. Seemingly, a workboat 1-whip only has a limited conversion of ->, and there is no advantage to the wonder from the overflow.

However, if the creation of a settler generates unresolvable pressure on workboat availability (as discussed in previous posts, we want five plantable workboats between about T110 and T130), it might be better to 1-whip a workboat in FH and focus the food onto CC to push another whip-cycle through there early. This delays settling (say) BF somewhat, but lets us get three workboats at about T124(FH), T126(SM), T135(SM) with the option of whipping one more in SM to arrive T137.

So, lets look at the value we get for that early WB out of FH.

First, we consider using FH to hurry to get one of the two WBs for PC
R1's spreadsheet suggests those first two WBs head off to PC. They arrive and plant ready for work on T127 and T130. We have to settle on T116 to have grown to 2 by the time both are available - but bc's plan settles it T117, and R1's plan settles it T120. So this second workboat is 1 or 2 turns early (respectively) if you crunch the numbers. It's then not until T137 or T138 (respectively) that we grow to size three to finally start working the Pmine - that's 20 or 18 turns after settling! The lighthouse still isn't finished until about T144, and it doesn't pay to whip off the population working the Pmine just to build a lighthouse, so we grow to size four at about T146 or T147 (respectively) before 2-whipping the granary to regrow to size 3 on about T151. (This is a whole population point better than R1 actually did in his game - he seemed to whip the granary three turns later than I think is best.)

Meanwhile, CC built a settler for MC, and then a settler for BF. BF settled T145 and got its fish improved T146 in R1's spreadsheet.

Second, we consider using FH to get another settler out earlier
Now let's take PC settled on T117 as in bc's plan, and have it receive the T126 WB from SM to be available T129 and have it build its own second WB. This means we start working the unimproved seafood, working the PM when built, and then the nets when improved. PC then completes its own WB on T134 and grows T135 to use it. It keeps growing through the lighthouse build (done T149), 2-whips the granary T150 and regrows to size 3 T155. This plan has ten fewer base on whatever build comes next, so we'll say that PC is 10+4*5=30 behind under this plan.

Meanwhile, CC built a settler for MC, and FH has been free to build a settler for BF, which plants T130. SM took its own second WB from T135, and chose whether to whip out another WB for BF to use on T138 (about 8 turns after it settled), or to slow-build it for BF to use on T145 (about 15 turns after it settled).

In the second run, BF got settled a bunch earlier, so even if it gets its WB only on T145, it's still got 15 turns on the unimproved fish to show for itself, i.e. (4-2)*15=301575 less about 4*15=60 in increased city maintenance. It also busted some more fog, preventing barb city, galley and unit spawn chances, and providing more warning of a raid on SM's fish. An earlier WB for BF whipped from SM would show a bigger profit for BF, but would hack SM down somewhat.

So considering just PC and BF from the point of view of the empire, I think the second plan is slightly ahead. The fact that CC's next settler can go somewhere else to get the next city faster seems like a clincher.

 

[mabraham]

One interesting thing about that last post is that I started it having reflected over dinner on the fact that the WB whipped out of FH is generating food elsewhere sooner that should offset the failure to Duckweed something onto the wonder. However unless someone can show that the net effect on the wonder is signficant, I think my numbers show that whipping the WB out of FH isn't best.

 

[bcool]

Okay I have 1 idea that addresses one of Ron's concerns.

If we have Bronzed Fish get the 1st workboat from Stone Mountain, then have Bronzed Fish whip out a workboat at the 1st opportunity and give it to Paired Clams...

I did a quick test (but didn't implement the working of the unimproved plains hill to get the 1st workboat out of Paired Clams faster that mabraham recommended )
Anyways, with the workboat from Bronzed Fish, Paired Clams was Pop 3 12/39 food in bin, 28/90 on a granary with a lighthouse both clams and the PMine worked on T150.

Gets Paired Clams up faster than leaving it to fend for itself. Bronzed Fish was slowed down after whipping its 1st workboat at pop 2 and then whipping a LH at its 2nd pop 2 it was up to 2 pop 6/36 food bin and 16/90 on a granary on T150.

 

[Ronnie1]

The discussion seems to have gone a bit adversarial between these two test game approaches. That's OK - a hot forge makes good steel - but I thought I'd make a friendly reminder that we're all trying to get the same objective

I apologize if my tone got adversarial...that was not my intention.

I think I just got a bit confused about the goals we were shooting for possibly. I have been working on the assumption that we wanted cities to be able start their own LH's out of the gate so we could whip those ASAP in order to gt to the second build ASAP and make the seafood resource = max value at the same time. In order to do that, the WB's need to come from somewhere else.

From what I see know...my earlier test where I plant cities 4/5/6 on T 120/129/138 seems even better, and it was NOT optimized as best it could be for a number of reasons.

I also ran another test last night that went the opposite direction a little. I settled Copper a tad later, but all 3 had better tile use sooner.

Maybe we just need to define a little better EXACTLY the short term goal we are trying to hit!

Also...here is the T108 test save altered slightly to allow English trade routes when we open borders. I just added a few land tiles to make a route.

 

[bcool]

I think expanding out to 8 cities instead of delivering an early workboat to Paired Clams is currently the most desirable path. We can confirm this with the numbers (and it looks like mabraham has already done so). The cost of expanding as fast as possible is currently still being offset by the additional trade route and ability to work even unimproved seafood.

I think it definitely is preferable to have the cities get workboats the turn they are settled, however given the choice of 1 more city or having a city start with a workboat it seems the 1 more city is better. And in my plan it does deliver Bronzed Fish and Crabby a workboat the turn it is settled. But Paired Clams, Marble Clams, and Stone Mountain are left to largely produce their own workboats. There are solutions to get at least 1 workboat to PC ...

I think my suggestion of giving SM's 1st workboat to BF and then having BF whip out a workboat for PC mitigates the cost of settling PC without its own initial workboat.
This sets up PC a little bit better than my test game shows and allows us to get either the Parthenon, Moai, or even the Colossus there if we decide that is best.

Earlier lighthouses also might improve my results but away from the game as I am I don't know by how much or exactly when/where we would whip lighthouses. I'm guessing Marble Clams and PC could have built LH faster, but I did whip it first thing in Bronzed Fish as I recall.

 

[adrianj]

Could we narrow down the discussion to the next 10-12 turns?

Importantly, I don't think there is much to improve or change in about the T108-T120 window, so perhaps we should lock something in soon.

If both Ron and bcool are playing these turns the same, then do we have a concensus?