After looking at the balance report, I'm going to take the save and make a couple more tweaks.
|
[SPOILERS - NO PLAYERS] The Final Clusterfuck: The PB38 Map and Lurker Thread
|
|
February 5th, 2018, 19:09
(This post was last modified: February 5th, 2018, 19:10 by Cornflakes.)
Latest iteration attached ... Missed gems at one start and had two more outliers.
Note: I removed gold entirely from the mainland and added many copies of gold to the Astro islands. I attached the full .txt balance report. Here is the output from the balance report on the latest map (sorted first by the luxury column, and then by number of land tiles): Looking at the luxury column, there are no clear outliers. Looking at the higher-rated compared to the lower-rated I think I'm satisfied with the luxuries. T could maybe get another duplicate or two added that they could trade away. Each player has guaranteed access to 5 luxuries (gems + [starting tile] + [silver OR furs] + 2 more). Some have contested but reasonable access to a 6th. Take a look and let me know what you think.
Wow, that looks like a huge improvement! I'll have a closer look tonight, but it looks like we're really on track to get this thing out the door by the weekend!
If somebody pops gold from a mine, how big of a game changer will that be? I think it should be okay, since everyone with astro access will get gold eventually anyway, but it should at least be considered since 25 players and lots of hills make it a lot more likely than it otherwise might be.
February 6th, 2018, 06:13
(This post was last modified: February 6th, 2018, 06:28 by RefSteel.)
Yeesh, I keep forgetting in between editing sessions how my machine takes 3-5 seconds per click for adjusting resources or water/land on this map. If it would let me queue clicks up without skipping half of them, I'd just do everything I want to on a screen and then while it slowly gets executed, I'd go read a book. Anyway!
I moved/added/changed lots of marble and stone more or less as described, with ~roughly equal numbers of each rather than concentrating on stone. I also added about 10% of the mainland rivers I wanted to add, and made a couple smallish adjustments to islands and food. (Not in that order.) I tried to help start W in particular since it seemed to need it the most, but I doubt if I did enough for it; waiting on the balance report for more information on that, and on what else needs fixing. Just looking at the map from on high with strategics only turned on, I feel like I've added way too much iron over the course of my editing passes (e.g. replacing copper) but I don't know if that means I should replace half of it with other stuff, remove a bunch of it altogether, leave it and figure it'll be fine, or what. Probably part of the answer is "If a start looks generally too strong, this might be something to take away from it." Dunno. I do feel like we're very close, and my computer can handle most of the aesthetic/global-warming stuff just fine (e.g. trees for lakes and mountains, sea ice, etc.) - or at least way better than what I did on this pass. (I'll keep the number of deep and distant whales and fish low for several reasons, including this one.) (February 6th, 2018, 06:13)RefSteel Wrote: Yeesh, I keep forgetting in between editing sessions how my machine takes 3-5 seconds per click for adjusting resources or water/land on this map.  oh my!Quote:I also added about 10% of the mainland rivers I wanted to add Where are you planning to add more rivers? Quick math on the balance report give about 700 riverside tiles now. Quote:Just looking at the map from on high with strategics only turned on, I feel like I've added way too much iron over the course of my editing passes (e.g. replacing copper) but I don't know if that means I should replace half of it with other stuff, remove a bunch of it altogether, leave it and figure it'll be fine, or what. Probably part of the answer is "If a start looks generally too strong, this might be something to take away from it." Dunno. Opening the map with a text editor I did a count of the iron ... 83 total ... about 10 of those are on astro islands ... leaving just under 3 per player. That doesn't sound too unreasonable. The heaviest concentrations of Iron are also in the northwest (more cramped for land area) and desert bands (poor terrain anyway). It won't be revealed until the point where an extra couple of hammers per turn is super relevant so I'd just leave it.
Here are the results from novice's script without water tiles and islands:
https://pastebin.com/U80W6BqU In the results below, I manually calculated uncontested seafoods to get a slightly better picture: Starts that have 1 stdev higher land quality score than average: Start R (158 usable mainland tiles, 21 food resources, though quite much jungle) Start E (192 usable mainland tiles, 22 food resources, though very much jungle) Start Q (158 usable mainland tiles, 24 food resources, though 5 neighbours) Start H (174 usable mainland tiles, 26 food resources, though 5 neighbours) Comparing these results with GJ's balance tool's, I think it's clear that start E doesn't need to be nerfed and start R isn't overpowered either. That's why I'd say that there are no clearly overpowered starts. Starts that have 1 stdev lower land quality score than average: Start D (120 usable mainland tiles, 16 food resources) Start X (124 usable mainland tiles, 15 food resources) Start V (122 usable mainland tiles, 18 food resources) Start B (131 usable mainland tiles, 15 food resources) I think that it's interesting that GJ's tool doesn't identify these spots are particularly bad. Instead it shows A, W and Y as the worst ones, so I'm not really sure about the conclusions to be made. However, I think it's still pretty obvious that these spots (and W) aren't as strong as some others, so I'd think about making the two tails of the snake slightly better by adding more land and/or food. If adding land is impossible, I'd at least seriously consider giving them some extra food. I'm not really sure either what to think of starts A and Y (the only starts with only 2 neighbours). Novice's tool doesn't identify them as particularly bad (average land quality scores, A has 134 usable mainland tiles, Y has 142 usable mainland tiles). However, they clearly have weaker astro island access than most other starts which has to be taken into account. I'm not sure either if having 2 neighbours is that much (or any) better than having 3, because the options for conquest are much more limited (see JR4 in PB37). In other words, there is an argument to be made for slightly buffing these starting positions as well. Regarding luxuries, I think that start O is missing one classical happiness resource, as I see only 4 luxuries (ivory, silver, spices and incense). I see no other alarming problems with luxuries. Regarding food and health resources, it's not that important as health isn't that big of an issue, but I'd do a few changes anyway: Start U has 6 (other starts have at least 8) different health resources, I'd turn at least one of the pigs into a cow and some other pig into a corn/rice to give start U one granary resource more. Start C has 7 different health resources, I'd turn one deer into a pig/cow. Start I has 7 different health resources, I'd turn one crab into a fish/clam Regarding strategics, I'd consider moving the aluminiums around starts O and M, there are 3 aluminiums on the border but neither of them has a backline aluminium. I don't know if that's a problem or not though. Starts X and V have BFC aluminium, whereas no other starts has BFC metals. I don't see this as a problem, but do you see some kind of a problem with it? I didn't spot other problems with the strategics. Edit: I missed that you had buffed start W after Cornflakes posted the balance report, and this time novice's tool didn't identify it as considerably weaker than the rest, so I'm not sure it needs to be changed anymore.
February 6th, 2018, 12:47
(This post was last modified: February 7th, 2018, 10:13 by Cornflakes.)
To further expand on my rivers comment ... with about 700 riverside tiles now, that is between 15-20% of the main continent land tiles. That sounds about right for river coverage. The balance checker indicated a variance of about 20 riverside tiles on the low end to about 35 riverside tiles on the high end. Maybe extend some of the rivers at the lower end of the spectrum. If each start has 30-35 riverside tiles that would be right about 20% (or 1 in 5 tiles).
The astro islands do need a few rivers added ... but those can be drawn mostly at random with little forethought. I made a debug that I can look over at the office. Looking at the rivers there are 2 that I noticed (A & J) that make a "U" with both ends touching the coast. That is an un-natural occurrence that never generates under normal scripts. "A" should be easy enough to split into 2 rivers towards the south away from the BFC. "J" looks a little odd , but the below layout give 33 riverside tiles. (February 6th, 2018, 11:48)Erwin Wrote: Regarding luxuries, I think that start O is missing one classical happiness resource, as I see only 4 luxuries (ivory, silver, spices and incense). I see no other alarming problems with luxuries. Thanks for catching that! O is actually missing Gems which all other starts have. There are multiple Gems to the east of O that we can grab one and move closer. The Dye in the middle (currently closer to K) can be moved 2-3 tiles north/NW where it can be more contested by O. K already has 2 calendar sources and all three ancients (fur, silver, gold, incense, wine) so they don't need another guaranteed source from Dye. Quote:Regarding food and health resources, it's not that important as health isn't that big of an issue, but I'd do a few changes anyway: Hadn't considered health yet, thanks! Quote:Regarding strategics, I'd consider moving the aluminiums around starts O and M, there are 3 aluminiums on the border but neither of them has a backline aluminium. I don't know if that's a problem or not though. Starts X and V have BFC aluminium, whereas no other starts has BFC metals. I don't see this as a problem, but do you see some kind of a problem with it? I didn't spot other problems with the strategics. There's a desert hill north-ish of both O & M that would work nicely for an extra Aluminium. I added those 2 BFC aluminum at X and V, those 2 starts were clearly lacking it so I just stuck them there since it was easy. Aluminium is so late that it shouldn't matter.
February 6th, 2018, 13:20
(This post was last modified: February 6th, 2018, 13:33 by RefSteel.)
I'm glad to hear that the iron (and most of the land in general according to Novice's tool) isn't as bad as I feared. I'll not mess with the iron any further and I'll panic less about start disparity, although I'm still worried about starts W (135 land tiles per GJ's tool) and H (5 neighbors and I think insufficient compensation).
Good question about the rivers, Cornflakes. I wanted to add more just so that there are more rivers that don't pass through players' BFCs, but glancing at the balance report suggests that the starts with the fewest river tiles are not currently among the weakest overall, so with one or two exceptions, I'm not sure where to add them. Probably I'd just add another short river somewhere near H, but I'll assemble a little more information and look at it again later today. (February 6th, 2018, 12:47)Cornflakes Wrote: The astro islands do need a few rivers added ... but those can be drawn mostly at random with little forethought. Yeah, I haven't touched the astro islands at all yet except to sink two tiles to prevent galley bridges. Rivers aren't random though; they should flow from highland to the sea (or to lakes, but the sea is better). Quote:Looking at the rivers there are 2 that I noticed (A & J) that make a "U" with both ends touching the coast. That is an un-natural occurrence that never generates under normal scripts. The trouble with the debug is that it doesn't show the direction of river flow. Both of these "U" rivers start by a mountain and have an early fork, then both forks flow down to the sea. It's possible that normal mapscripts don't do that, but I'd rather have these rivers than a normal mapscript's. |