Login
Ratjoy.com » Forums » EoS Club Rats Den » Store Balance

Store Balance


Previous 1 2 3 4 [5] 6 7 8 9 10 Next
Innocent Bystander
RJ: Matthew Matician

Post Rating: 0
+ / -

Total Posts: 76
Karma: 54
Joined: Apr 2, 2012
I think float/double would be great for quality, but I'd recommend against using it for money if the reason is to store numbers that are too large for an int-like type. Mostly because if the number is so large that a 64-bit signed int can't store it, the difference between consecutive numbers stored in a 'double' format will be at least around 2000 or more. (If it's for some other reason, then ignore my post. :)
Scott (Admin)
RJ: Ratan Joyce
CO: Ratan Joyce

Post Rating: 0
+ / -

Total Posts: 1175
Karma: 5083
Joined: Jan 13, 2012
Yup, it's for the same exact reason - I was just thinking if your $ is up in the 64-bit range, then certainly you don't care about losing/gaining $2000 here and there.

The 1-quality-sale compromise sounds reasonable too me, but I'll probably need more votes on this issue.
Mister Death
RJ: McFlono McFloninoo

Post Rating: 0
+ / -

Total Posts: 266
Karma: 300
Joined: Feb 6, 2012
I doubt that the difference between integer and floating-point operations is going to seriously impact performance, given that the vast majority of CPU time in a game such as this goes to memory/disk retrieval and storage.
Cian Kemp
RJ: Cian Kemp
CO: Cian Kemp

Post Rating: 0
+ / -

Total Posts: 183
Karma: 58
Joined: Apr 9, 2012
How about make it 1 quality for each slot, and have an option to auto-mix everything you buy? It could be a company option in the settings page. That way people just running retail businesses can turn it on and simplify their life, while people who want to keep things separate can turn it off and keep multiple stacks of different qualities. I'd probably turn it on in my retail store companies, I just wouldn't want it for my electronics company since I strive for high quality :)
Andrew Turner
RJ: Thomas Lazygun

Post Rating: 4
+ / -

Total Posts: 44
Karma: 26
Joined: Apr 9, 2012
I like Innocent Bystander Matthew Matician's approach. I really want to see different qualities kept in the game.

And consider this: how often in the real world do you see a supermarket shelf empty apart from a handful of goods priced ridiculously high. But that happens all the time in EoS because the only alternative is having the shelf completely empty and generating no revenue at all. In the real world, shelves aren't "use it or lose it" but a vital resource that must be managed. When an item is no longer available from the wholesaler, the store owner will price it low, ship it to a discount store, hold a stock clearance sale, do whatever it takes to *get it off the shelves so the shelves can be used to stock something else*.

An alternative model, that I believe warrants consideration, is that that stores have a limited number of shelves (or 'cabinets' or 'slots' or 'bays') that is just a little too low to sell *everything* that the store sells. If you're selling two different qualities of something that takes up two of your sales slots, and furthermore gives a mild penalty to your sales of that item.
Fish Bike
RJ: FishBike
CO: FishBike

Post Rating: 0
+ / -

Total Posts: 22
Karma: 51
Joined: Mar 29, 2012
The 1-quality-sale compromise sounds reasonable too me, but I'll probably need more votes on this issue.

I'd prefer this to the "quality averaging warehouse" design. One problem with averaging item quality is if you're improving the quality of items you make, it'll take basically forever for your inventory to actually reach the higher quality level unless you dump all your stock first. It seems more logical to me that you would have lots of varying quality in the warehouse and can decide which to sell in your store, which to put on B2B, etc.

It also seems like a less disruptive change, as it's only a change to the store and not to inventory and every place quality is used. Easy to undo if necessary, whereas once inventory is all mixed together, it's going to be pretty much impossible to go back to how it was.

I may also be biased by a form of PTSD I have from working with an inventory control system that was supposed to do the accounting based on the "average cost" method that never quite worked correctly. I spent many hours looking through transaction histories and verifying calculations by hand to figure out where it had gone wrong...
Karl Katzke
RJ: SpecialK

Post Rating: 0
+ / -

Total Posts: 3
Karma: 10
Joined: Mar 29, 2012
I agree again that I'd prefer to be able to have different qualities of stuff to store in my warehouse and to have on B2B market. I think managing quality is an essential part of the game, and you can't do it if it's all one number.

Now, only having one quality for offer at your store makes sense ... although it does require a bit of a suspension of disbelief because you'd normally in a normal retail store find different qualities at different price points...

Throughout all of this discussion with store balance, especially with a retail business model making an incredible amount of sense yet the game admin penalizing people who are following that model... I still haven't figured out if EoS is a retail simulator or a manufacturing simulator or a supply chain simulator.
Adam Snider
RJ: Adam Snider
CO: Sad Adder

Post Rating: 0
+ / -

Total Posts: 7
Karma: 10
Joined: Apr 3, 2012
I echo Innocent Bystander and Fish Bike.
Alexia Perdhaer
RJ: Alexia Perdhaer

Post Rating: 0
+ / -

Total Posts: 100
Karma: 30
Joined: Apr 6, 2012
@scott: is there not a big decimal type in PHP, or some PHP library? It works a lot better than float for money, and is natively supported by databases.
Alexia Perdhaer
RJ: Alexia Perdhaer

Post Rating: 0
+ / -

Total Posts: 100
Karma: 30
Joined: Apr 6, 2012
there is http://www.php.net/manual/en/ref.bc.php which might be less than ideally performant as it uses strings as it's underlying representation.
Scott (Admin)
RJ: Ratan Joyce
CO: Ratan Joyce

Post Rating: 31
+ / -

Total Posts: 1175
Karma: 5083
Joined: Jan 13, 2012
I just realized Thomas Lazygun's comment would solve everything, and make the game more fun, literally.

An alternative model, that I believe warrants consideration, is that that stores have a limited number of shelves (or 'cabinets' or 'slots' or 'bays') that is just a little too low to sell *everything* that the store sells.

Using this approach, I can have all shops start with 4 shelves, and in addition to bigger size generating more sales, each 100 m^2 can add 1 more shelf, up to 20.

Pros:
1. No more going out to buy everything - much less micro-management.
2. Rewards specialization, you can assign the same Q50 Wines in 5 different slots if needed.
3. Still allows simplified store interface.
4. And it makes sense to me.
5. Less bloat generated per tick in the DB.
6. Same products of different qualities will not compete with one another, therefore reduce calculations.
7. All stores get automatically balanced without needing to balance each. (The total number of possible products is no longer as important.)

Cons:
1. ... Less impressive pop-up?
2. You will need a way to get rid of all those products in your warehouse.
3. I'll need to make a minor change to the equation to increase sales.

If this is implemented... then I'll have to say: d*mn, I can't wait for the new store interface!
Jim Frazer
RJ: Kenneth Noisewater
CO: Kenneth Noisewater

Post Rating: 0
+ / -

Total Posts: 25
Karma: 31
Joined: Apr 14, 2012
Damn, that's a great idea. I think that would make things so much more interesting and less micro-manage-y. Plus, makes for some interesting decisions. Do I build 3 supermarkets and have them each stock different items, or do I make a giant supermarket that sells less variety but more volume.
Innocent Bystander
RJ: Matthew Matician

Post Rating: 0
+ / -

Total Posts: 76
Karma: 54
Joined: Apr 2, 2012
I like the idea of being able to diversify strategies (sell all or just a few), it seems similar in spirit to Tex's idea from a few weeks back.

A few thoughts that come to mind:
1) Instead of having slots where you can place items, perhaps a text box next to each item for you to enter a relative weight for how much shelf space you want to give it in your store(s) you're devoting to that item (maybe have a "reset all items to equal" button up top). The rationale for this would be that it would prevent having to restock or fiddle with the same item in a bunch of slots: you could just assign that item a high weight and be done.

2) I don't think that larger stores need another advantage over smaller ones (ie: have more slots to sell things in)--if we did the % of store for each item, that would effectively give more shelf space to items as your store size went up without having to do discrete slots for items. This means you don't have to worry about how all the stores that have less than 4+20 expandable slots worth of items get handled.

3) Whatever way you decide to go, I would implore you to make everything (except perhaps advertising, expanding and selling) all do-able without having to leave the page (or perhaps the option to do it all from one page and not have to do pop-ups). Kind of like the warehouse or B2B pages work. I'm not sure what information others like to have handy, but my wishlist of things to include on main page for each product (table-style is fine with me, I don't need pretty :) would be: cost,price,quality of item, # sold in last day.

4) However you decide to do it, I also really hope that one aspect stays the same as now, which is that all stores of type X share the same price, shelving options, etc. Not having to monkey with every shop of the same type is much appreciated.
Scott (Admin)
RJ: Ratan Joyce
CO: Ratan Joyce

Post Rating: 0
+ / -

Total Posts: 1175
Karma: 5083
Joined: Jan 13, 2012
Response:
1 & 2. I'd still have to place limits on the available slots. (I'd like to delay sharding the DB for as long as I can)

3. Of course, that's one of the main goals

4. That's doable.
Andrew Turner
RJ: Thomas Lazygun

Post Rating: 0
+ / -

Total Posts: 44
Karma: 26
Joined: Apr 9, 2012
Scott, how much of the sales calculation can you pre-compute? How feasible would it be to have your tick processing being something like 'update warehouse set stocks = stocks - precomputed_sales' ? You'd end up doing more work overall but it wouldn't all be concentrated at the tick.
Previous 1 2 3 4 [5] 6 7 8 9 10 Next


You need to register or login to post a reply.