PDA

View Full Version : Economy Discussion



Silvio
10-08-2012, 01:28 PM
Now that Nicolast and those who blazed trails before him have left (i.e. Trampstamp), I wanted to revisit some economy discussion. With respect to economy, I feel that the following three philosophies should be employed:

1. Always have something building, upgrading, and expanding.
2. Always choose the most expensive build/upgrade you can afford (with some certain exceptions)
3. Ensure you vary your upgrades to accomplish the most widespread wealth generation (i.e. this means that having 5 buildings generating 10K each is better that having one that generates 50K, because it limits your exposure loss to robberies)

While there are huge variations in how you apply these philosophies with respect to which buildings you build or upgrade, (ex. based on ROI, GFI, etc.) I think that these philosophies give you the best mechanism to build economy the fastest, regardless of methods of application.

Thoughts?

Dravak
10-08-2012, 01:33 PM
No if you have too much empty space stop expanding is the secret , money not used up , is money wasted .
If you have money to keep everything going , then that is it .

Paulio
10-08-2012, 01:37 PM
Silvio, I think you have some good thoughts. There are probably more efficient ways to upgrade than "Always choose the most expensive build/upgrade you can afford," but I think it's a good rule of thumb for those who don't want to over think it.

mxz
10-08-2012, 01:39 PM
I'd disagree with all 3 points. There's not always a reason to expand or build. Upgrades are a good thing to always have going, though.

The most expensive upgrade is not the best upgrade. There's also plenty of times when it's better to save for an upgrade than blow everything just because you have money.

I don't agree with spreading out payouts. My 2 NCs make me over half my daily income. The exception would be to keep a diversified portfolio of 12/24/48s, at a minimum, with NCs, 8hrs, and 1hrs depending on your activeness and strategies.

Dravak
10-08-2012, 01:44 PM
no in general keep upgrading and building isn't a bad strategy , expanding is good too ( but the least important since money legitly is short always) but keep building missile turrets and upgrading money building is the most important aspect of a good hood.

Sorry if you have money to upgrade a money building , is always beter then upgrading defenses ..
Once you are the 6/12/24 strategy , simply means you are already covered for most of the upgrades ..

Silvio
10-08-2012, 01:47 PM
@mxz - just read you NC analysis post. Nice work. However I would always advocate expanding/building. Only exception would be if the expansion time for the next expansion exceeds the build time of the next few buildings required to fill the space.

WRT upgrading - I wholeheartedly agree with the most expensive isn't guaranteed the best, but as a general rule of thumb I've found that when I played against a short cash flow sinking all my cash into upgrades my IPH went up faster than I would have expected. Personally, I use the GFI metric against my limiting factors (combination of cash available and build time - often I cannot get to the game when an upgrade is complete, and I like to maximize the time).

I have also found that once I obtained sufficient higher income buildings, the few robbers that were successful were taking less, due to their limiting factor of available robs. Hence my 3rd philosophical point.

Silvio
10-08-2012, 01:49 PM
@Dravak - wrt upgrading defensive buildings - I wholeheartedly agree. The only ones I have upgraded are the ones I mistakenly upgraded to meet game challenges before I found the forum defining defensive statistic calculations.

mxz
10-08-2012, 01:55 PM
Economically speaking, there aren't enough worthwhile buildings to justify continual expansion. Also, there are economic ramifications (opportunity cost) of using money to expand rather than upgrade. So, from just that economic perspective, it's tough to justify a $15m upgrade when you could just sell off more useless buildings to create the space for better ones.

Robbers are going to be present...if an attacker really wants to rob you they will. And if you have a high IPH you're more likely to be a sought after target, regardless of whether your income comes from a myriad of buildings or your 2 NCs, Office Buildings. By the time you're putting up Office Buildings and Casinos everyone remembers you and your collection times if they get them. The best strategy for not getting robbed is collecting on time. Some almost never get robbed - the exception is when they can't log in regularly. There's nothing you can do about it, then...people are going to take the goods that are available whether you have 2 awesome buildings or 12.

Dravak
10-08-2012, 02:03 PM
The whole trick to this game is grab the 30% if you cannot collect on time , 30% of something beats beter then 30% of nothing , people love to decorate there hood , is that space wasted definetly ..
But nobody is complaining about that , if those space for decoration is used for a example a flower shop 10 ..
It brings in money ..

Space and Decoration and Strategy is up to people to decide .
But yeah if your game is about preventing robberies as much as possible , well then time strategy is usefull ..
But in general economical tactics , more money buildings more income !

Delete money buildings you don't like for more room if needed , but if you are flooded with space ..
Then you don't have to delete anything , it is also important to note , do not overinvest in defense buildings.
When your have too little money buildings .

Since without cash you can't upgrade or build .. let alone expand .
At endgame lol if people want to waste stamina robbing a 4k building let them , stamina is worth a lot more then 4k ;)

dudeman
10-08-2012, 02:06 PM
I like to always have an upgrade going. Priority goes to money buildings, but you can use defense building upgrades to push your start time if you want a money building to finish upgrading closer to your normal collection time.

Once you have all the money buildings you want, building isn't really that important unless you want to build defense buildings. If you decide it's no big deal, building and expanding takes a back seat to upgrading money buildings.

If you decide you want the defense, at least building the turrets will take over 100 days if you build the Gatlings too and you shouldn't be too pressed for space in that time. When you are ready for the larger defense buildings, you will have upgraded your money buildings to the point of being able to expand a few times and not even think twice about it.

I don't upgrade based on most expensive or biggest IPH increase. I upgrade based on completion time and work the upgrades so that they finish closer to convenient times rather than, say, in the middle of the night. I'm not really sure how to properly describe my method, but it works for me and I don't have much trouble saving for upgrades.

I have pretty much spread out my upgrades. Nothing to do with robbery prevention though (just use an alarm), it's all about generating income. If I rely on two NCs as the foundation of my economy, upgrading one of them can be a serious blow to your income, for DAYS! That's madness! Lol. I find it a lot more efficient to have many high paying buildings. This gives me the ability to chain upgrade a building while making and saving loads of cash, instead of starting an upgrade and crippling my economy because of it.

Different strategies for different people though. I'm sure my game plan wouldn't work for everyone, just the same as other people's strategies don't always fit my style.

Ramshutu
10-08-2012, 02:16 PM
For economy, it is all about overall income per hour gain in the long term. It doesn't matter if you build twenty buildings and upgrade a bunch of buildings up to level 10. If in the same time, you could save all that money in the same time, you were spending on the upgrades and buildings, and spend it on a single build that nets the same IPH gain, this is the best approach to take.

Jtstar7439
10-08-2012, 02:27 PM
I try to put the bulk of my upgrades into my 12/24/48's because they are synced and never get robbed. I upgraded my night clubs to level 2 but think I will stop there. I'm at 157 and they only get robbed once every few days but that will change when I hit the shark tank in the next few weeks.

So while I have a few hundred million banked and could easily upgrade them to level 4 all I think that will do is falsely boost my IPH cause I'm never gonna get 4 collections out of them.

mxz
10-08-2012, 02:44 PM
At 157 you've still got a ways to go to get to the shark tank. Last couple weeks it was pegged around 176.

Euchred
10-08-2012, 03:06 PM
Economy discussion is way to confusing to me, I'm more of a do'er.

Jtstar7439
10-08-2012, 03:22 PM
At 157 you've still got a ways to go to get to the shark tank. Last couple weeks it was pegged around 176.

Well it sounds stupid but I'm looking forward to getting beat up lol. I'm farming TLC's and have been leveling about every other day so I'm guessing around 3 weeks. Down around here I have only about 1 maybe 2 challengers per level. Also I'm really hoping that the hoods are better up at the top cause 95% suck here.

emcee
10-08-2012, 03:44 PM
1. Always have something building, upgrading, and expanding.
2. Always choose the most expensive build/upgrade you can afford (with some certain exceptions)
3. Ensure you vary your upgrades to accomplish the most widespread wealth generation (i.e. this means that having 5 buildings generating 10K each is better that having one that generates 50K, because it limits your exposure loss to robberies)
Thoughts?
1. Always have something upgrading.
Always have something building only applies to certain circumstances. Some buildings are not worthwhile to build in the long run and your finite upgrade time (before you quit) is better spent on select buildings.
Never expand beyond your means. Little to no empty space in your hood. A lot of type B buildings aren't even worth building and the money spent expanding your hood would be better off upgrading type A's, purchasing MTs, Lofts and eventually NCs.

2. This is too broad a statement. I agree when it applies to NCs for instance but laundromats are one of the cheapest upgrades and one of the best in the long run.

3. I agree to a point depending on what building mix (type A yes and B not so much). On my level 73 account I am actually doing the opposite. I have already sold 42 buildings and are keeping buildings for now that are producing above $13/hr/square. This goes back to using every square inch of your hood efficiently and keeping hood expansions in check until economy has flourished.

Silvio add me and you can see what I am doing on my Miguel account 998-375-471.

KemoKidd
10-08-2012, 07:53 PM
I agree with upgrading being better than building. I also do what DudeMan says in its not always about the best iph but also at what time it will be finished. I will expand when necessary but I am always upgrading.

kitzune
10-08-2012, 08:23 PM
Should you not be looking for best Income per hour with lowest return on investment? That way the upgrade or building is going to pay itself off earlier and helping your income grow faster?

mxz
10-08-2012, 08:38 PM
Should you not be looking for best Income per hour with lowest return on investment? That way the upgrade or building is going to pay itself off earlier and helping your income grow faster?With that logic you'd be stuck on Italian Restaurants to 10's while you could afford better buildings (MT, Lofts).

617Pats
10-08-2012, 09:03 PM
Economically speaking, there aren't enough worthwhile buildings to justify continual expansion. Also, there are economic ramifications (opportunity cost) of using money to expand rather than upgrade. So, from just that economic perspective, it's tough to justify a $15m upgrade when you could just sell off more useless buildings to create the space for better ones.

Robbers are going to be present...if an attacker really wants to rob you they will. And if you have a high IPH you're more likely to be a sought after target, regardless of whether your income comes from a myriad of buildings or your 2 NCs, Office Buildings. By the time you're putting up Office Buildings and Casinos everyone remembers you and your collection times if they get them. The best strategy for not getting robbed is collecting on time. Some almost never get robbed - the exception is when they can't log in regularly. There's nothing you can do about it, then...people are going to take the goods that are available whether you have 2 awesome buildings or 12.

Agree. I am nearing 1mil/hour (@lvl137) and I have discarded the majority of my beginning and low level buildings. I had wasted too much money on unecessary expansions. Once you start getting to larger upgrades and incomes those low level buildings are worthless. The money the beginning buildings (pizza parlors,souvenir, electronic, etc..) generate becomes so significantly small, they are not worth focusing on collecting. My lvl3 NC generates $1,155,000, which I collect 3 times a day. The previous buildings incomes are typically so small they become irrelevant. Burn the buildings and use the land for something better.

Coldjoey
10-09-2012, 05:39 AM
All points are valid, the big question should not be how high can you get your iph but how much you personally can collect per 24 hours. Two level ten matts collect 10k ph but if you only collect 15 times aday they are not going to pay dividen. No biggie, but if you upgrade same two wedding chapels to level ten you iph is X. but, if you only collect 4 times in a day, the income collected is, half X.
Personally I won't do another upgrade untill lofts and MT's are at level 10 and N/C's are at level 4 (level 3 at present)

TRAZ
10-09-2012, 08:41 AM
I want more space please...lol. I upgrade whenever I get the funds I want for a big project. Sometimes I just give in and do the shorter upgrade time as well. I build whatever d I happen to be on at the time. When it finishes, the next one goes in the queue.

murf
10-09-2012, 10:15 AM
Should you not be looking for best Income per hour with lowest return on investment? That way the upgrade or building is going to pay itself off earlier and helping your income grow faster?

This is exactly what I do...I've had a few discussions on this and only ShawnBB sees this point.

mxz - if my next Italian upgrade has a better ROI then my Loft/MT (adjusted for collection rates) then I would not pass up the upgrade just to save for the Loft or MT upgrade, at some point the Loft/MT ROI is better then I skip the Italian and do the next highest IpH boost that still has a better ROI then the Loft/MT upgrade.

I think this is an inherently simple concept, when projects are non mutually exclusive.

mxz
10-09-2012, 10:48 AM
mxz - if my next Italian upgrade has a better ROI then my Loft/MT (adjusted for collection rates) then I would not pass up the upgrade just to save for the Loft or MT upgrade, at some point the Loft/MT ROI is better then I skip the Italian and do the next highest IpH boost that still has a better ROI then the Loft/MT upgrade.

I think this is an inherently simple concept, when projects are non mutually exclusive.Here's an example from my game.

I'm looking for my next upgrade after NC3. I have something like $10M saved since starting the upgrade.

Based on gain calculated break-even days (what some people incorrectly refer to as ROI), the Pizza Parlor is the best upgrade at 3.2 days before it breaks even. It nets me an extra $4k a day. Even the Italian Restaurant which breaks even (based on gain) at 22.7 days doesn't look too shabby but its only making an extra $12k per day ($50k total). The cost of these two are $12.6k and $272k. The problem? I collect around $15M a day...that chump change they're adding that mean pretty much nothing. Those buildings are essentially decorations.

That's why you really have to throw out "ROI" - its a very incomplete metric. It works well at the very low levels but after you have a decent economy it's mostly useless. More likely, I'll choose the PT2 or Loft7 upgrade because they add meaningful dollars to my economy. They may not break even as quick, but they are far better investments.

PawnXIIX
10-09-2012, 11:27 AM
That's why you really have to throw out "ROI" - its a very incomplete metric. It works well at the very low levels but after you have a decent economy it's mostly useless. More likely, I'll choose the PT2 or Loft7 upgrade because they add meaningful dollars to my economy. They may not break even as quick, but they are far better investments.

Yep. Whenever I want to determine the next upgrade I should pursue, I tend to use this metric - GFI. Ramshutu was the person to post it, i'm not sure if anybody even uses it. I've seen it pop up in a few threads but I personally like it :)

http://www.funzio.com/forum/showthread.php?33363-ROI-is-dead.-Long-live-GFI

kitzune
10-09-2012, 11:48 AM
Here's an example from my game.

I'm looking for my next upgrade after NC3. I have something like $10M saved since starting the upgrade.

Based on gain calculated break-even days (what some people incorrectly refer to as ROI), the Pizza Parlor is the best upgrade at 3.2 days before it breaks even. It nets me an extra $4k a day. Even the Italian Restaurant which breaks even (based on gain) at 22.7 days doesn't look too shabby but its only making an extra $12k per day ($50k total). The cost of these two are $12.6k and $272k. The problem? I collect around $15M a day...that chump change they're adding that mean pretty much nothing. Those buildings are essentially decorations.

That's why you really have to throw out "ROI" - its a very incomplete metric. It works well at the very low levels but after you have a decent economy it's mostly useless. More likely, I'll choose the PT2 or Loft7 upgrade because they add meaningful dollars to my economy. They may not break even as quick, but they are far better investments.

with your income the Pizza parlor or Italian restaurant or any low level will never be the best or highest IPH per ROI.

If you had 6 buildings with similar 50k+ IPH and then you would choose the upgrade that has the fastest ROI payoff. Otherwise you would be tossing more money for a smaller upgrade.

PawnXIIX
10-09-2012, 11:54 AM
with your income the Pizza parlor or Italian restaurant or any low level will never be the best or highest IPH per ROI.

If you had 6 buildings with similar 50k+ IPH and then you would choose the upgrade that has the fastest ROI payoff. Otherwise you would be tossing more money for a smaller upgrade.

The biggest problem with ROI is that it only determines the return from that building alone. The other problem is that it assumes perfect on time collections as well, and there's only a few people that I know who are capable of that level of collection. Usually I miss a nightclub every once in a while from work or sleep :/

ROI is not a perfect system and honestly it shouldn't be counted upon. I rob more than 5x my IPH and that isn't factored into the ROI, and neither is the rest of our economies.

Ramshutu
10-09-2012, 12:14 PM
with your income the Pizza parlor or Italian restaurant or any low level will never be the best or highest IPH per ROI.

If you had 6 buildings with similar 50k+ IPH and then you would choose the upgrade that has the fastest ROI payoff. Otherwise you would be tossing more money for a smaller upgrade.

As buildings increase, the income increase to cost decreases. Eg: You have to pay more for each $\hr as the buildings get higher level. This means, that by its very nature, low cost low increase buildings are preferred in terms of ROI. This means that you may want to use that 10m scratcher payout on upgrading your laundromat to level 10, rather than to upgrade your loft up to level 4, for example.

Another example, is looking simply at $\hr increase. However this means that with an income of $357/hr the best build you could make is the nightclub, which would take 8 billion years to save for.

GFI, as pawn said, is a fusion of the two. For a given upgrade, and current economy, it tells you how much your income will increase per hour, inclusive of upgrade time AND save time.

If you plugged it in for all possible builds and upgrades, it will tell you the upgrade that will net the biggest economic gain, even including the time that it takes you to save for it.

murf
10-09-2012, 10:13 PM
Here's an example from my game.

I'm looking for my next upgrade after NC3. I have something like $10M saved since starting the upgrade.

Based on gain calculated break-even days (what some people incorrectly refer to as ROI), the Pizza Parlor is the best upgrade at 3.2 days before it breaks even. It nets me an extra $4k a day. Even the Italian Restaurant which breaks even (based on gain) at 22.7 days doesn't look too shabby but its only making an extra $12k per day ($50k total). The cost of these two are $12.6k and $272k. The problem? I collect around $15M a day...that chump change they're adding that mean pretty much nothing. Those buildings are essentially decorations.

That's why you really have to throw out "ROI" - its a very incomplete metric. It works well at the very low levels but after you have a decent economy it's mostly useless. More likely, I'll choose the PT2 or Loft7 upgrade because they add meaningful dollars to my economy. They may not break even as quick, but they are far better investments.

You are clearly missing my point, I'm not saying that the Pizza Parlor is the better upgrade, I'm saying that if you sort your next upgrades by IpH increase. Your top one is your base, so in your case NC3. Then as you move down the list, I would upgrade the next one that has a better ROI then your NC3. It might be your next Loft upgrade, DR upgrade, MT upgrade etc. And if my next best ROI upgrade is very cheap, that's ok, because then I'm saving for my NC3, because I don't have any better investment options.

Basically my point is don't skip better mid-high levels ROI upgrades, strictly to save for your highest increase in IpH.

I try and use real life examples. If you have $10,000 and are offered 2 options.

a) $100,000 investment - 10% return - $10,000 income
b) $10,000 investment - 20% return - $2,000 income.

if these choices are not mutually exclusive, I would never (and I hope no one would) pass up choice (b). And by not choosing your mid-level, better ROI upgrades, to save for your high-level lower ROI upgrades, so your in fact passing up investment (b) to save for (a), which IMO is lunacy.

murf
10-09-2012, 10:16 PM
with your income the Pizza parlor or Italian restaurant or any low level will never be the best or highest IPH per ROI.

If you had 6 buildings with similar 50k+ IPH and then you would choose the upgrade that has the fastest ROI payoff. Otherwise you would be tossing more money for a smaller upgrade.

Thank you!!! Thank you!!! Thank you!!!. Finally someone who sees the world the same way I do!!!!!

mxz
10-10-2012, 04:22 AM
You are clearly missing my point, I'm not saying that the Pizza Parlor is the better upgrade, I'm saying that if you sort your next upgrades by IpH increase. Your top one is your base, so in your case NC3. Then as you move down the list, I would upgrade the next one that has a better ROI then your NC3. It might be your next Loft upgrade, DR upgrade, MT upgrade etc. And if my next best ROI upgrade is very cheap, that's ok, because then I'm saving for my NC3, because I don't have any better investment options.This may actually show Italians as the best upgrade...

Rams' equation is a good method. But sorry, murf, I couldn't recommend anyone past an IPH of $50k use ROI as a major part of their decision making since it favors the smallest investments.

murf
10-10-2012, 06:12 AM
This may actually show Italians as the best upgrade...

Rams' equation is a good method. But sorry, murf, I couldn't recommend anyone past an IPH of $50k use ROI as a major part of their decision making since it favors the smallest investments.

You not listening to me. I'll give an example later to prove my point. But to state it simply again, I order everything by largest IpH gain, then scroll down the list until there is a better ROI investment then my top IpH gainer. It doesn't favor the cheaper upgrades because I'm sorting by IpH.

If my only upgrade with a better ROI then my top IpH gainer is a cheap upgrade, then that's ok. I've exhausted all good potential medium upgrades and now I can focus on saving for my top IpH gainer while doing cheaper upgrades.

murf
10-10-2012, 06:17 AM
This may actually show Italians as the best upgrade...

Rams' equation is a good method. But sorry, murf, I couldn't recommend anyone past an IPH of $50k use ROI as a major part of their decision making since it favors the smallest investments.

I use to do something very similar to Ram's equation. Then when I got to over $1mm/hr. I realized that it kept telling me that the NC was my best upgrade, even though Casinos, Palms, GC, etc all had better ROI and all were expensive enough that I needed to save for them. So, I realized that it would be much better to do those first, then to keep saving for NC upgrades. My income went from $1mm to $2.5mm in a very short period after I switched my methodology.

You don't have to agree with me, but please look back at my real example a few posts ago and tell me why I should not do investment (b) to save for (a).

murf
10-10-2012, 06:39 AM
Here's an example:

Let's assume these are my next 7 highest IpH upgrades.

BUILDING LEVEL IpH gain ROI (days)
Nightclub 6 $224,583 101.16
Palm 3 $49,843 105.46
Casino 2 $41,250 103.61
G Club 3 $24,750 90.89
Hotel 2 $13,062 91.85
Loft 6 $7,058 61.99
MT 8 $3,125 71.70

OK, so in my methodology I sort it my IpH as above. Then I set my ROI benchmark to my highest IpH upgrade, which in this case in NC -> L6 (or 101.16 days). I scroll down my list and I see the GC -> L3 is the first better ROI upgrade, so if I have the money for the GC upgrade, but not the NC upgrade I will do the GC one first. If I don't have enough money, I then scroll down the list for a better ROI then my GC, so my new benchmark is 90.89 days. Loft is then my best upgrade, if I have enough money for the Loft upgrade, I do that.

What happens is I save money while doing the Loft upgrade to be able to do the GC upgrade when I am done. And I continually rinse and repeat this process until all my top upgrades have a worse ROI then my NC-> L6 and only at that point am I'm doing cheaper upgrades like MTs, which will allow me to save for my NC.

To say ROI isn't meaningful is the most ridiculous statement I've ever heard. The inverse of ROI is yield, and if you don't care what your investments yield, I'm not really sure what else you can care about. At the end of the day, you want to earn the highest return of your invested $$'s. That is what I am doing here.

mxz
10-10-2012, 08:25 AM
Highest return percentage is a bad stat, though. Your method, as far as I can tell, relies on an arbitrary cut off of buildings. That's fine...but I don't see a reason for stopping at MT (maybe there is one?). At the point you're arbitrarily making cutoffs of which buildings, I don't understand why even sort by IPH gain.

When saving for larger upgrades I generally set my bar at cost needs to be between what I make in 6 hours and 1 day. 6 hours because its not insignificant and 1 day because it's less than 20% of weekly income. Those numbers are highly arbitrary based on my style of play - I wouldn't recommend them for everyone. My goal is not to earn the highest amount compared to what I spend - its to earn the highest amount I can (period). For me, then, NPV of upgrades is the most important metric. ROI does make sense if you care more about spending minimal amounts of money compared to what you make...but my experience is that's a very small percentage of players that care about their economy.

PawnXIIX
10-10-2012, 08:49 AM
The only problem I have is that ROI uses a perfect collection scheme. I'm usually asleep for the 4th nightclub collection and I get it about 1/2 the time when I wake up.

However you want to apply it is fine, opinions are opinions. It's just not a perfect science.

mxz
10-10-2012, 09:51 AM
Some of us estimate collections and extrapolate from there :)

My spreadsheets take into account realistic collections/day. For instance, I use 2.5 for 8 hrs and 3 for NCs. I actually get closer to 2.9 and 3.5....but it's not hard to convert to (real) IPD and back to (real) IPH if you need to.

Ramshutu
10-10-2012, 10:32 AM
We are all most concerned with building are economy as large as possible fast as possible.

GFI, while not including mutiple upgrade path is a measure of just that, the numbers it gives, is the speed at which a particular upgrade (and saving) will improve your economy.

ROI is not directly tied to increasing the economy, it deals with the upgrade in isolation; how long it takes an upgrade to return your cash on hand to the level it was before it was purchased. Now this is okay, but it has some massive conceptual failures:

- Every building will, eventually, pay for itself. In the typical sense of ROI in a business sense, all buildings have an infinite ROI. For a dollar invested, you make an infinite amount of money.
- Once the building has paid for itself, the cash is surplus and gets plowed into your economy. At this point, a massive income building that takes a long time to recoup its investment will earn free cash at much faster rates than a low ROI item. A good modifier to ROI would be to change the equation to say 'how long it takes an upgrade to pay for itself, AND earn you an additional $1m' I guarantee you, that this would show some interesting things.
- ROI is calculated as a relative value of the upgrade cost, not your economy as a whole. There is nothing in the calculation that deals with economy increase.
- what it is measuring is not economic benefit, but short term affordability. It is inherently trying to minimise time spent saving over increasing output. This is bad, because even taking into account time spent saving, some upgrades with bad ROI make massive contributions.

PawnXIIX
10-10-2012, 11:51 AM
We are all most concerned with building are economy as large as possible fast as possible.

GFI, while not including mutiple upgrade path is a measure of just that, the numbers it gives, is the speed at which a particular upgrade (and saving) will improve your economy.

ROI is not directly tied to increasing the economy, it deals with the upgrade in isolation; how long it takes an upgrade to return your cash on hand to the level it was before it was purchased. Now this is okay, but it has some massive conceptual failures:

- Every building will, eventually, pay for itself. In the typical sense of ROI in a business sense, all buildings have an infinite ROI. For a dollar invested, you make an infinite amount of money.
- Once the building has paid for itself, the cash is surplus and gets plowed into your economy. At this point, a massive income building that takes a long time to recoup its investment will earn free cash at much faster rates than a low ROI item. A good modifier to ROI would be to change the equation to say 'how long it takes an upgrade to pay for itself, AND earn you an additional $1m' I guarantee you, that this would show some interesting things.
- ROI is calculated as a relative value of the upgrade cost, not your economy as a whole. There is nothing in the calculation that deals with economy increase.
- what it is measuring is not economic benefit, but short term affordability. It is inherently trying to minimise time spent saving over increasing output. This is bad, because even taking into account time spent saving, some upgrades with bad ROI make massive contributions.

+1 for the issues on ROI. It causes tunnelvision in the sense that your economy is most like not staying stagnant.

Another aspect of ROI is that lets say you have a building that costs 250k to build, and your IpH is 25k. It would take your economy 10 hours to recuperate that cost. When in reality you can take 2 minutes, find a no banker and hit him for the money back right away.

I also like your GFI ramshutu. I did have to say NO to it though, because nightclub to level 3 is 101 while hotel to level 2 is only an 85 for me considering the nightclub costs nearly 5x more. I just don't have $111M to spend xD

Ramshutu
10-10-2012, 12:09 PM
I think it needs a little tweaking, as it compares upgrades in isolation. Not as part of a whole sequence.

Im working on a application that does some serious algorithm work to work out most efficient upgrade path. First step is to work out income using ROI method and GFI.

prostouser
10-10-2012, 02:01 PM
Here's an example:

Let's assume these are my next 7 highest IpH upgrades.

BUILDING LEVEL IpH gain ROI (days)
Nightclub 6 $224,583 101.16
Palm 3 $49,843 105.46
Casino 2 $41,250 103.61
G Club 3 $24,750 90.89
Hotel 2 $13,062 91.85
Loft 6 $7,058 61.99
MT 8 $3,125 71.70

OK, so in my methodology I sort it my IpH as above. Then I set my ROI benchmark to my highest IpH upgrade, which in this case in NC -> L6 (or 101.16 days). I scroll down my list and I see the GC -> L3 is the first better ROI upgrade, so if I have the money for the GC upgrade, but not the NC upgrade I will do the GC one first. If I don't have enough money, I then scroll down the list for a better ROI then my GC, so my new benchmark is 90.89 days. Loft is then my best upgrade, if I have enough money for the Loft upgrade, I do that.

What happens is I save money while doing the Loft upgrade to be able to do the GC upgrade when I am done. And I continually rinse and repeat this process until all my top upgrades have a worse ROI then my NC-> L6 and only at that point am I'm doing cheaper upgrades like MTs, which will allow me to save for my NC.

To say ROI isn't meaningful is the most ridiculous statement I've ever heard. The inverse of ROI is yield, and if you don't care what your investments yield, I'm not really sure what else you can care about. At the end of the day, you want to earn the highest return of your invested $$'s. That is what I am doing here.
+1. But I'm using GFI instead of "IPH gain" to determine main goal and calculate priorities of secondary goals. ROI of main goal is used in the same manner: non-arbitrary filtering out of ineffective updates.

The difference between our decisions will be visible only for cheap building when "Upgrade" time is much longer than "days to save" time.

For example, you would prefer Italian restaurant 10 (IPH gain ~800, Upgrade time 101 hour,GFI ~5, price price 454632 ) over Fancy Restaurant 2 (IPH gain ~300, upgrade time 7 hours, GFI ~ 13, price 450900) and I wouldn't.

BTW I think that using rule "ROI_acceptable_updates <= ROI_current_main_goal" is not perfect, but it is intuitively understandable and must be very close to the really best decision.

On my opinion GFI is pragmatical and adapted for CrimeCity hood model (need to save money, and have to use only one upgrade simultaneously). Some problem is that is not wide-known economy term like ROI, NPV and so on :)
It has insignificant flaws and can't be accepted by purists: profit decrease when upgrading or is not accounted in GFI itself, for example.

Ramshutu
10-10-2012, 02:14 PM
+1. But I'm using GFI instead of "IPH gain" to determine main goal and calculate priorities of secondary goals. ROI of main goal is used in the same manner: non-arbitrary filtering out of ineffective updates.

The difference between our decisions will be visible only for cheap building when "Upgrade" time is much longer than "days to save" time.

For example, you would prefer Italian restaurant 10 (IPH gain ~800, Upgrade time 101 hour,GFI ~5, price price 454632 ) over Fancy Restaurant 2 (IPH gain ~300, upgrade time 7 hours, GFI ~ 13, price 450900) and I wouldn't.

BTW I think that using rule "ROI_acceptable_updates <= ROI_current_main_goal" is not perfect, but it is intuitively understandable and must be very close to the really best decision.

On my opinion GFI is pragmatical and adapted for CrimeCity hood model (need to save money, and have to use only one upgrade simultaneously). Some problem is that is not wide-known economy term like ROI, NPV and so on :)
It has insignificant flaws and can't be accepted by purists: profit decrease when upgrading or is not accounted in GFI itself, for example.

I think that if you do pure comparrisons, those flaws are insignificant. The thing I think is the biggest flaw, is that you cannot easily compare multiple upgrades vs a single upgrade. For example, upgrading an UC from 1-4 may stack up against level 1 NC build, but GfI doesn't show it.

The important thing to note with NCs, however, is that while it is a bit disheartening to save/upgrade/build, they are the biggest income building in the game, and if you had all game buildings to level 10, I think a significant proportion(if not the majority, I forget which, as I read it a while ago) is made up of only your level 10NCs, so it is not a big stretch to work out that getting them to mid numbers is actually a good use of cash, but I am willing to state that it is possible that the maths won't back it up.

I will know in a bit, not knowing the best way to do things bugs me, so I am building an economy computer model, I intend to make it so that you can compare, and create multiple algorithms to determine new buildings, and see whether it really stacks up.

Written class outlines already!

murf
10-10-2012, 02:42 PM
Highest return percentage is a bad stat, though. Your method, as far as I can tell, relies on an arbitrary cut off of buildings. That's fine...but I don't see a reason for stopping at MT (maybe there is one?). At the point you're arbitrarily making cutoffs of which buildings, I don't understand why even sort by IPH gain.

When saving for larger upgrades I generally set my bar at cost needs to be between what I make in 6 hours and 1 day. 6 hours because its not insignificant and 1 day because it's less than 20% of weekly income. Those numbers are highly arbitrary based on my style of play - I wouldn't recommend them for everyone. My goal is not to earn the highest amount compared to what I spend - its to earn the highest amount I can (period). For me, then, NPV of upgrades is the most important metric. ROI does make sense if you care more about spending minimal amounts of money compared to what you make...but my experience is that's a very small percentage of players that care about their economy.

I made the cutoff at MT for sake of an example....I use all the buildings I own, however my methodology is still the same.

murf
10-10-2012, 02:52 PM
+1. But I'm using GFI instead of "IPH gain" to determine main goal and calculate priorities of secondary goals. ROI of main goal is used in the same manner: non-arbitrary filtering out of ineffective updates.

The difference between our decisions will be visible only for cheap building when "Upgrade" time is much longer than "days to save" time.

For example, you would prefer Italian restaurant 10 (IPH gain ~800, Upgrade time 101 hour,GFI ~5, price price 454632 ) over Fancy Restaurant 2 (IPH gain ~300, upgrade time 7 hours, GFI ~ 13, price 450900) and I wouldn't.

BTW I think that using rule "ROI_acceptable_updates <= ROI_current_main_goal" is not perfect, but it is intuitively understandable and must be very close to the really best decision.

On my opinion GFI is pragmatical and adapted for CrimeCity hood model (need to save money, and have to use only one upgrade simultaneously). Some problem is that is not wide-known economy term like ROI, NPV and so on :)
It has insignificant flaws and can't be accepted by purists: profit decrease when upgrading or is not accounted in GFI itself, for example.

Thank you for taking the time to understand my strategy. And thanks for the feedback.

The way I incorporate upgrade time is 2 fold: 1) it is added directly to ROI 2) the amount of income lost during the upgrade period is added to cost of upgrade. I haven't modeled it further as I viewed it as insignificant for the majority of cases, and couldn't find an elegant way to model it, without introducing random assumptions.

I have to be honest, I've only glanced at the GFI calculation awhile ago, and I remember it being very similar a model I had built in where I tried to weight ROI and IpH gain, using my current IpH as an input as the weight, so as I earned more, IpH was more heavily weighted. My model wasn't exactly this, but this was the gist of it. I am going to take a closer look at what GFI actually does before I make any more generalized statements about it.

PawnXIIX
10-10-2012, 03:11 PM
Thank you for taking the time to understand my strategy. And thanks for the feedback.

The way I incorporate upgrade time is 2 fold: 1) it is added directly to ROI 2) the amount of income lost during the upgrade period is added to cost of upgrade. I haven't modeled it further as I viewed it as insignificant for the majority of cases, and couldn't find an elegant way to model it, without introducing random assumptions.

I have to be honest, I've only glanced at the GFI calculation awhile ago, and I remember it being very similar a model I had built in where I tried to weight ROI and IpH gain, using my current IpH as an input as the weight, so as I earned more, IpH was more heavily weighted. My model wasn't exactly this, but this was the gist of it. I am going to take a closer look at what GFI actually does before I make any more generalized statements about it.

http://www.funzio.com/forum/showthread.php?33363-ROI-is-dead.-Long-live-GFI&highlight=Gain+Investment

Just in case you needed a little bathroom reading :)

mxz
10-10-2012, 03:15 PM
I have to be honest, I've only glanced at the GFI calculation awhile ago, and I remember it being very similar a model I had built in where I tried to weight ROI and IpH gain, using my current IpH as an input as the weight, so as I earned more, IpH was more heavily weighted. My model wasn't exactly this, but this was the gist of it. I am going to take a closer look at what GFI actually does before I make any more generalized statements about it.For anyone looking for the simple, quick, and dirty...GFI is definitely the way to go. It's fairly intuitive to understand and produces, roughly, what higher order analysis does. Like Rams said, it's not perfect...but it's a good approximation.

How's this for a list of basic, fundamental economy focused upgrade goals:
Minimize save time Minimize upgrade cost Minimize upgrade time Maximize income gained
The reason GFI works well enough is because it incorporates all 4 factors. In the long term ROI and NPV only really hit on 2 of them (3 in the short term). You could make the case NPV does all 3 long term after you've processed and input your cost of capital but I'm guessing most CC players haven't added that equation to their sheets. So in order to effectively utilize either you have to do some higher level processing that most people won't bother, care or even understand enough to do.

PawnXIIX
10-10-2012, 03:22 PM
For anyone looking for the simple, quick, and dirty...GFI is definitely the way to go. It's fairly intuitive to understand and produces, roughly, what higher order analysis does. Like Rams said, it's not perfect...but it's a good approximation.

How's this for a list of basic, fundamental economy focused upgrade goals:
Minimize save time Minimize upgrade cost Minimize upgrade time Maximize income gained
The reason GFI works is because it's incorporates all 4 factors. In the long term ROI and NPV only really hit on 2 of them (3 in the short term). You could make the case NPV does all 3 long term after you've processed and input your cost of capital but I'm guessing most CC players haven't added that equation to their sheets. So in order to effectively utilize either you have to do some higher level processing that most people won't bother or care to do.

The only problem I found with GFI is that it doesn't take into account the time it took to save up for the upgrade itself. Right now I have a hotel upgrade to level 2 that gives me a reading of 85, and nightclub to level 3 which is 101.

The nightclub on income alone and perfect collections would take me 496 hours to save up
The hotel on income alone and perfect collections would take me 124 hours to save up.

So basically it's telling me that the best upgrade for my economy would leave my economy stagnant for 21 days to save up. There needs to be some scale that should be added to GFI that takes into account the time to save up for the upgrade.

My hotel upgrade starts tomorrow, sorry GFI xD

Euchred
10-10-2012, 03:33 PM
Reading these analysis' is interesting but to me if you care about your economy enough to stay at it the only factors that matter are; how long you've been working on your economy for and how good you collect the rest pretty much does itself. I understand why you guys are doing these but 100% efficiency doesn't mean that much to me.

mxz
10-10-2012, 03:35 PM
Reading these analysis' is interesting but to me if you care about your economy enough to stay at it the only factors that matter are; how long you've been working on your economy for and how good you collect the rest pretty much does itself. I understand why you guys are doing these but 100% efficiency doesn't mean that much to me.Over an infinite time, I agree...inevitably you'll have every building to level 10 - but if you're trying to produce an economy quickly there's certainly more rhyme and reason than blindly and arbitrarily picking upgrades.

Euchred
10-10-2012, 03:44 PM
Over an infinite time, I agree...inevitably you'll have every building to level 10 - but if you're trying to produce an economy quickly there's certainly more rhyme and reason than blindly and arbitrarily picking upgrades.

Ya I don't really give starting out too much thought my camper just hit 1 million per hour and the upgrades I need to do next are painfully obvious same story with my high level account everything falls into line on its own.

prostouser
10-10-2012, 03:52 PM
...
My hotel upgrade starts tomorrow, sorry GFI xD

With my collection profile Nightclub 3 has both better ROI and GFI than Hotel 2 (ROI is 60 days vs 95 days) -> no way I'm going to build Hotel 2 before Nightclub 3. Sry hotel 2. :)

But if my collection profile is 1 collection per day, Nightclub 3 would have worse ROI , and decision will be opposite.

-sry, mxz, I read you post that it is not ROI but break even, just used to it in my spreadsheet.

PawnXIIX
10-10-2012, 04:14 PM
With my collection profile Nightclub 3 has both better ROI and GFI than Hotel 2 (ROI is 60 days vs 95 days) -> no way I'm going to build Hotel 2 before Nightclub 3. Sry hotel 2. :)

But if my collection profile is 1 collection per day, Nightclub 3 would have worse ROI , and decision will be opposite.

-sry, mxz, I read you post that it is not ROI but break even, just used to it in my spreadsheet.

I feel like the 100 million for the upgrade though is what is the issue though. My economy is only about $220k right now, so saving up that money would cause my economy to stay stagnant for too long despite the fact that the nightclub has the highest GFI value. There comes a point where the upgrade cost is too much even if the figures say it's the right upgrade for the best return.

That's what i'm trying to get at :p


Nightclubs are a special deal. They make so much money compared to the other free buildings that they're always going to be the best value or upgrade at low levels. They're type A so level 1->2 is 3x the previous income, and compared to level 1, level 3 is a total of 6x the original income. Due to the nature of their income, it throws off the GFI figure to lead people to believe that this is the best upgrade regardless of their IPH. Like mine, there's no way I can support a $111M upgrade when I make $1.2M on my 24s and just under $600k from my single level 2 nightclub. I'm struggling just to pull together $40M for a second nightclub let alone upgrade the lone one I have to level 3. GFI says do it but I know I can't comply with the figures.

Euchred
10-10-2012, 04:18 PM
That's the interesting and fun part of high level nightclubs i'm currently saving for NC 10 while upgrading palm hotels, office buildings, rock cafes ect.

PawnXIIX
10-10-2012, 04:19 PM
That's the interesting and fun part of high level nightclubs i'm currently saving for NC 10 while upgrading palm hotels, office buildings, rock cafes ect.

I bet you must be enjoying that multi-billion dollar upgrade while there's still players down in the low levels just struggling to get what little money they have in order to unlocks goals :p

mxz
10-10-2012, 04:26 PM
I bet you must be enjoying that multi-billion dollar upgrade while there's still players down in the low levels just struggling to get what little money they have in order to unlocks goals :pAfter your second NC2 is up you're basically Jay Z. Start makin' it rain cause money ain't a thang.

Euchred
10-10-2012, 04:27 PM
I bet you must be enjoying that multi-billion dollar upgrade while there's still players down in the low levels just struggling to get what little money they have in order to unlocks goals :p

I'm not trying to show off, ok just a little. But my point is I didn't put my hotel to level 2 until my nightclubs were level 6.

erik
10-10-2012, 05:29 PM
I bet you must be enjoying that multi-billion dollar upgrade while there's still players down in the low levels just struggling to get what little money they have in order to unlocks goals :p

That is in part a game choice that prevents the multi-billion upgrades :)

Ramshutu
10-11-2012, 04:09 PM
As an interesting side point, I'm going to try and have my computer model ready by the end of the weekend (I know how to kick back and party).

It currently sucks in the money building spreadsheet, and simulate a perfect scenario (processing income, 100% collections, immediate upgrades/buildings) using any algorithm of choice to determine the next upgrade or build. The first pass version will compare GFI and ROI to see how many hours it takes to get from 1 level 1 laundromat and pizza parlour to $1m. I may throw in Murfs idea if I can boil it down into code.

The code here may actually end up being more complicated than the game it is modelling ;)

kitzune
10-11-2012, 06:10 PM
As an interesting side point, I'm going to try and have my computer model ready by the end of the weekend (I know how to kick back and party).

It currently sucks in the money building spreadsheet, and simulate a perfect scenario (processing income, 100% collections, immediate upgrades/buildings) using any algorithm of choice to determine the next upgrade or build. The first pass version will compare GFI and ROI to see how many hours it takes to get from 1 level 1 laundromat and pizza parlour to $1m. I may throw in Murfs idea if I can boil it down into code.

The code here may actually end up being more complicated than the game it is modelling ;)


if you use collections per day you can then have a collection for building that have a payout period less then 24 hours. As an example on 6 hour payouts if you put 2.75 3 out of four nights you collect 3 times and the other time 2.

Ramshutu
10-12-2012, 01:01 PM
if you use collections per day you can then have a collection for building that have a payout period less then 24 hours. As an example on 6 hour payouts if you put 2.75 3 out of four nights you collect 3 times and the other time 2.

I have implemented quite a neat program, that will use "ticks", each tick represents a small portion of time (I have 5 minute ticks at the moment). Each tick, the program will go through and re-calculate what's going on, whether any upgrades/builds have completed, collect all buildings (if they are available).

I wanted to do this so I could do the following later:

- Model the effects of waiting till collections vs immediate upgrade.
- Model the effects of being robbed if buildings are uncollected.
- Model the effects of spending 8 hours asleep, and being unable to collect, and how choosing your upgrade to span evenings can helper.

I have started getting data out on three different algorithms now:

1.) Pure ROI - The upgrade/build order is determined purely by the lowest ROI building that can be purchased with current cash.
2.) $/Hr/Hr - The upgrade/build order is determined by choosing the HIGHEST $/Hr/Hr building that can be purchased with current cash.
3.) GFI - The upgrade/build order is determined by choosing the item with the Highest GFI only, if it cannot be purchased with current cash, the system will save.

The collections are modelled on how the game collects (IE: collecting at different times), and thus models real income loss when upgrading mid-cycle.

I use level caps on buildings, with the level of the simulated player increasing (on average) 1 level per game day. I also ignore the lower level gold buildings, but assume all other gold buildings will be purchased immediately.

On my first data sweep, and I need to double check everything is working properly, it seems that ROI starts falling significantly behind after about 10k IPH. $/hr/hr and GFI are about the same, but strangely, $/hr/hr seems to be slightly ahead in general.

The reason I'm not sure about the algorithms yet, is that it shows that getting to $1m for GFI and $/Hr/hr in ~200 days, which seems a bit on the steep side. ROI gets to $1m in 2 years :*)

mxz
10-12-2012, 10:06 PM
The biggest problem with GFI is saving time is weighted equally as upgrade time. Putting a factor on it makes it more serviceable to anyone also using higher order analysis (I think my own spreadsheet equations weight save time 15x compared to upgrade time). That might be an interesting knob to tweak....

Edit: using that weighting I tend to get very close to what higher order analysis confirms.

Deluxe
10-12-2012, 10:33 PM
Ive got an approach where i do all buildings to same level consecutively. I dont run away with one set of buildings like many do with ncs. Ill save a couple hundred million to start and then ill upgrade the most expensive and time consuming set of buildings first. Because of my initial save i am left with a buffer for upgrade #2 in that bldg pair, plus money earned while bldg #1 completes. Once the ncs are done, ill go to palms and will have enough re-saved to upgrade immediately out of nc upgrade. Buffer shrinks but still exists to extent where i can do palm #2 immediately. Then down to office, gent clubs, hotels, pirate, credit, rock, condos...
What is really nice about this is that w appropriate initial save (will upgrade D bldgs for a week or so) i can upgrade every single remaining bldg to same level without breaking at all. The financial buffer starts big, gets eaten away at substantially with initial, costliest upgrades, but then swings back to increasingly larger buffer as i move into lowest tiered upgrades for bldgs to a certain level. By the time i take my condos to level 4 for a nominal cost, my financial buffer has grown to point where i can recommence cycle of upgrading ncs to level 5 before moving on down and upgrading all to same level.
I only started this for end game bldgs to level 4 so i dont have much background but as my second pirate tavern finishes to level 4 im sitting on 200 mill again with next upgrade only costing 42 million and taking 35 hours. At 2 mill/hour ill make 60-70 million in time it takes my 42 mill upgrade to finish so im +20-25 million return on this building (40-50 mill on the pair) and that will only increase as i move thru rcs and condos like i said. Ill come out w all bldgs to level 4 and ill be sitting on probably 400 million ready to start the cycle over again for all bldgs to level 5. I dont have hard data/analysis but all indications are that i can continue to upgrade without ever having to take a break, not even for the costliest of building upgrades
^^^^^^^^ and the reason that is so important is i want a balanced economy thru all my bldgs so if im waiting 5 days for nc to upgrade, im making serious cash from all my other bldgs and not taking a 20-30% iph hit while a monster econ bldg sits idle and upgrades.

Ramshutu
10-15-2012, 11:10 AM
Okay, so I got my geek on, and created a hood simulation. This basically recreates the typical upgrade/build/collect cycle you experience in your hood, and allows you to test the economy benefit of the above metrics by plugging the metrics into the simulation.

The important things to note are as follows:

1.) It assumes a perfect collection cycle, including overnight.
2.) No lower level gold buildings, with the exception of the basketball court and ice cream store will be purchased, the assumption is that gold is no object.
3.) Space available in your hood is ignored for the purposes of the test.
4.) I've double checked the algorithms, there *could* be something wrong somewhere, but it doesn't appear so from my checks.

I ran the above algorithms together, until all metrics yeilded a total hood economy of $1m.

For comparrison, I also added a test metric, which is simply choosing the quickest upgrade out of the lot.

The program sucks in values from the building spreadsheet that can be found here (https://docs.google.com/spreadsheet/ccc?key=0AgEHrvnoijXudHFVU2RxYUQzYXYtOURRd0d3M1hVR 2c&hl=en_US&pli=1#gid=0)


Results

The results I found were more interesting than I thought.

The following graph shows the different metrics overall economy output vs time. Note: All graphs are displayed with a Logarithmic X axis (1k -> 10k is displayed as the same height as 10k -> 100k).

http://i1063.photobucket.com/albums/t511/ramshutu/IncomevsTime.jpg

What you can see here is fairly interesting, in that there is really not *that* much different between GFI and $/hr/hr. The reason for this, is that while $/hr/hr focuses on affordability, GFI focuses on saving, the former will pick a significant number of cheaper upgrades over saving for the uber big one. This is why GFI is quite bumpy (I remove IPH from buildings undergoing upgrade to show what you are ACTUALLY earning).

What really suprised me, is exactly how much ROI sucked. I mean, really sucked. At the end of the test, after 600+ days, GFI and $/hr/hr were sitting at 7m, with ROI hitting $1m. Not only that, at no point does ROI outperform the others, it stays at about 60-75% of the output of the others up until around 30k when it starts really loosing out. When the main contenders are at $1m, ROI is sitting at a paultry 450k.

This may seem difficult to explain until you see the graph of "cash on hand" (below)

http://i1063.photobucket.com/albums/t511/ramshutu/COHvsTime.jpg

This explains a lot about what is going on, and tends to match what Deluxe was saying previously. Because ROI is minimising the investment, and maximising returns, in a perfect world it ends up generating lots of surplus cash. This cash is not doing anything, there is no interest and nothing really gained by it doing nothing. The other metrics are constantly forcing ALL of the money to be generating more money in the economy, thus generating big gains.

Re-running the test adding "tax" to earnings, EG: 50% of of cash income from buildings disappears (either on other things, or on robberies,etc), but while this lowered the ending income, there was still a significant underperformance by the ROI metric

Conclusion

Realistically, while I put a lot of stock in GFI over $/hr/hr, realistically, there is no big difference between chosing GFI, and chosing the building with the higest $/hr/hr that you can afford when a previous upgrade completed. What surprised me the most, is that these two metrics got the same answers, but did so through different upgrade routes.

Most importantly, is that this test shows that "in general", ROI is a very poor metric; it significantly underperforms all other metrics, to the point that after only a month, it is almost 60% of the income of the other metrics, and it only gets worse from there.

Further tests

I do not think that this is fully the end of ROI, but it does go a long way to destroying it. I am thinking of running a different suite of tests with different spreads of buildings, timings and collections to build the scenario into something that more closely represents what happens in your hood normally.

The Simulation software (it's more of a test harness, really) is relatively flexible, and I plan to expand it to cover more scenarios, if anyone has any suggestions for algorithms to use to decide buildings, or any other novel idea about a scenarios to test, let me know and I will have a pop.

The code is ~ 750 lines, written in C++, using Visual Studio 2008, although uses the QT Toolkit for UI and other bits. I am happy to provide source code if anyone is interested, but onyl once I can be bothered to setup source forge :(

Deluxe
10-15-2012, 11:16 AM
Ram you make my head hurt...in a good way!

Auxilium
10-15-2012, 11:21 AM
I love to read a summary on these statistic threads. Thanks for sharing the info!

mxz
10-15-2012, 11:28 AM
Amazing.

What's SU/G?

I'm guessing income / hour decrease at times because you're taking into account specific buildings being unavailable during upgrades? Why is GFI the only one seemingly affected?

What did you see as the optimum upgrade strategy? Does that get output or logged somewhere?

Next steps, in my mind:
Allow for tweaking of collections / day.
Allow for a weighting modifiers on GFI factors.
Add a NPV curve (unless that's SU/G)

Ramshutu
10-15-2012, 11:38 AM
Amazing.

What's SU/G?

I'm guessing income / hour decrease at times because you're taking into account specific buildings being unavailable during upgrades? Why is GFI the only one seemingly affected?

What did you see as the optimum upgrade strategy? Does that get output or logged somewhere?

Next steps, in my mind:
Allow for tweaking of collections / day.
Allow for a weighting modifiers on GFI factors.
Add a NPV curve (unless that's SU/G)

S U/G is missed named, but is "Shortest Upgrade", and orders things by the smallest time to upgrade.

The income/hour decreases as I calculate the "actual" IPH rather than what the game would report. GFI is *more* affected because it prioritises BIG + LONG upgrades over short and sharp upgrades that $/hr/hr does.

The upgrade strategy is currently logged, but to my debug window :). I am working on some more framework improvements and tidy ups so I can record it.

Deluxe:

Just FYI; your suggested algorithm, roughly translated into code as "Prioritise upgrade of lowest level buildings", actually isn't too bad, it is marginally worse than using the $/hr/hr approach, but not that much. It still blows away ROI.

Silvio
10-15-2012, 07:32 PM
Ram you rock! If I could add my two cents, the reason ROI doesn't work as a good metric in this game is because it is a business metric used to maximize cash flow for things like food, rent, mortgage, entertainment (otherwise known as living).

I love the charts. Thanks for the work. I appreciate the geek inside.

Deluxe
10-15-2012, 07:39 PM
Deluxe:

Just FYI; your suggested algorithm, roughly translated into code as "Prioritise upgrade of lowest level buildings", actually isn't too bad, it is marginally worse than using the $/hr/hr approach, but not that much. It still blows away ROI.

Yay! So even tho i dont have the capacity to justify it with super sweet charts and the like, im happy to know that (what i believed was) my mathematically inclined brain can still compute rough and raw to a pretty solid approach :)

PawnXIIX
10-15-2012, 07:54 PM
The Simulation software (it's more of a test harness, really) is relatively flexible, and I plan to expand it to cover more scenarios, if anyone has any suggestions for algorithms to use to decide buildings, or any other novel idea about a scenarios to test, let me know and I will have a pop.

The code is ~ 750 lines, written in C++, using Visual Studio 2008, although uses the QT Toolkit for UI and other bits. I am happy to provide source code if anyone is interested, but only once I can be bothered to setup source forge :(

Would you mind sending me some of that code? I love seeing how other people program :)

murf
10-16-2012, 06:11 AM
Ram,

Thank you for doing all that work. Your results are interesting. I just want to be clear that the ROI strategy that you used in your simulation is NOT what I am espousing. If you want more detail on my economy strategy let me know, and I'll try and write it up.

Dipstik
10-16-2012, 06:20 AM
You made a simulator to simulate a simulation game?

Ramshutu
10-16-2012, 07:59 AM
You made a simulator to simulate a simulation game?

And played simultaneously.

Dipstik
10-16-2012, 08:17 AM
Mind = blown!

Ramshutu
10-16-2012, 01:46 PM
Would you mind sending me some of that code? I love seeing how other people program :)

Yes, but I would say that I'm in the middle of refactoring some of the bits to make it faster for some other bits I'm intending to try (Read: It's not good enough to submit for formal code review yet!) . If you remind me at the weekend, I will send it on.

In the mean time, I will leave you hanging with the following function:


void CMappedAlgorithmList::recalculateList()
{
//Use the algorithm to recalculate the weightings it will assign
//to all buildings.
// TODO: two foreach loops? WTF: Move into algorithm base, and optimise to a single operation.
m_algorithm->recalculate( *this, m_allBuildings );

//Create a temporary map that can be sorted.
QMap<double, Building*> sorted;

foreach( Building* b, m_allBuildings )
{
//Add all buildings, with weighting as the key.
//NOTE: insertMulti allows two buildings with the same weighting
//to be added into the same list (non-unique hashing).
sorted.insertMulti(b->algorithmWeighting(), b );
}

Q_ASSERT(sorted.count() == m_allBuildings.count());

//Magic: Map is sorted from low to high weighting. Simply wop out
//the list of values from the map and re-assign it into the m_allBuildings list.
m_allBuildings = sorted.values();
}

mnju_03
10-16-2012, 01:48 PM
Interesting...