I just realised why Tauron is so bad. He's got a horde of hellhounds roaming his island, and thanks to lizardmen, no boats. That's a reason why lizardmen are so bad with chaos maybe? But, my aggressive strategy has prevented him from summoning them on the continent, so that might just be good play from my part.
|
Test games played
|
|
Right I'm aware that different settings and races can make a difference. But generally, they should (in my opinion) be less of a difference tan realms and retorts. For instance if you choose all the wrong settings and race for your realm/retort, I would expect you to play one difficulty worse, at worst. Similarly, choosing the right race and the best settings, should only allow you to move up one difficulty at most. For instance, my build without barbarian (in fact with any race) should still let me play at master level.
So if we assume a decent but not perfect race, and decent but not oerrfect settings, with the knowledge that settings + race MIGHT cause a fluctuation of one difficulty I think we'd be fine.
There are pairings where you probably drop more than one level. Playing Gnolls without a realm that can deal with flying units - for example doing all Life books - is a massive disadvantage.
Likewise, playing all Chaos or even worse all Sorcery with a race that has no early game units (or no good units overall) is very bad. Of course this only matters if you had to build settlers of your primary race. Finding a good neutral can completely negate this part. Turning power to Max when playing noding retorts, race, realms and Sprites probably matters more than a single level of difficulty too. Lizardmen on Tiny can be a massive advantage too. Rich with Dwarves...even more if you can transmute to turn everything into gold and crystal...and adamant... I'd say the race and map settings are +1/-1 difficulty in most cases, but there are exceptions where it might be two or more. Although all of this is just guessing. There is no way a human can try all the possible combinations in their lifetime or even 10% of them.
September 25th, 2017, 16:39
(This post was last modified: September 25th, 2017, 16:40 by Nelphine.)
Oh sure but that escapes the point. So, let's say this fair land, climate, minerals, good (high? Whatever 1.5 power is) power.
Now based on your previous you think 10/1/1 no retorts is weak. So this should win normal 100% if you are experienced, advanced 90%, expert 75%, master 40%, lunatic 20. Now are there any races that you think would preclude that? You mentioned mono life gnolls, but really, they should be able to conquer cities, whether neutral or AI, pretty early to overcome that. I don't think there is a rave by itself that would drop normal below 100%. However I do think right now you could lose a game on advanced, so 90% might work? I dont know, i probably wouldnt do that well. I don't think you could ever win lunatic. Master.. I don't it would be above 20% but I could be wrong. Which means expert is probably closer to 50 or 60%. So let's use a mono realm with 1 book from each of 2 other realms as our base. From there, we figure out: at what point does adding retorts or focusing on other realms make it stronger? Should 6/6 be comparable to 10/1/1? Or is it better? OK things I'd like you to put numbers on for how often they should win each difficulty, so we can see how the difficulty actually is. If you need to, pick the second best race you can think of for each. 10/1/1 10 myrran 6/6 5/5 myrran 5/5 +2 excellent synergy choices (don't need to specify, just magically assume they're wonderful). Once we have numbers for these, then pick a build you think you personally would like to play at lunatic and give numbers for that. Then give numbers for my impossible build. Once we have numbers for all that we should be able to guesstimate numbers for most things and see how far off some things are.
Gnolls with life....once you committed to the strategy, built several gnoll cities, produced a bunch of buffed Jackal Riders and are ready to go to war, someone with many flying units appears and then you lose because you have no way to attack them and no spells to deal direct damage. So no, it wouldn't be 100% even on normal, maybe 90 but not 100.
It's not like there are no ways around it, but there are no "guaranteed" ones. If the first enemy uses the flying units, you're stack. There is an escape route of hiring a ranged/thrown/flying hero but that's up to pure chance. (although the AI being less aggressive on attacks might help you there - you can't conquer anything but they can't attack you either, gnolls have high military power...) Quote:10/1/1 These all fall into the "Average" category above. Quote:5/5 +2 excellent synergy choices This would be "Good". Squeeze in one more retort and it should be "Top" tho there probably are realm combinations where 2 retorts would better. To play on Lunatic, let me see... A. 5 Death 4 Sorcery, Alchemy, Archmage , Chaneller and...haven't thought about a race for this one yet. B. 4 Chaos 4 Life, Runemaster, Archmage, High Men C. 4 Nature 5 Sorcery Astrolger Archmage Sage Master, max power, High Elf Note I haven't actually tried playing either of these yet on Lunatic. I only remember two games I played and won there - one was 5 Chaos, 5 Sorcery, chaneller and I won it through lucky diplomacy and the other was, Guardian Warlord Myrran Dwarves with 4 Life 2 Nature I think? Wouldn't play this one again, too few books... Your impossible build is clearly in the "top" category. In fact it performs better than it should be but we already know it's due to the AI not attacking berserkers. Maybe a new feature that allows the AI to perform specialized attacks based on stack composition against human player targets only could help this but I'm not doing that until we are at least considering the current version stable. ...anyway, sleeping now, and won't be online for a day so you have plenty of time to think about it.
You guys are IMHO approaching this from a too complicated direction, it's simply impossible to take in consideration every variable in this game, which is why it's so good. Just let the player assess the difficulty of a random setup based on his experience. In this case, decent books but no retorts should have been a 50%. However.... Take a look at this:
![[Image: LRk1adP.png?1]](https://i.imgur.com/LRk1adP.png?1) This is in expert, not master. I laughed when I saw this, Tauron has more than 50 hell hounds, his island is now hell island, and keeps corrupting one or two tiles per turn: that's what, 70-80 mana per turn?... He's got only the capital - lizardmen, without even the shrine - and no nodes. He has equal to my power production on the graph. I would have had 0% of victory if he weren't constrained on that island. Now all those hell hounds are becoming mine mbwuhaha go ghouls go! :D I'll send them to self destruct against lairs and nodes for great justice, and win. The game AI is now good enough to deal with the player without so much resource cheating. It should definitely be toned down. We can then start nerfing the lunatic level strategies, so that lunatic returns as difficult as it is now, and other strategies become feasible at lunatic as well. We could even reduce the building razing ratio proportionally with the AI cheating reduction: my troops razed a good neutral city almost completely, I'm still dubious if there's a difference between AI and neutral cities for razing. This - the last paragraph - would be a good objective for this testing activity. Do you agree?
The problem is two fold: one, how much do we need to reduce the cheating to make seravy's numbers become reality? Two, lunatic strategies are usually based on a core that is very hard to fix. Draconian bowmen for example. We've discussed removing them before, I still think its a good idea, but its purely a balance decision, and seravy is right to doubt the removal of something purely on balance. Bezerkers are due to strategic combat - which we've proven doesn't have the space to be solved (seravys idea for composition won't change my game much, except to require me to build some other units in cities that otherwise literally never build units, which seems to be something the game is designed for anyway). Ghouls/sprites are a core part of their realms, and only the most practiced players can use them at lunatic.
So I think the first thing is to actually determine where the cheating for advanced and expert should be. Lunatic is an afterthought. Master is, to me, what we should really be aiming at as the ultimate goal. Right now my suspicion is that only the lunatic strategies have a good chance at master, and they of course destroy master.
September 27th, 2017, 05:23
(This post was last modified: September 27th, 2017, 05:24 by Seravy.)
Now that we defined some goals might as well do things in order.
...or not, I'm skipping easy for now. Normal, 3 N, 3 S, 2 C 2 D, 1 L, Halfling. so far I'm losing. Admittedly I'm not playing very well - probably better than what a player on normal can do but I'm not trying too hard. There was a wizard near me - I took two of their small cities, as well as a neutral, then gathered 9 bears and attacked the capital. The bears lost 3 nagas 4 ghouls and some crap was enough to beat them, I didn't expect that. Next I sent 7 slingers with flame blade, 1 bear and a hero - those lost again. Meanwhile a random wizard's doomstack came and razed two of the 3 cities I got. Now it's 1406 and I still only have 2 cities. I might need to come up with a real plan to be able to win. ...plan C 9 ghouls of my own, all 9 died before killing a single defending ghoul. eh. Might as well not use the second stack... Time to try again after a morning cofee and this time not take it easy. It seems playing Normal difficulty is serious business now...
Advanced difficulty is easy but aside from that, the wizard plays Dwarves who grow very slowly, and didn't get Stream of Life. It also started on an island and didn't have a spell to leave the island. No floating islands or water walking (3 books are needed in nature or sorcery to pick those). It did have Sprites doomstacks but those aren't very effective against the stronger Myrran nodes and neutrals.
|