Something Ain't Right with Installation Upgrades [RESOLVED AND IN PROGRESS]

[notorious_BTC]

I’m pasting both coderdan’s words from the Discord chat as well as my thoughts so they’re better recorded

coderdan:

Why did we make it suboptimal to upgrade all the way to Lvl 9? It comes down to a few factors:

1. Opportunity cost
   If you only upgrade to Lvl 4-5, yes you are getting the “optimal” yield, but you’re not getting it very fast, compared to a Lvl 9. In farming, time is everything. We all hope these tokens will do well and not get farmed to 0, and we’ll be doing our best to have a wide variety of fun and useful sinks to offset, but in the end it’s still farming. People are gonna dump. So there’s still an advantage to getting to Lvl 9 faster than others, as you can clean your parcel out faster.

2. GLTR
   From the conversations above, it seems like maximizing your APR is the goal. But what if your goal is to maximize GLTR? You’ll want to build a large position of Alchemica to LP with, and to do that you’ll need to compete against others doing the same. Thus, you’ll need to move faster. That is how the higher level installations make sense.

3. Strategy
   A good game has multiple winning strategies, and we felt that “get to lvl 9 as fast as possible” was a bit too simplistic. There definitely is a rewarding strategy for power-building to Lvl 9, but it may not be for everyone.

That’s a general summary of our thoughts on the issue. We did consider many “taxes” on various aspects of gameplay (such as harvesters breaking down over time and needing to be repaired), or an installation that would let you expand how many harvesters you could build, but those seemed unnecessary and would probably lead to Spacious owners protesting, so we chose not to include them.

Happy to entertain more thoughts from our smartest minds though!

me:

Thank you for sharing the thought behind the logic that went into the decisions made.

I agree with you that there ought to be more than one way to “win” the farming game. However, with the current build, there is no win for a lvl 9 installation, especially for humble parcels.

There simply isn’t enough alchemica in the land to warrant anything above level 5 at the absolute maximum.

For instance, a level 6 FUD harvester at today’s rates costs 284 GHST.
The total amount of FUD in a humble across all acts is 28,473.
If you look at the cost of the parcel (100 floor) + the cost of just one lvl six harvester and a lvl 1 reservoir (horribly inefficient setup, btw) nets you a profit of just 3.65 GHST.

To me this is a bit sad for players with smaller bags as their land will never reach a pinnacle. The math clearly shows it is advantageous to build out as many low level harvesters as possible. Even with regards to the rate of extraction, there reaches a tipping point where a large qty of mid lvl harvesters outpaces in rate + cost that of a smaller qty of max lvl harvesters.

I think a good game gives players something to aspire towards. If the “end game” of the farming installations is actually far worse in terms of returns then I find that incredibly discouraging. Not to mention we’ll never get to see all the hard work that went into the artwork for higher tier installations!!

The low lvl spam building strategy also promotes a Matrix movie-like setup where the whole realm is nothing but rows and rows of harvesters as to use the space for anything else impedes our ability to earn.

A few ideas that have been suggested that make sense to me:

Cap the qty of harvesters allowed on each land. Something like 8, 32, and 64 for each size.

Make the higher levels scale better. Even if we keep diminishing returns, the severity of the scale might be better tweaked.

Increase the length of build time for each upgrade. This prevents faster inflation while also promoting GLTR & the currently underwhelming Maaker.

[notorious_BTC]

It’s also worth mentioning that it becomes economically more sensible to simply buy a fresh land and spam build it than to upgrade further. Even if there becomes scarcity in the lands over time and alchemica become sparse, it still will be better to build a huge array of cheap harvesters as the higher levels will become exponentially more expensive as alchemica is too.

Perhaps I lack imagination, but I can’t think of a scenario where a lvl 9 maaker (valued at 20k GHST right now) is a better decision than buying 5-10 more parcels and building them up with cheap harvesters all with an instant build.

[HARDKOR]

Land is grossly undervalued. It’s not fair to use todays prices at all, to value land.

L9 has no place on a humble.

What is the best way to run these numbers?

I like that you chose FUD, as it is the edge case, on valuation.
What are all the parameters we are actually trying to solve for here?

RIO, opportunity cost, Net Alchemica Yield, Net GHST yield?

What is appropriate time frame to solve for? 1 spacious parcel, average rolls, left alone for a whole act?

Is there an excel solver we can build out, to answer these questions, so the discussion can be done with better tools? Without that, the best we can do is make a hypothesis, and then run those specific numbers.

@notorious_BTC

The low lvl spam building strategy also promotes a Matrix movie-like setup where the whole realm is nothing but rows and rows of harvesters as to use the space for anything else impedes our ability to earn.

I feel like you provided the counterargument right there…the land is valuable in and of itself, and the harvesters are taking up too much space.

[HARDKOR]

Sign up | Miro | Online Whiteboard for Visual Collaboration

I started a Miro for this. This is the link for if you want to edit it.

I pulled everything on the forum and discord apart into sections. Feel free to Arrange things differently, if you feel there is a better way.

[notorious_BTC]

We may ultimately decide to keep things as they are, but it is of my opinion we at least lay out a few different scenarios and decide on one which seems, based on our forecasting:

• the most fun strategy-wise. Ideally we have a few ways to play where different routes lead to success in their own right

• The best visually for the realm. Huge arrays of lvl 1-3’s I don’t believe is something anyone wants to see. We want to have space for aesthetics on our land

• The best economically. We want a design structure that keeps alchemica semi-scarce where the players are gobbling up the supply to advance, thus minimising inflation.

[notorious_BTC]

There’s a lot of juice in the DAO discussion, so I’m pulling this here from Coderdan

How open would ppl be to a patch that:
-Reduced harvest rate for lower lvl harvesters
-Increased harvest rate for higher lvl harvesters
-Made lower lvl harvesters a bit more expensive
-Made higher lvl harvesters a bit cheaper
-Possibly played with build time as well

[Umami]

i would be openly against any changes to recipes unless we notice they are imperative, since people have been accumulating specific kinds of alchemica towards their own goals for over a month
I would consider imperative to change anything that would be either unsustainable long term or that would hurt small investors in the short term and couldnt be changed with some other mechanic that hasnt been established yet

[notorious_BTC]

Agree fren. One of the challenges with the current model is that it actually hurts the small investor as space is much more important than leveling harvesters. As such, humbles get rekt the hardest as spacious have so much land to build as many lvl 1’s as possible.

[SlickBB]

Definitely add tiered harvester quantities to that list (i.e. aaltar or maaker level up increases number of harvesters). That was a really nice idea from discord.

I know it may not be ideal to keep referencing CoC as we are different games but another mechanic they had was you couldn’t upgrade your town hall until you’d built a certain level of harvester, barracks, reservoir etc. Perhaps we can implement that here too?

Maybe something like this:

Altar Level Up Requirements
Level 1 - None
Level 2 - 1x L1 Harvester, 1x L1 Reservoir
Level 3 - L1 Harvester of each type, L1 Reservoir of each type
Level 4 - 2x L2 Harvester, 2x L2 Reservoir
Level 5 - L3 Harvester of each type, L3 Reservoir of each type
Level 6 - 3x L4 Harvester, 3 x L4 Reservoir
Level 7 - L4 Harvester of each type, L4 Reservoir of each type
Level 8 - 4x L5 Harvester, 4x L4 Reservoir
Level 9 - L6 Harvester of each type, L6 Reservoir of each type

Please note these are examples and I have not crunched the numbers on these at all.

This system would need to be carefully tuned but necessitates higher level harvesters to get to those higher daily UBI levels of an altar (and higher levels of buildings in general).

The other thing we can do to encourage higher harvester levels is give them a lot of HP in comparison to the lower levels. So when we get to the grid it makes way more sense to have a few high tier, more defendable harvesters than bulk L1’s.

[MikeyJay]

I do think there’s been a bit of a misunderstanding about the math at play here, so I’d like to restate this problem slightly and also address one (of many) possible solutions.

The current harvester mechanics discourage leveling up

As the harvester upgrade economics stand currently, each upgrade has diminishing returns. Because of this, a rational player would not upgrade any harvesters past level 1 until they have exhausted their limits on all parcels. Please consider the below chart showing the ROI of each upgrade of a FUD harvester.

(Note that 1 FUD means 1 FUD worth of any alchemica relative to their supply- where 2 FUD = 1 FOMO, 4 FUD = 1 ALPHA, 10 FUD = 1 KEK)

Consider you have some alchemica to invest and a decision to make:

1. Craft a new L1 harvester - you will have 21x your investment after 1 year.
2. OR – Upgrade an existing L1 harvester to L2 - you will have 11.24x your investment on the upgrade cost after 1 year.

While 11.24x is a pretty sweet return, you could have almost twice the return by just crafting a new L1 (provided you have the space within the chapter 4 limitations to do so).

As you can see in the chart, the higher the level, the less incentive you have to upgrade your harvesters. This doesn’t improve over a longer timespan. Here are the same numbers shown over a period of years:

Don’t larger parcels have more incentive to upgrade than smaller parcels?

The same issue applies to any parcel of any size. Parcel size has nothing to do with harvester leveling because the limitations on the # of harvesters allowed on a parcel scale according to their size. The spacious owner is equally as unlikely to upgrade as a humble owner.

Is this just a temporary issue until all parcels have reached their harvester limits?

Even if your parcel space is full, a rational player is still going to consider other uses of those funds, even within the aavegotchi ecosystem… the 10 year return of 5.44x on a L9 parcel is the equivalent of about 18% return per year. It takes 2 years to breakeven on the cost of upgrading from L8 to L9.

It could be that this risk / return is unattractive, even compared to other investments within the aavegotchi ecosystem.

Why this is a problem for gotchiverse tokenomics:

It has been noted many times that GLTR is the key innovation in gotchiverse tokenomics, and I agree. The utility of GLTR is very closely tied to giving players incentive to level up their installations. Harvesters will likely be the most plentiful installation in the gotchiverse.

In addition, Bible Chapter 4 added the Maaker installation, which will serve as an important alchemica sink in the gotchiverse economy. If players are not incentivized to level up their harvesters, they will not be incentivized to level up their maakers.

I think this issue has the potential to cause the alchemica economy to run too hot in the early days. There are a LOT of parcels to be developed, and if we need to wait for parcels to be full of harvesters before GLTR and maakers to have strong utility, even a short term overheating could pose long term problems.

One Possible Solution

Coderdan brought up one possible solution recently:

What would that look like?

Consider a scenario where the short-term ROI favors crafting new L1 harvesters, but higher level harvesters are a better investment the longer your time horizon. So the ROI chart from earlier could look something more like this:

You could achieve the above with the following recipe:

The main change here is that instead of having diminishing returns with each harvester upgrade, you instead have increasing returns - balanced by an increasing build time.

I do not propose using exactly these numbers. There are other considerations at play in terms of emmissions, how long it would take to farm an average parcel, etc… I haven’t looked at all yet- this is more to illustrate the ratios at play - these numbers could easily be scaled and the slopes adjusted in various ways.

If you would like to see (and double-check!) all of the math behind my calculations above - please feel free to check out this spreadsheet - make a copy and play around with the numbers if you’d like.

This is not the only solution

I have heard several other good possible solutions mentioned. Any solution that gives players further incentive to level up and not just spam L1’s until the gotchiverse is full is an improvement in my mind. I just wanted to throw this out there to hopefully help further add to the conversation as to why a change is needed to rebalance harvester levels.

[soloegoist]

I really like the work done by MikeyJay, thank you fren! I would like to express my opinion about the decreasing/increasing ROI. I think that we can’t do ROI that increases with each level of upgrade because the upgrade should be a price in case the farmer wants to get alchemica faster. Speed is very very important and valuable because these are not just game tokens and we cannot give the same high ROI on high levels as on low levels. It’s inflationary.

If we leave the current recipe then no one will be motivated to upgrade to level 9, and if we make the recipe something like MikeyJay showed, then no one will have a reason NOT to upgrade. I think we should find a middle ground.

We need a recipe with a declining ROI but not as much declining as the current recipe. The ROI should drop slightly with each level, BUT the farmer should be willing to pay for the rate of production. With the current recipe, the farmer is not at all interested in buying production speed at such an expensive price (big loss of ROI and a lot of alchemica to upgrade). But the recipe should be such that the farmer has a DIFFICULT CHOICE: it should be difficult for him to sacrifice a small decrease in ROI for the sake of accelerating production, but at the same time it should be difficult for him to stay at the current level there must be a dilemma !! if the recipe is such that both the production rate and ROI increase with each level, then the upgrade will be too easy and obvious solution.

[MikeyJay]

@soloegoist Really appreciate your thoughtful feedback, fren!

Regarding the issue of the speed of harvesting alchemica, we must consider both the cost of the harvester, as well as its output. If a player had to choose between harvesting 1 FUD a day at an upfront cost of 10 FUD vs harvesting 10 FUD a day at an upfront cost of 1000 FUD, a rational player will choose the former, irrespective of the fact that the latter option technically harvests faster - because the player is motivated by their net position.

Making the difference less severe doesn’t really change a rational player’s decision. The same holds true if I offer the player an investment of 1 FUD = 10 FUD a day vs 2 FUD = 19 FUD a day. The player is still incentivized to choose the former investment, because they can just do it twice and achieve a better yield of 20 FUD a day for an investment of 2 FUD.

Having the alchemica being harvested faster is irrelevant to a rational player, because in the end, their net position is still “slower” with higher level harvesters. The player would need an external factor that put the actual speed of the harvester at a premium - for example, if alchemica that was left unharvested at the end of each act reset to zero, being lost forever. My understanding is that is not the case, and that surveying is an additive process (but please correct me if I’m wrong about this).

The trade-off in the increasing ROI model is that for a period of time, your ROI is worse before it is better. In the example model I showed earlier, if a player is forced with the decision of whether to upgrade an existing harvester from L4 → L5 vs crafting a new L1, the breakeven of ROI between those two decisions is about 5 months- after 5 months, L5 upgrade would be the more profitable route. So the player must choose to take less return now in favor of more return later.

The example model I showed uses about 8-9 month spread between the ROI of crafting a new L1 to upgrading L8->L9 - the model is scalable - so that time period could be as long or as short as we’d want it to be by changing the slope of the line, but the key innovation really is the increasing ROI.

Now I want to be clear that I believe a model with decreasing ROI of harvester levels could work, but I think in that case the player needs an external limiting factor that ties the number of total harvesters you can have on a parcel to something other than the parcel size. One suggestion would be tying this to the maaker level - by adding this external cost you’re indirectly creating an increasing ROI system, it’s just hidden in the cost of the maaker level. I think that is also a good option to explore.

A model like the one I’m describing has no single optimal strategy, and depends on the player’s value of time. This is why the increasing returns model will not lead to everybody maxing out every harvester. Some will value the time value of money different than others. Further, it rewards players for investing on longer timeframes. The current model in the bible offers no trade-offs for crafting new L1 vs upgrading an existing harvester on any timeframe, except for the limit on the # of harvesters per parcel.

I’d lastly like to note that this question of decreasing ROI affects players choices at every upgrade level. We need not focus on the extreme case of upgrading to level 9. Even a level 2 upgrade is currently a suboptimal decision for the player vs crafting a new L1. This will always be the case with a decreasing ROI model, barring some other external limiting factor on the # of total harvesters.

[soloegoist]

Thanks for the answer @MikeyJay now it’s clearer to me how the ROI changes and that at first it still decreases (“less return now in favor of more return later”). But I would like to clarify what you mean when you say that a rational player would not upgrade L1 to L2 with a production of 10 vs 19 for doubling the price? If I can build 128 harvesters, then of course everyone will build 128 L1 first, but then after all, any player will have a reason to upgrade L1 to L2 if it’s 10 vs 19 because LAND IS NOT FREE. If the land was free, then of course everyone would take another land and build 128 L1 on it, but with a recipe with a decreasing ROI, people will still upgrade if this decrease is smooth and insignificant.

Do I understand correctly that what worries you about the recipes is that they do not leave the player a choice and, as it were, force him to first build the maximum number of L1 and then only allow him to choose to what level he should upgrade them? and would you like the strategies to differ even in such a way that some player from the very beginning would choose a strategy NOT WITH THE MAXIMUM AVAILABLE amount of L1 (of course, provided that this number is not limited by anything, for example, a maker)?

[MikeyJay]

My first and foremost concern with letting land scarcity solve the issue is that it feels like that is a long ways off- given that there is a lot of land yet to be released. I wonder if this would give GLTR too little utility in the meantime, and that the short term damage this would cause would lead to a much longer term threat to the economy.

Second concern is whether the ROI of higher level harvesters will be sufficient to attract capital, or whether a rational player may simply stop upgrading at some level and dump their alchemica rather than re-investing it in upgrades.

[Quincy1971]

Does anyone have any views on the chart Dan placed in the DAO Meeting thread? From the chart that was shared you will be able to recoup the alchemica needed to upgrade in a short length of time after you upgrade. Now the costs of these upgrades is still a little pricey the higher you go, but it is nice to see that any upgrade would “pay for itself” in less than 25 days of completing it. For me just having a clear “payoff chart” is a good incentive to level harvesters.

[HARDKOR]

What if we combine a slightly improved ROI curve with making sure that the hit points ramp up significantly at about the same point that the ROI starts to drop?

Also - now that we know spillage is HUGE, are these numbers taking into account the part where you are losing nearly 50% on an L1, to spillage? We may be grossly miscalculating these numbers, as prior to todays answers on spillage radius, we were mostly thinking it would be achievable to get some or most of it, but it is now clear that most of that is just gone, until you get past the mid level reservoirs.

[Quincy1971]

Are hit points part of the bible for installations or is it something that was later proposed?

[HARDKOR]

We know hit points exist because we know that higher level walls are both taller, and have more hit points. Licks can temporarily disable installations, which confirms that from another angle.

If the curve on the hit points is set properly, it could add more utility to the upgrades without affecting the tokenomics.

This is just like the real world… you can get a bunch of cheap crap off Amazon, that breaks easily and is disturbingly lightweight, or you can get pro stuff that lasts a lot longer, or you can get the industrial version.

What if, at level 9, a harvester makes the parcel not count, when calculating estate size? It’s so big because it’s like a headquarters now. That last surge of money, was to move the administrative staff in.

EDIT: Link to board where hit point alpha was documented.

Making level 9 make sense.

[Quincy1971]

I was aware that the walls had hit points but have not read anywhere about licks being able to temp disable installations could you please post the link

[HARDKOR]

What if level 6 and up can be moved to a new parcel? That gives you a reason to max it out, suck a parcel dry, and then send your big harvester off to mine out a new parcel.

No need to mess with the rates then, as what you are buying, is mobility.

The 1-5 are hard installs, but the 6-9 are basically like the shipping container generator units