Qsc18-Celts - Results and Strategy Discussion

[cracker]

Rabies, I'm not 100% sure what you refer to by "a lot of color coding" but some of the fields in the scoresheet are color coded a little bit to help with sorting and to try and avoid braincramps as we load that massive amount of info.

There are color codes tracking the versions and prior versions and I believe there are color codes on a few worksheets to help with group sorting. There are also a few names highlighted to help keep track of non-timeline games.

In the tech sections and sometimes in the productions section yo may see an outlined box to indicate current research and production targets.

We may use more highlighting in the future if it can be done without cluttering and biasing the results interpretation too much.

 

[rabies]

Thank you cracker. This explains a lot. I can imagine that this spreadsheet is a bear to deal with. Braincramps indeed. In restrospect, I guess there is not as much color as I first percieved.

One question if you can remember. I noticed the various internal notes in some of the columns. Some were obvious ('low explo'), mine was rather confusing, I believe 'monarc' was the note... I realize in the multitude of games you processed, mine was probably not particularly noteworthy, but if you could remember what this note referenced, I would appreciate it. I gained a TON of insight from this sheet.

 

[cracker]

I think that "monarc" may just be an artifact from the recycle process.

We begin the game analysis with the results from last month's event and then update thisng toe reflect the specifics of the game setup before we start adding new players and deleting particiipants taht may have passed on this game.

One thing we have not done alot of publicly yet is to slice across the data matrix in a different direction and to look at data on a player by player basis. There are a lot of really revealing and cool pieces of information when we do this cross game comparison and hopefully we can start to do more of this in the future. This cross game comparison seems to potentially have more value in the lower and middle thirds of the player pool because in the upper third we can see the more experienced players doings some radically different things each month as they pursue extreme implementations of different strategies.

There are some players that actually seemed to be following a preprogrammed recipr for civ3 game play that produced very similar game profiles in the QSC results for Qsc16, Qsc17, and Qsc18 even though the games reflect radically different conditions.

We can also do some more detailed slicing of specific game parameters and compare the results back to specific independent variables. An example would be to look at the correlation of total accumulated shield output versus the number of full speed worker equivalents.

Hidden inside all this data we have been accumulating is another whole level of discovery that we are only just beginning to appreciate (even if we do all sorts of shifting with the scoring formulas over time.)

 

[Ribannah]

Originally posted by SirPleb
It seems to me that after starting by moving the worker SW:
1) Moving the settler NW will bring five previously unknown tiles into the first city's eventual radius.
2) Leaving the settler at the start position brings six previously unknown tiles into its eventual radius.
3) Moving the settler S (the only remaining choice which makes the worker remain productive after its first move) brings seven previously unknown tiles within its eventual radius.
So I'm puzzled by the logic of moving NW - it seems to me to reduce the odds of a bonus at the same time as losing a move.

Option 3 I never considered since it was a move away from two bonus grassland tiles. I don't agree with 2 revealing six squares either. The partial view was already somewhat informative.
It was a close call though. May have been partly out of habit from Civ2 where the starting position would almost by definition be one square away from the optimal location.

Re it not amounting to much, I don't agree

The problem with fish is that it produces no shields. After 12 turns you can catch up and then pass the alternative in growth, but not with the production. You still have to actually build that Settler. Now IMHO the difference between 13 or 14 cities by 1000 BC is no greater than having one more Warrior early on (I could have found Greece early and stolen some Workers from them, too).

But if it were a cow, I'd gladly move 3 steps.

@Ribannah - I know this is low odds but just in case: do you happen to have more detailed notes of your QSC period you could publish? I'm trying to understand high aggression early starts better, it is an area I've never done well at. I've been trying to replay your opening sequence but I'm not getting something right (perhaps the archer or warrior path), things keep coming out differently in my attempts. It would be great if you wouldn't mind publishing a bit more detail, for this game if you still have it, or for a future high aggression start.

What I remember is that I purposely ignored the hut that the Archer passed on the way, and also that I waited a turn with either my first or my second attack on Carthage to see if workers and settlers were entering or approaching the city .
But I don't think my QSC is very instructive. It was the first time I ever tried anything like this. I saw the opportunity and took my chances, and could just as easily have failed (but still come out ahead with those initial slaves!). It was spectacular, but the potentially instructive part of my game came after the QSC period.

Edit: The warrior, as I recall, went due east first until it found a nice mountains/hills series to follow south.

 

[hbdragon88]

I placed second to last for my third qsc and mysteriously scored lower than in my qsc17. What did I do so different? For the third time I used Moonsinger's approch to start off the game, but only got four cities built; I had six or seven in qsc17.

About 2000 points in qsc16; 3000 in qsc17; only 2700 points in qsc18...

 

[civ_steve]

Great Work! What a grouping of game information the QSC is.

The city placement map is very interesting. I didn't realize the Lake Fish food advantage , and ended up placing city#2 near the cattle, which was too far away for quick development.

Also, the distribution of results seems a lot smoother than in past months. Prior QSC's had a small group of leaders, and then a large, smoothly-distributed, group of followers. This month's results are almost continuous, which indicates to me that the difference from one game to the next is just a choice or two.

 

[Bamspeedy]

Hb, you spotted some of your own mistakes in the last part of your timeline. I don't have the art files to load up vanilla civ3, so I can't look directly at your save, but...

By looking at your map and then reading your timeline, it looks like you sent just 1 warrior southwest to explore, then when another warrior found the volcanos to the north, you devoted much of your manpower to destroying the volcanos. You wasted, what 2-3 archers or so trying to kill them? Once you found out they had a better defense, you should have left them there until you had swordsmen, at least. Prior to 1000 B.C., it was unlikely that there would be something up there that was of so much importance that you would want to get a city all the way up there anyways.

Build more cheap warriors for exploring the maps and finding people, so A.) when other civs get mapmaking, your map is so valuable, you can cash in on it bigtime. (I sent 3 warriors exploring at the start and when another civ got mapmaking, I gave him my map +40 gold for mapmaking, then went and made everyone else broke by selling my map and mapmaking.) and B.) You can make contacts with civs yourself instead of paying for them. and C.) you have a better idea of where you want to expand. A typical build pattern is warrior-warrior-settler, or 3 warriors then a settler (how many warriors depends on shield output-you want a settler out the door when you hit size 3,or shortly after). Or 2-3 warriors, then a granary (granary in only 1-2 cities usually). Some warriors can go explore, and some stay in the city for military police.

Would you have been able to sell mysticism (that you got from a hut) to anyone?

Try to get more than 1 worker by 1000 B.C.

Gpt deals and techs. If you are going to buy techs, then either A.) buy them after several of the AI already knows the tech, or B.) try re-selling them to other AI and make some money back, or to get other techs. I usually don't like to pay for a tech that only 1 civ knows and nobody else has anything to offer, unless I'm just going for all-out fast tech pace.

Before 1000 B.C., I usually just like to build warriors only. Spearman are awfully costly and usually I only build them in an emergency, and archers I would only build if you plan on attacking somebody very early.

 

[Smirk]

Originally posted by SirPleb
I dunno, can't second guess others, but it does seem an odd move to me. Regarding the defensive consideration, moving to the hill would be superior.

You are thinking locally to the city. My original thoughts when I posted to the pregame thread had in mind cracker's warnings about intense barbs. While you would get a 50% bonus to unit defense, having a lake on one side of your city delays approach quite a bit. While there certainly isn't any specific defense bonus there are tactical defensive effects of a lake. This was what I had in mind when I originally considered the hill or nw for defense.

About not being very risky I agree, a lost move is not much if there's something to be gained. But the odds on gaining something are not maximized by a NW move.

I agree with you about judging, and I hope some of the eleven that moved there can elaborate on their thoughts, even if they aren't mathematically sensible. One other point is that it was known that there was a large land mass, this means less water so it seems highly likely that a fish would appear in a lake. This is pretty common actually. Of course I wouldn't move on a prayer, but this thought went thru my mind when I saw the lakes. As it turned out there were two fish nearby.

I've seen fish in lakes in other games and it seemed common enough, this map was obviously modified to have more lakes and I wonder if cracker purposely put fish in there or they were generated?

Re it not amounting to much, I don't agree - I'd move my settler two steps if it would mean a certainty of getting a food bonus. Food is the most important thing in the early game I think. I'd happily trade shields, gold, having to build an aqueduct in the long run, and throw in two settler moves, all just to get a food bonus

I'm not sure I share your entusiasm for food in the capital. I tend to prefer a high production capital. Of course had I known the fish was there I most likely would have moved, but for a fish thats questionable.

I would guess that one bonus food for the capital translates to at least two additional towns built by the 1000BC QSC cutoff;

True, but what have you lost? Its much harder to get a second or third city to the magic 5 shield production to churn out warriors [1]. In this game I might had ignored the fish simply because I knew I was going to try and get a lot of warriors built. If I were to replay this game I would still not move, and then settle and share the fish in my second and third city like you did. Meanwhile I would skip the granary in the capital and build an early barracks instead. And maybe get the fish cities with a granaries, again like you did. I only built one granary in my game.

Having just finished my gotm19 QSC I'm questioning my own strategy which is usually expand, expand, expand. Its good for milking and score, but power is far more important IMO.


Note 1
The key problem is getting a managable 4+ pop, while not too hard is certainly not something easy to do early. You'll need gold and slider for most cases, and this map was really weak on gold, and also you tend to waste a lot of gold using the slider in too high a percentage, especially when its only one city you are trying to balance.

 

[LeSphinx]

Congratulations for the GOTM/QSC team to bringing us the QSC#18 results so quickly!

Congratulations to the top score players and to all the players.

LeSphinx

 

[Borealis]

Smirk- I moved my settler for the reason you suspected, that there would likely be fish or something interesting in the larger lake. I still had 5 (IIRC) bonus grasslands in the eventual radius, and defensively it was good to have that lake to the left. I don't know if cracker sets up the maps deliberately to trigger this every time (he said he did in GOTM-Carthage), but I've found that I've moved the settler in every past GOTM/QSC I've participated in, unless there's a food bonus staring me in the face.

 

[ControlFreak]

Originally posted by LKendter
ControlFreak comes in ahead of me?
This is the second time someone in a training game beat the teacher. I wonder what this is telling me?

@Lee
Sorry I missed your post. I didn't mean to dis you. You are another one of the reasons I am playing at the level I am. It's no coincidence that my GOTM resulted in a Domination win modeled after our Dom win in LK33. You're a great player and teacher. I fully expect you to stomp me and all your trainees in next month's QSC.
[punch] [dance] {CF lays down the gauntlet.}

 

[Moonsinger]

Originally posted by hbdragon88
I placed second to last for my third qsc and mysteriously scored lower than in my qsc17. What did I do so different? For the third time I used Moonsinger's approch to start off the game, but only got four cities built; I had six or seven in qsc17.

I'm sorry. When it comes to QSC, Bamspeedy has always been the best in any type of map. If you want to take more risk, you can always follow Ribannah. If you follow my game, the 22nd place is the best I could get out of my QSC. So basically, Bamspeedy and Ribannah paths are good; Moonsinger path is not so good.

 

[mad-bax]

Originally posted by a space oddity
I almost fell of my chair seeing the results finding myself 22nd

It came as no surprise to me. Well done Space!

 

[OneFastWarrior]

Congrats to all the top players.

I was 79th in my 1st QSC. I wish I had the patience to have waited to read some timelines before I started gotm 19, but alas, I do not. I will be reading some of the top ones over the course of the month and hopefully I will do better in the following months. I know that I am getting better, as I have went back to some games that I started a month ago and found myself sayin "What was I Thinking?"

 

[zagnut]

Originally posted by Moonsinger


I'm sorry. When it comes to QSC, Bamspeedy has always been the best in any type of map. Moonsinger path is not so good

I think your start advice is good. (Warrior, Warrior; research Pottery; pre-build Granary; build Settlers. Then second city to build barracks and military units). It just seems to me that one additional factor has to be added - the sharing of resource squares between the first two cities. I have historically spread my cities and it has lead to low scores on the recent QSCs. You get a real production boost by micromanaging the resource squares between the first two cities.

I really did not know how to do this until I read SirPleb's comprehensive timeline for QSC 17. I heartily recommend his timeline to anyone who wants to know how it is done. He analyzed every move on every turn in great depth. He played on PTW but you can recreate the moves in vanilla civ.

 

[a space oddity]

@mad-bax: Thanks! You didn't do so bad yourself. Too bad Mad Swede and Raw Power didn't enter this month. It is good to see our tourny conscript class-of-season-4 do well here in the GOTM.

 

[ControlFreak]

Originally posted by zagnut

It just seems to me that one additional factor has to be added - the sharing of resource squares between the first two cities. I have historically spread my cities and it has lead to low scores on the recent QSCs. You get a real production boost by micromanaging the resource squares between the first two cities.

I really did not know how to do this until I read SirPleb's comprehensive timeline for QSC 17. I heartily recommend his timeline to anyone who wants to know how it is done. He analyzed every move on every turn in great depth. He played on PTW but you can recreate the moves in vanilla civ.

Yes, as I said at the beginning of the thread, sharing tiles, especially bonus ones is a huge advantage. That could be another factor in the success of the ICS style that Bamspeedy uses. I've never gone for the I Crammus award because I didn't want to manage all those cities but tile sharing can be accomplished with tight build patterns.

@Zagnut are you giving out more zagnuts? Like ...oh say... the top 3 QSC finishers.

 

[Moonsinger]

Originally posted by zagnut
I think your start advice is good. (Warrior, Warrior; research Pottery; pre-build Granary; build Settlers. Then second city to build barracks and military units). It just seems to me that one additional factor has to be added - the sharing of resource squares between the first two cities. I have historically spread my cities and it has lead to low scores on the recent QSCs. You get a real production boost by micromanaging the resource squares between the first two cities.

Thanks for your kind words. Overall, I think my strategy is good for winning the game (because if it isn't good, I would have changed it by now); however, if anyone want to try for a top QSC score, there are many other styles that would yield better result. My strategy has always been very conervative. I do not trade techs or resources unless I have something to gain. Since I'm not aiming for the fastest in spaceship, I like to keep a slow tech pace. I don't build a lot of defensive units. I like the horseman, knights, cavalry upgrade path. I'm a builder most of the time and waiting for the right monent to strike. That's pretty much how I handle myself in real life too. I spend most of my time studying and preparing myself for that one opportunity. The opportunity may never come, but when it comes, I'm going to be ready for it.

 

[zagnut]

Sorry, Control Freak, no Zagnut bars for QSC. Only for the three people that score just above me in the GOTM results.

After all, I wouldn't want to dilute the value of these wonderful prizes.

 

[ControlFreak]

Originally posted by zagnut
After all, I wouldn't want to dilute the value of these wonderful prizes.

I completely agree. I was just hungry.