Skip to Content
 

Stock market game: Can't make 2 different strategies balanced

9 replies [Last post]
Mon, 08/09/2021 - 09:24
larienna
larienna's picture
Offline
Joined: 07/28/2008

I am working on a stock market game in my free time. I am I am trying to balance the outcome of the game. I did a lot of digital testing so far, now I am at a point where I want my game to have 2 strategic path.

A: Simple path: Buy low, sell high
B: Complex path: Buy high to benefits from higher dividend by switching stocks more often.

The game use cards instead of dice (compared to St0ck T!cker) to add some predictability to the outcome.

I have 3 AI:

A: Buys the 4 lowest stock out of 8 whatever the conditions.
B: An advanced AI that uses "Multi Criteria Decision Analysis" (google it) to determine the best stock to buy.
C: Simply sell and buy random stocks each turn.

Even if I change certain game rules, the outcome is always in that ratio:

A: 100% output
B: 80% output
C: 10% output

So for example, if A ends with 50000$, B will end with 40000$ and C with 10000$. This is the average results generated from 10000 games. The good news is that doing random actions strategy does not work.

Now I would want B to be slight more efficient than A, Maybe like 125%. The reason is for new player to use strategy A and more experienced player use strategy B which is more complex to pull off. Of course, you could mix both path.

Now the problem is that I cannot make path B better. I also tested different AI ponderations to find the most performant AI, and it seems that the market value of the stock is better NOT beign considered.

So I am trying to find a solution to make that alternate path more efficient. A simple solution could be to increase even more the dividend output.

Right now the stock value starts at 100$ and moves between 10$ and 280$. The Dividend percentage range between 1% and 28%. The higher the stock the higher the dividend. But maybe the dividend % is not high enough to be worth buying high stocks. You are gaining much more by stock growth than dividend outcome.

I also added a mechanism that you cannot buy stocks lower than 50$. That prevents easy growth where buying at 10$ and selling at 50$ grows you fortune 5 times. But even with such restrictions, AI A still surpass B.

Do you have any other solutions?

Top
Mon, 08/09/2021 - 11:12
#1
let-off studios
let-off studios's picture
Offline
Joined: 02/07/2011
Stock Sale Bonuses

One way to incentivize frequent trades is to provide some kind of in-game benefit to trading more often. What options have you thought of for this?

  • More stocks to choose from once you have recorded X number of trades.
  • "Blue Chip" options or bonds available for a limited time, available only to those who have recorded a certain number of trades.
  • (on the flip-side) Penalties for holding on to stocks too long.
  • Bonuses for diversified portfolios, with stocks from several different companies.
  • Reputation bonuses for certain stock purchases, that provide bonuses to score at the end of the game.

Think about the behaviour you want to encourage, particularly for venteran players and those who pursue counter-intuitive strategies, and you might have a few more options to consider.

Top
Mon, 08/09/2021 - 13:26
#2
larienna
larienna's picture
Offline
Joined: 07/28/2008
I like the idea of

I like the idea of opportunity. Like special events that could occur that the player could benefit from if they make the move.

I still want to keep the game simple as it`s intended for non-gamers, but I`ll give it some thoughts.

Maybe something like "This stock act differently for the rest of the year". It's the popular stock.

Else higher dividends or more often dividends could be a quick solution.

Top
Tue, 08/10/2021 - 06:43
#3
X3M
X3M's picture
Offline
Joined: 10/28/2013
Administration costs

???

What if buying a lot of cheap stocks simply cost more administration costs. Thus lowering the efficiency for these?

Not sure if it makes any sense.

Top
Tue, 08/10/2021 - 10:31
#4
larienna
larienna's picture
Offline
Joined: 07/28/2008
Have a limit on the number of

Have a limit on the number of stocks owned could be an idea. A bit like in 18XX where there is only 10 stock max per company.

I made additional tests. For AI B, I tried to maximize dividends, and it seems that buying 6 different stocks instead of 4 helps maximizing dividend. Else all preponderation values for the AI gives the same results. So diversifying could help this strategy, but it might not be a more complex strategy as you could buy all stocks like many player did in the old game.

So I decided to change the AI A to buy only 2 out of 8 stocks instead of 4. and the profits increased by 250$. Which is just, WOW. It now makes it even harder to balanced.

It seems to have partially answered my question of diversification VS concentration strategies, since I also wanted to make both strategies viable.

----------------------------------------------------------------------

Now one thing I though for opportunity, is that each year (6 rounds) 3 out of 8 stocks are selected as being "hot stocks". Those hot stocks pays double dividend and generate speculation that can make the stock eventually burst. The objective is to make it a double edged sword, yes you are going to get more profit, but there is a risk the bubble will eventually burst and the stock value will drop. I'll give it a try and see the results.

The idea is to force players to change wagon once a while and prevent buying everything. It could be a way to inject more dividend money without changing the rates.

Top
Tue, 08/10/2021 - 11:29
#5
larienna
larienna's picture
Offline
Joined: 07/28/2008
By the way, I forgot to

By the way, I forgot to mention, if you like stock market games, you can play the digital prototype with your computer and a spreadsheet. It runs in Java, so you need Java to be installed. Just PM me and I'll send you a link with the password. All the files and instructions are there.

For some reason, my friend cannot run it on his Windows 10 PC which is weird since Java is supposed to be portable and I am only using the text mode, so nothing fancy required.

Top
Mon, 08/16/2021 - 09:12
#6
larienna
larienna's picture
Offline
Joined: 07/28/2008
If concentrating on a single

If concentrating on a single stock is too powerful, I could limit the number of stocks a player can own.

This could be justified to reduce component cost, instead of using cards to keep track of stocks, I could use a chart with a track ranging from let says 1 to 9 and sliding a token to mark the number of shares you have. Then you can flip the token to mark stock quantity from 10 to 19.

That would probably reduce production cost as you only need 1 chart and 8 tokens. It would also reduce the ambiguity of: "Will I print enough shares". But it will feel less like you are actually trading something.

It could also alter the experience as if a player max out his shares, he needs to invest else where. I could even prevent a share from splitting if a player would get too much shares. All those mechanism would be methods to force players to split up their assets. It feels a bit artificial, but it could do the job to reduce the power of concentration.

I'll explore a bit more and see if there are other parameters that I could change.

Top
Mon, 08/16/2021 - 10:39
#7
let-off studios
let-off studios's picture
Offline
Joined: 02/07/2011
Business Obligation

larienna wrote:
If concentrating on a single stock is too powerful, I could limit the number of stocks a player can own.
One way to enforce this and keep it streamlined and at least a -little- realistic is that the more stock someone owns in a company, the more ownership they have of that company. So someone may be limited in owning too much of the stock of a company because they cannot become a "majority partner" or something like that.

You can also address this mechanically by making stocks more expensive with each stock purchased by that player. So even though there's a market price, since someone already has a large volume of that company's stock, they need to pay an increasing percentage higher after they own 5, 10, 15, 20, etc. This way, there's no "hard limit," but there's a diminishing return - at least initially.

Top
Mon, 08/16/2021 - 11:30
#8
larienna
larienna's picture
Offline
Joined: 07/28/2008
The company majority

The company majority justification limit, seems like a good thematic explanation.

I want to keep the game relatively simple to handle, since players has to do the math. It's already a pain to balance, if I also have to add variable cost, that would be too complex.

Top
Mon, 08/16/2021 - 12:14
#9
larienna
larienna's picture
Offline
Joined: 07/28/2008
Still playing with

Still playing with variables.

I decided to change the number of cards that get drawn. Formally I drew 8 cards per turn, and 19 cards out of 67 ended up getting discarded. Since there are 8 stocks, and 2+ cards of the same stock could be drawn, it made sure that not all stocks moved every turn.

But I think the problem is that there was still too many stocks that moved each turn. So I decided to reduce the number of cards drawn to 6 and discard 31 cards. Both AI tried to buy the 3 lowest or the 3 best stocks. and surprisingly, the results flipped. Here is the results:

Player:Jimmy Low
Fortune average:       22027.928$
Nb DIV missed average: 74.53
Nb DIV hit average:    44.846
Nb Crash average:      1.15
Nb Split average:      0.026

Player:Johnny High
Fortune average:       26902.717$
Nb DIV missed average: 78.098
Nb DIV hit average:    41.278
Nb Crash average:      0.125
Nb Split average:      0.922

So I'll have to explore this change. The positive is that it would make less cards for the players to apply. But since I wanted to give each player a card as a clue on the market movement, I am limited to 6 players.

I will compare the payout according to concentration VS diversification to see how the numbers progress and see if both strategies are valid. I might try lower of higher nb of cards for optimal nb of cards draw. It makes a lot of variables to play with to get the results I want. Hope it will be fun in the end.

Top
Syndicate content


forum | by Dr. Radut