Playing the AI in Daily Challenge Games
This was originally a reply to: https://forums.offworldgame.com/490448/i-need-help-underground-nuke
But I found it useful even for myself to learn from my own replays. I hope this helps Soren sell a whole lot more copies of this game, and I hope that makes the Daily Challenges extremely popular. I always enjoy ichbinenkonto's replays (when the game's Replay system doesn't break down and lose his replays), but I hope for more people with more creative strategies to share!
Some day, maybe Soren will work with me to create a proper tutorial --- or even a textbook! --- for this game. The economics concepts, even certain logarithmic arithmetics, are not quite amenable for explanation in quick tutorials. (My explanations below require almost zero math ability to grasp.)
--- Original Post follows ---
So, let's analyze my Daily Challenge replay on 2018-08-16. There are too many concepts I wrote down for myself, could even be a complete book named "Offworld Economics and Takeovers"! Much easier for me to just analyze actual gameplay in order to explain the concepts I learned about this game.
HQ Resources First
As explained with the Price Mechanics, buying HQ input resources will quickly raise their prices and delay your HQ upgrade drastically. (To illustrate, consider that producing 360 Si, aka Silicon, will take you 30s whereas buying from $1 would take $13K that has to be earned in 500s. You can see this in midgame in my replay, when Si and Al are at $1 thereabouts.)
Colony consumption: Po 2.80. Wa 0.64. Fo 0.96. Ox 0.72. Fu 1.20. Ir 2.40. St -1.20. Si 1.60. Gl -0.80. El 2.00.
The strategy is to wait and see what HQs the AI will plonk down.
I wanted the Si and Ba (Basalt Lakes) just south of the Colony. The Colony eats Si (to make Gl), so Si would be a great AR/CR in early game. Also, I'm willing to fly a bit southwest for Al, since I reckon I'll be making Fu. But alas, P3 takes that as Scientific at -$38K in debt.
Elite needs St (secondary) and Si (primary). Basalt lakes do not produce Si fast enough. St is intrinsically slow to make (0.5 per turn compared with 1.0 for Si/Al quarry). I don't see any Si/Ir cluster for Elite. That leaves me with Nomadic, which needs Al (secondary) and Si (primary).
There's a ton of Si just east of the Colony, but P2 takes that as Nomadic at -$28K in debt. I thought all was lost, and I had to start with hefty debt to secure the best spots (don't bother going zero debt in lousy spots, won't work). But wait, there's a similar cluster of Si south-southeast of the Colony. So, Si is secured.
There's only 1 spot we should take for Al. It has Al_vh (very high) right beside Ox (oxygen) quadrio (a term I coined, as opposed to trio). Note that trio has the best adjacency. See the end of this post for an explanation of Adjacency Effects. (We will need Ox to feed Wa production. Again, recall that we really should not buy inputs for our production because the Price Mechanics will quickly reduce profitability).
So, we settle our Nomadic HQ on those 2 spots.
I put down 3 Si quarries and 1 Al mine. Why? Firstly, recall the trio. Secondly, notice how none of the other players have Al now. That is why I dumped everything (but Al, Si, El) to buy up Al. I want to keep Al under-supplied for now, which will further accentuated the under-supply for the other players.
The Si builds up quickly, and I go level 2 at S1:16:30 (Sol 1, 16:30 hours).
Still on Primary and Secondary Inputs
I complete the Al trio, and create a Si quadrio. We need a bit more Si than Al for HQ upgrades.
Chase Rising Prices. El is being consumed by the Colony at 2.00 per turn (anything above 1.5 is hefty), plus P2 would be buying it up for his own upgrade. Since we know that El will rise every turn, we sell Al and Si to buy up El asap.
Once I have 40 El, I stop selling Si and Al.
Paradox of the Expensive Tertiary: Notice how quickly I could buy 40 El? Paradoxically, tertiary inputs seem expensive but constitute the lowest cost in HQ upgrades.
I reach level 3 at S1:22:50.
Still More Primary and Secondary Possible
We're lucky that we have a Si/Al cluster. We can make Si and Al even faster for our next upgrade. I build 3 more Si and 1 more Al.
Repeat the previous tactic for our next upgrade, except that we don't sell Si at all. Why?
Again, the Price Mechanics is at play here. Sellers who sell at low prices lose more (stockpile) than they gain (in cash); buyers gain more (stockpile) than they lose (in cash). The Price Mechanics means that buyers can buy large quantities while increasing the price by only a little; conversely, it takes less quantity bought to move a high price by $1.
I stop buying El when I have 60 El in stockpile. Now, I wait for the Al and Si production to push me to level 4.
Free Production Facilities First
We should always fill tiles with some free production facilities (mines, quarries, water pumps) rather than leave them empty. Yes, I know my Si production is overkill; 762 at S1:30:00 when my Al is just 199. Don't worry. We will be replacing our Si quarries with more valuable production later on. But for now, we don't have anything to sell for cash in order to build anything at all (eg Ion Collector). This is where your problem may have been; you ignored the value created by free production facilities while paying up for advanced production facilities that need time to break even.
But wait, we have a lucky break! At S1:30:00 everyone is given 1 free Claim. With the over-abundance of Si, I build 1 So (Solar Panel). Note how Po is rising fast? The Colony eats a hefty amount of Po (2.80). St is the biggest culprit, eating 0.3 Po per turn; P3 and P4 are making St in earnest. Ba (Basalt Platforms) eat 0.2, and so do water pumps (yeah, not all that "free", this one).
I reach level 4 at S1:31:00.
Here, we move into income production, no longer just HQ inputs production. You'll notice it's never a bad idea to over-produce the HQ inputs, since we'll be needing truck loads of them to upgrade to level 5.
I complete the So trio. That extra Si sure came in handy, right? I was at this point hoping I can use Po for income. But from experience, I know that Po will be crashed soon. The paradox is that the most profitable resource is often the most quickly crashed. This is possibly why you found it unprofitable to chase after the most profitable resources.
I'm actually out of cash at this point. With everyone over-producing Al (P2, P3, and P4 have Al_vh) and Si, no one could get much cash from Al or Si. But should I have built Ion Collectors and bought Al and Si? No. Again, recall the Price Mechanics. If I had used the fantastic CR from Fu (fuel) to buy up Al and Si, the other players would suddenly find their Al and Si (free mines and quarries) immensely profitable. I would then have enriched them by building Ion Collectors too early (Ca was expensive earlier on, compared to quarries and mines that are free).
Fortunately, P3 was impatient and upgraded to level 3 at this point, raising the price of Al. Notice the green highlight on Al price, indicating a sudden rise in price. I managed to dump Al at $24. I also dumped Si because I needed more cash for Ion Collectors.
At this point, I figured that everyone will be selling Si and Al (we all over-produce it now). So I auto-sold Al and Si to by up Ca, so that I can have my first Ion Collector asap.
Replace Poor Returns Facilities
I replaced 1 Si with my first Ion Collector instead of building the Ion Collector on a new tile. Why? Po price is high now. Rather than have 1 Si quarry eating 0.1 Po and 1 Ion Collector eating 0.3 Po, I will have just 1 Ion Collector eating 0.3 Po. Moreoever, with Si price at $1 now and a global oversupply, my Si tiles are not worth the Po cost they incur. But whoa, don't delete all Si production now! Recall that we don't want to buy our HQ input resources. Even with Si at $1, it will still take $5.1K to buy for our next upgrade.
So, with our first Ion Collector, we have formally entered into our Income Production phase.
Wait Out High Prices
The income from our first Ion Collector quickly gave us enough cash to start buying up Ca for our 2nd Ion Collector. Why didn't I buy up Ca right now?
Scroll through the map and see that my purchase of Ca had prompted the others to commit to making Ca. See their inbound freighters carrying Ca? Expect Ca to be sold down soon. P2 and P3 already have Ion Collector.
I build the 2nd and 3rd Ion Collectors when Ca price comes down from $53 to $40+. With 3 Ion Collectors, I'm paying the Po cost for 3 while reaping the earnings for 5.25. See Adjacency Effects below at the end of this post.
While waiting for Ca price to come down, I had enough cash to build an Ox trio. See beside Al.
Anticipating Demand (Market Movements)
See how P2 responded to the Po high profitability by over-committing 4 Nu (Nuclear Plants)? Wa price will go up. Wa production needs Ox, but Ox in this map is produced at half rate (solar facilities operate about 50% of the day). I created an Ox trio to start stockpiling Ox for Wa production.
At this point, I was tempted to replace my 4 Si production with Wa. But hold up, recall the Price Mechanics again. I don't want to buy HQ input resources, especially when I have to buy over 300 in quantity!
Within 10-20s, I have the 480 Si and 320 Al needed for level 5. The income from Fu is more than fast enough to chase the rising El price, so I didn't buy up El right away. (This is not an APM, aka Actions Per Minute, game ala Starcraft!). Moreover, I don't want to prematurely raise El price for P2 to dump; say P2 wants to quickly build a 2nd Ion Collector while delaying level 4 upgrade.
Draining Opponents' Cash
Notice that I don't sell any Ox at all? Firstly, I am stockpiling for Wa production. Secondly, P4 is making Gl, and must spend cash (not debt) to buy Ox to keep Glass Kilns online.
I reach level 5 at S2:20:00.
Complex Goods Need Complex Inputs
Time to profit-take from that high Wa price that P2 had been jacking up. I build 2 Wa right away, given that Ca price has been sold down below $40 (was $53). Water Processors need only 10 Ca each, anyway.
Why do I build only 2 Wa? Firstly, I need 1 tile for Hack (Hacker Array), 1 for Opt (Optimization Center). That leaves me with only 1 tile! But more importantly, I am producing only 4.72 Fu per turn, and 2 Wa will eat 2.0 of that. Since Fu profitability is still high, I don't want to cripple my Fu sales volume.
Again, I'm sure you know by now why I don't want to buy Fu to feed Wa production. Price Mechanics! Yeah, I'm sure you get the picture by now.
Dampening Opponent's Earnings
Another important reason I don't want to reduce my Fu sales volume much at all is because P3 is right now building a 2nd Ion Collector. Given adjacency effects, P3 will earn 3 (selling Fu) at the Po cost for 2.
I need to sell down Fu to the best of my ability now. In doing so, I also get cash which expands my options.
Primary Resource Heavily Needed for Most Buildings
I built the 2 Wa on new tiles, instead of replacing the Si tiles, and I'm already at HQ level 5. Why don't I replace the Si tiles since they have poor returns? I will still need quite a lot of Si to build Lab, Opt, Hack, etc. The amount of Si needed could easily run above 300, and I really don't want to buy that much Si even at $1. (Price Mechanics again! Yeah, you get the picture by now after I've repeated it to death, I'm sure. )
Insufficient Production Means No Income
Why don't I sell Wa? Isn't P3 already selling it, with P2 and P4 soon coming along too? No. Let's do a manual calculation.
P2 will have 1 Wa that eats 1.0 Fu per turn, which means P2 will fully take on his existing 1.73 Fu cost for freighters (was offset to 0.73 before). Also, P2 will only be able to sell 0.6 Wa per turn, after his HQ eats 0.4.
P3 is selling only 0.4 Wa per turn, and 1.13 Fu too (after subtracting his Fu cost for freighters).
P4 will also sell 0.4 Wa per turn, but eats 1.85 Fu per turn. His 1 Wa will nullify his 1 Fu production, as is also the case we discussed for P2.
Total Fu demand will be 3.58 (5.38 if adding Colony's). P3 and myself are selling 2.46. You will see Fu price still rising from here on.
Total Wa demand will be 1.60 (2.4 if adding Colony's). P2, P3 and P4 will sell 1.4 (I'm not selling). The price will still appreciate for my Wa stockpile.
What does all that mean? I want to stockpile Wa because I know the price will generally keep moving up.
When I first started playing this game, I used to be afraid of stockpiling resources when I see other players make these resources. Now you know how wrong that is!
Auction Prices By Standards
Anything below $10K is a steal. I see that P2 is about to get the Network Virus at $5K. I bid at $6K, and won it!
Let it go if the price goes above $10K. But there are more complex considerations depending on what you aim to do to your opponents. This short game didn't let me illustrate those considerations. Maybe next time. In short, the considerations have to do with CR and endgame. See my previous replies in this thread regarding when to go AR or CR.
Black Market Effects, Know Them
I buy up El, not worried that P2 will dump it. I had just used Network Virus on P2's 3 Nuclear Plants, forcing him to spend cash to buy Wa (P2 still has cash, won't take this Wa cost from debt). Even if P2 dumps El now, it will just be whisked away by Wa needs.
As an added benefit of my proper use of a Black Market item, Wa demand just increased by 1.5. This means my Wa stockpile value will appreciate even more!
Hacker First, High Demand Means Fast Game
The first thing I build is a Hack rather than an Opt. Why? (Not that it matters much in this game; my income is fast enough to build both almost at the same time.)
If you're playing Daily Challenge games, you're going for timing. The way top players (like ichbineinkonto) hit the best timing is to increase demand (artificially via Hacker Arrays) for goods.
I build an Opt soon after. The Fu trio income sure is coming along nicely.
Queueing Early (Enough) For Planned Future
I start buying up Ch (Chemicals) for my Opt to use. Why? P3 has a Lab. He'll be buying up Ch for sure.
But queuing too early for a planned future that is too far away? Bad. Strike a balance such that you queue early enough, but not too early.
Stop Exponential Growth
Opponents get exponentially more difficult to buy out they stronger they get. If you hit them early, they stay weak with limited options.
I dynamited P2's Pleasure Dome. El is expensive. He'll need half the resource (half of full price) to rebuild his Pleasure Dome. This is yet another way of draining your oppoonents' cash.
Moreover, his Pleasure Dome is the highest earning tile in the map now.
Revamping for Max Income
This is actually the late game phase. With the sizable income I have now, I revamp my HQ to maximize income so that I'm ready for the kill soon. (I better kill soon too, before every opponent locks out with 5 shares.)
In other situations, the late game may be much earlier, say at HQ level 3 for a 2-player game. The fewer the number of players, the lesser the demand and income rate possible, the more expensive level 4 and 5 HQ upgrades will be relatively.
HQ Powers, Know Them
Notice that I'm using Nomadic's "Return Claim". I get 5 at HQ level 5. I get to release 5 tiles and claim 5 new ones.
No other faction is so adept at HQ revamps as the Nomadic is!
The 3 So must go, of course; Po price is $2. I release the 3 So tiles.
Pick Your Targets
P4 can lock out easily (Elite). We will aim to drain his cash. Don't buy P4 first, or he might use spare cash to lock out, spare cash that was originally intended for HQ infrastructure investments.
P3 has the most income now, selling a small amount of Wa and almost as much Fu as I am. (This was calculated above in section "Insufficient Production Means No Income".)
P2's cash is still being drained by 2 Nuclear Plants, forcibly kept online by my Network Virus.
So, we hit P3, then P2, and lastly P4.
Draining Opponent's Cash, Case Study
P4 is now building 2 Fo, but has only 1 Wa. That means P4 will spend cash to buy 1.0 Wa per turn to keep the farms online.
Our main income will be Wa. We will hack up Wa price to further drain P4's cash, and we will build more Wa production to earn more quickly. We will also optimize Wa production, which means more Wa production given the same number of tiles we have now.
(The case is different for a 6-8 player game, where there is enough goods demand to give me a fast enough income to buy out, rather than merely majority buy, an opponent. Buying out an opponent gives me 2 free claims to further increase my production and income. For a 4-player game, an Opt is almost always necessary.)
Dampening Opponent's Earnings, Case Study
Should I build Wa production now? Not yet. First, I need to ensure that P4's farms don't earn too much cash. That's cash he can use to lock out! So I build lots of Fo to sell down Fo price.
Yes, that will seem counter-intuitive to newbies especially if they simply focus on "highest returns now". A newbie would probably build for more Fu income. But recall from our calculations before that P3 and myself are already currently selling over 2.0 Fu per turn. Although Fu demand is still under-served at this point, building more Fu will crash Fu price and instantly cancel Fu profitability.
I build 4 Fo over my 4 Si, and a 5th Fo for even more adjacency. I released 1 Al tile for that 5th Fo. I then further build 2 more Fo over my remaining 2 Al tiles; at this point, I don't have 2 "Return Claim" powers to release all Al tiles.
Should I build Fo production on all tiles? Then I would be buying Wa as inputs to my Fo production. That will quickly reduce the profitability of Fo. I do still want to earn from Fo even as I sell Fo price down.
7 Fo will require 7.0 Wa per turn to feed. I currently have 5.25 from a trio. So I build 2 more Wa.
What about feeding Wa production? 2 more Wa require 2.0 Fu more per turn to feed. I now have a surplus of 1.52 Fu. Building 1 more Fu will at least give me 2.52 Fu surplus (more if you calculate adjacency bonus). So I build 1 more Fu. 2 more Wa will also need 0.5 more Ox. My current Ox surplus of 0.75 will suffice.
And there it is. A HQ geared towards earning from Fo, but has 5 Wa, 4 Fu, and 3 Ox (and just 7 Fo). To a newbie, it will look like an assortment of facilities. To a trained eye, it's actually a Food-centric HQ.
Wait. Why am I selling Wa now? Firstly, this is the end game. Selling Wa gives me faster income to deal killing blows. The usual paradigm here is "sell everything"! Also, selling down Wa dampens Wa income for P2 and P3; yes, they're still quite far away from locking, but no harm dampening their Wa income.
But wait, won't P4's Fo profitability increase from lower Wa price?
I'm aiming to reduce P4's CR from Fo. CR is fixed by the current price of Fo and Wa. Yes, I'm selling down Wa. But that will simply "drain P4's cash less". Yet, at the same time, I'm selling down Fo faster than I'm selling down Wa. My HQ is a Fo-centric setup now.
In totality, I'm actually at least reducing P4's cash income, and hopefully soon draining his cash as well. (But make sure you don't sell down Fo so much that his 2 Fo turn off; he'll stop buying Wa by then!)
In this game, you won't often have an absolute monopoly over a resource (I would wish P4 is also mainly selling Wa, just like what P2 and P3 are doing). You often have to do some fine calculations and maneuver for an outrunning posture.
So P4 sells Fo. I sell Fo faster and better (more efficiently).
Ox is still stockpiled, not sold. I'm preparing to go all Wa later, so I better stockpile Ox; Ox is not a continuous supply.
Just about now, the Hack and Opt come online. I hack up Wa price, and Opt Wa output.
The income from all that adjacency --- Wa, Fo, Fu --- is incredible. The income rate is so pumped up that I find myself quickly buying up P3.
Avoid Killing Players Others Invested In
At the point when I have 3 of P3's shares, P2 buys 1 into P3. Why?
P2 wants to raise his own share price, thereby making himself more difficult to be killed off.
Soren programmed the AI extremely well (too well that it possibly hurt the game's sales). The AI knows that P3 is about to killed. Upon being killed, all debt is zeroed. So, if P2 has 1 share in P3 at B rating, that single share's price will leap up tremendously when P3's debt is zeroed.
At S2:42:10 P2's target share price is $16.01. I will kill P3 at this time.
Watch what happens when P3's debt is zeroed the next second. Erm. Ok, nothing. That's possibly because P2's debt keeps increasing, and also because P2 didn't have a substantial stake in P3. You may notice P2's share price sticking to $16.01 for quite some time before falling to $15.85 and lower. That may be the share price "jump" I was talking about.
If you kill off P3 when P2 has 5 shares in P3, there will be a substantial jump in P2's share price. I made that mistake as a newbie, killing off players that others have invested in.
Anyway, I'm glad P2 blew his cash on buying P3. He now has less cash with less options.
Keep Up Black Market Moves
I had Gooned my Opt because I know that a Wa hack up was already launched and on its way. My Opt was more important than my Hack at this point.
I got a Dynamite from P4 from this Goon. I quickly used it on P4's Lab. The Lab will cost him $4.5K (half of full price) to rebuild. That will take from his cash.
To Off or Not --- Production Draws From Cash
Earlier, I had bought EMP, but didn't use it yet. Why?
I was already outrunning P3 fast. I was selling down Fu and Wa, severely dampening his income from those resources. My own adjacency bonuses and enhanced output ensured that my Wa remains very profitable. My Fu adjacencies is better than P3's. In short, I made Wa and Fu a lot more efficiently than P3 makes them. I sold them more efficiently too, consequently.
P2 was in a much weaker state than P3, having a lone Fu with no adjacencies, which was entirely committed to feeding a lone Wa.
P4 is where the discussion is most interesting. Should I EMP P4's 2 Fo? But that would mean P4 no longer buys 1.0 Wa per turn. Inflicting EMP on P4's Fo will also mean turning off P4's Wa as well, which doesn't mean anything more than drawing more from debt; life support buys Wa with debt, not with cash.
I have found no use for the EMP on P4.
But later on (at S3:09:00 when the sun comes up) I did find EMP useful for P4's Ox trio. That would have given just enough cash for P4 to lock out! (There was a surprise Ox shortage a short while back, that took me by... surprise.)
But after inflicting EMP on P4's Ox, the stockpiled Ox (not yet at full freighter capacity) in those Solar Condensers were immediately flown back to HQ. P4 now has $17K cash, and can lock out at a share price of $16.36! Fortunately, I had used Dynamite on P4's Lab, and P4 now spends $4.5K to rebuild the Lab.
Optimization Center Equals More "Tiles In Play"
Upon killing P2, I wanted to replace some Fo with Wa before going for P4. But I found no need to. My Opt had in effect given me almost 9 Wa tiles in place of the 5 I physically own! Now my HQ is a Wa-centric production without me even touching the original Fo-centric setup!
Nothing much to say here. The 2 Fo keeps P4's cash suppressed at around $14K while I quickly bought up his shares and raised his share price out of his reach. (+5% for each share above 4 held).
And finally, some concepts mentioned above explained below.
Hope you enjoyed the replay and analysis!
The first factory you put down gives you 1 of that factory. The 2nd you put down effectively gives you 2 more. The 3rd will give you 2.25 more. A trio gives you a total of 5.25 factories. The 4th one you put down gives you only 1.95 more.
The "holy trio" is what you should always aim for.
Power Transportation Is Free
The reason I am willing to build Po production facilities away from HQ is because they don't take any money to transport. The Po produced is instantaneously "teleported" to me, free of charge.