The Trading Game - Development Log #346

One thing I’d like to point out is that the PrUn differs in one big way from IRL markets: orders cannot be placed algorithmically. Penny wars can unfold much faster if computers are doing the bidding vs humans. The limiting factor for how fast the prices converge in a game penny war are either how much time the involved parties are willing to spend on it, the fees imposed upon placing a new order or a hard limit on the number of orders that can be created. Those are generally quite high hurdles for commodities which have a high spread, resulting in very slow price convergence.

I am with Lowstrife. I think the slow price convergence is due mainly to low volumes in general; it has very little to do with minimum price change increments.

If you want to increase volume on the CX, then I maintain PrUn is not the game for you; the only way to increase CX volume is to eliminate LMs and that would just be a different game entirely.

Yeah without HFT, we’re never going to get 1bp spreads and infinite amounts of liquidity like we see in other markets.

image

The opinions I express are just opinions I think which would improve the health of the markets and make the most bester changes for the betterment of the mostest number of people.

Some individuals may not see beneficial changes for themselves, but as a whole, the health of the markets would improve for everyone. We’d only really see these changes impact the T1 markets. For the higher end stuff, most players already ignore the penny bid system anyway. People meaningfully operate on the 3 or 2 sigfig rule. Because it’s not about being “first” in line. It’s about offering a meaningfully cheaper price so that someone actually buys it at the cheaper price, when before, it may have been too expensive.

2 Likes

The reason people don’t like penny wars is that it’s demoralising if you’re a beginner player who logs in once a day, and you finally get your BSE to the CX and list it for sale, and then the next day when you log in you find someone changed their listing to be a penny less than yours, so you haven’t sold anything. It favours players who have more time to monitor the markets all day and adjust their bids.

The 5 order per 24hr limit helps, but a minimum tick size would also help, because the underbidding players would be forced to lose more money when they did it, so then underbidding wouldn’t be such an easy decision for them. That’s why players are asking for minimum tick sizes, it has nothing to do with liquidity or price convergence, even if those things would happen as a result.

3 Likes

You’re precisely right with how you demonstrate the problem.

You will see actual liquidity develop in the orderbook. You will see larger orders be more lazy, hanging out deeper into the book happy to wait. And you will see smaller orders torward the front of the orderbook. Now and then you may see a larger player move a large block and adjust the price.

But you’ll have compressed spreads. And you won’t have to actually be first in line to have any hope of getting filled. You can throw an order “lolidksomewhere” and it should have a better chance of not getting bypassed and forgotten about.

The more players and the more trading activity there are in the game the better this works. The game is… still small now. So the markets are pretty baren still.

1 Like

Won’t increasing the tick size mean more goods going through MMs instead of to real players? Tick size won’t stop me listing and re-listing at any price. So far I haven’t had to use an MM ever, but I can see it happening for sure if the tick sizes increase. To do stuff like HQ and ship upgrades, or build your corp headquarters you need revenue. If I sit around waiting for my goods to hopefully sell sometime in the future those things will take too long, and the game will become boring. I see penny-wars as players being pro-active so their revenues don’t stagnate. As far as demoralizing new players is concerned, that slow early growth stage we all had to go through, use it to explore and learn how the game works.

Tick sizes have nothing to do with the marketmakers. They impact the pricing and efficiency of the market - not the overall supply and demand.

If an item has no buyers or sellers and runs into the MM - that situation has nothing to do with whether the tick size was 3 or 5 sig figs.

1 Like

I think spread is a huge issue, it makes people determined to sell to sell order and buy to buy orders, because of FOMO.

I reckon making it easier for people to flip or update order, will close the spread which encourage people sell to buy order and buy to sell orders. I spent a week flip items on market, I have noticed a this in play.

It makes me very happy that you respect the open market.

The tick size idea is alright but given there is a limit on delete orders already I don’t feel it is necessary and is actually a risk - if you concede on this point and implement regulation, you’re more likely to be pressured to concede again. This would not be conceding if it replaced (or increased) the delete order limit, as then the level of control in the market is maintained.

Finally, the point about “but players with more time still have an advantage” is technically correct, but why is this bad? It is an organic property of a market - increasing the effort of your participation (time) can result in an increased reward.

Because PU is a game catered to people who don’t have the time to spend hours in front of the game every day. As the game is right now, whether you spend 30 minutes or 10 hours a day in-game will not affect your results by a lot. I assume the devs want the game to stay like that (else they wouldn’t have made it like that in the first place).

1 Like

I agree with everything you’ve said.

As the game is right now, whether you spend 30 minutes or 10 hours a day in-game will not affect your results by a lot.

Given the above, playing longer is a non-issue, but has been used as an argument for implementing tick banding. Given we already have a limit of 5 order deletions - this change doesn’t feel necessary.

My wider views on implementing any controls or changes to the market are in my original post.

Tick sizes are superior to a deletion limit, though.
If you had to have one of them, which one would it be? Which one is less annoying?

Tick sizes just feel much more natural to me and much less limiting. And given that we will have one of either (presumably they would remove the deletion limit), I support tick sizes in that particular context.

But also, there are other arguments supporting tick sizes. The main ones are faster movement of prices due to a wider distribution of prices across the order book. Which, arguably, is better even if you place it on top of the deletion limit. There’s a reason exchanges IRL use tick sizes, too, and I assume it’s not due to regulation, although I might be wrong.

Edit: After a bit of googling it turns out I probably am wrong… so nvm what I said regarding “theres a reason” lol. So ignore the part about it arguably being better even on top.

1 Like

Tick sizes are superior to a deletion limit, though.
If you had to have one of them, which one would it be? Which one is less annoying?

The delete limit is very effective and stops the penny under-bidding pretty quickly. It also doesn’t mess with the volatility of the market. Tick banding may be less effective, as those willing will simply underbid by larger units of currency, and the volatility of the market will increase.

However it’s not that I fully don’t like idea - in my original post I proposed we opt to replace (largely or fully) one form of market control with another, if the new control is deemed completely necessary. This would be either increasing the delete limit or removing it, in place of the tick banding.

I’m coming at this from a very high-level point of view - the robust choice when altering any (complex) system is to “remove” and not “add”. When we do add to a system it becomes more fragile and the chance of us adding more increases (particularly as there are more players) as our allegiance to ideas of purity (i.e. open unregulated market) become less solid.

If we don’t do the replacement, this is a big “add” - and given it is in response to a perceived problem in the market (penny under-bidding) it becomes a regulation to firstly prevent certain forms of market behaviour and only secondly to benefit the market as a whole.

Even the statement “benefit the market as a whole” is so subjective that it should be avoided in any decision making.

1 Like

It’s not that clear.

Median volatility is slightly lower with the smaller tick size, and there is no
evidence of changes in the extent to which price changes are subsequently reversed

Now this is in real markets.

In my opinion, for us in PRUN, volatility will go down. Not up.

Why would people offering better prices sooner than they otherwise would have cause volatility to go up?

Another piece of information about tick sizes causing spreads to be tighter and causing them to collapse further

This analysis allows for a substantially
larger effective sample size as compared to before-versus-after studies of marketwide tick size reductions, and allows for separate examination of tick size increases
and decreases. The empirical results indicate that bid–ask spreads are decreased
by three to five cents per share with the smaller tick size and provide no evidence
of a reduction in liquidity.

Now this is just one paper, thankfully done before electronic trading and HFT was systemic. But it puts some data points here.

This is why stocks well, A) do splits, but B) have specific tick sizes they operate in. That’s why Tesla doesn’t trade at $297.17573845834587. You pick a point relative to the volatility where you limit the precision of orders. Otherwise people will cluster up at one point and artificially keep the spread wide. When you want the spread to be narrow and tight, a healthy market.

4 Likes

Interesting @lowstrife and thanks for linking the paper.

Obviously if those that penny-underbid (and/or want to shift their goods quickly) were to $1 underbid, the price may fluctuate (volatility) to a larger extent that the current PU market.

For example, current bids on DW:

  • 50.00 → 49.99 → 49.98 → 49.97

Bids with example tick banding:

  • 50 → 49 → 48 → 47

However what I hadn’t considered was the positive effect it would have on spreads (i.e. making them tighter) as you’ve outlined. I also had not considered that the volatility is capped once it reaches the highest/lowest opposite order.

Anyway the wider (philosophical/high-level) points about implementing market changes/controls still stand and is my primary concern.

Am I being stupid? It seems to me the paper says “smaller tick size = tighter spreads”, which is basically equal to “no tick size = tighter spreads” in PU since no tick size is the smallest tick size there is.

Unless the argument for introducing ticks into PU is to create larger spreads, which would be pretty unique, that doesn’t seem to bring home a lot, no?

Regarding volatility: It will undoubtedly go up. That’s the whole point, I thought. To increase the speed in which prices change.

Well by definition, yes, smaller tick size = tighter spreads.

If you have a two markets, both are valued at $100. One has $1 ticks, one has $0.01 ticks.

As is quite common for low IV (implied volatility) stocks, spreads will collapse and people will be offering on both sides of the spread. As close as possible as things can be without matching.

In market one, you will have bids at $100 and asks at $101.

In the other market, you will have bids at $100.01 and asks at $100.02.

Technically, yes, the spread is “tighter” between the two in absolute terms. But there is a lot more that goes into the mechanism than just the spread.

Here’s the catch though. In volatile markets, where the spread might look something more like this:

Bids at $99 and asks at $101

Bids at $99.01 and asks at $101.01

Once volatility is high enough, like we see here in PRUN, the effective spread, regardless of tick size, increases. That’s why we have a bid at $55 and asks at $65. That spread will be maintained across both markets, all else being equal. However, any change in price dictated by someone wanting to offer a better price actually moves the price by $1 instead of $0.01. This means prices “react” and “correct” faster, increasing the speed at which prices match and someone is willing to accept someone else’s price offer instead of throwing up their own bid.

This also ties into implied liquidity or whatever it’s called. Here’s a great example:

The orderbook is as we are used to it here in PRUN. But the green and red lines are a graphed visualization of the orderbook.

The main point I’d like to make here is this concept - if you have no restrictions on tick size, and the volatility of the market is far lower than the spread that is maintained by the tick size, you will see events like this. Where all of the liquidity is focused up directly at the front of the orderbook, but there is nothing behind it. Notice how the orderbook behind this large cluster of orders is completely empty.

This is in contrast to what normal, healthy orderbooks look like:

image

Regarding volatility: It will undoubtedly go up. That’s the whole point, I thought. To increase the speed in which prices change.

Now this is quite interesting. I don’t think we can count prices bouncing back and forth between the best bid and ask to be “volatility”. Volatility, in my mind, is a meaningful change in the offering and structural price difference in the market. If a market is just bouncing back and forth between bid at $55 and asks at $65, that is a lot different than a market with very tight spreads moving from $55, to $65.

We’re never going to have penny-tight spreads in PRUN. But I think we can make meaningful change in some of the more extreme examples by implementing measures that will improve the health of the market for everyone.


Now let us look at actual PRUN markets.

Here is Bfabs and SF. Both of these orderbook snapshots are typical for moria, and they are also very similar in the total # of orders they both have.

I then graphed the spreads and the orderbook structure of these two markets. I formatted the orderbook so that it looks the same as the previous picture. Asks ontop, bids on bottom.

The numbers on the axis are only a guide to show where you are - we’re looking at the dots and the slope of each orderbook as it’s graphed here.

Please excuse the 90 degree rotation, I didn’t want to slam my face into Excel to figure out how to rotate the axis points of my data.

Now let’s take a look at the SF market.

A few things I’d like to point out.

Notice how each of the dots represent an order? Notice how there are fewer “clusters” of dots at exactly the same price? They are far more spread out across a range of prices.

And note how so many of the orders are clustered much much closer to the spread, which in itself is far smaller?

And notice how these is a far more even distribution of orders in the SF market, with a far more linear slope across a range of prices with less “Stair stepping” as we see in the BFAB market?

This is a sign of a much healthier market.

Why is the SF market much healthier? Because it has a far larger effective tick size resulting in meaningful price differences anytime anyone wants to be closer to the front of the orderbook. This, over time, tightens the spread in this real-world example, and offers a more even distribution of supply & demand across a range of prices in a healthier manner.

4 Likes

I think your market depth charts (at least that’s what a quick google told me it’s called) are exactly what we need. It’s a quick way - as you said yourself - to see how healthy a market is. Too often we don’t realize how thin an order book is till it’s too late. Unfortunately our growth is so fast, there isn’t enough reason to player with long term offers and bids (IMHO).

Final addendum that I’m going to have as another post since it’s topically different

That post is just the technical opinion I have on the matter. It doesn’t address the philosophical or “Free market should be free” aspect of things. I think because of my real profession (can you guess what it is by my knowledge here? lol) I’m a bit too “close” to be objective about it. But I guess, to be clear, I’m a proponent of (specific) market controls.

Ok here’s an example.

We lock the ability for users to place wildly mis-priced orders. You need PRO and to confirm you actually want to before it’s allowed. This is generally a good thing, it improves the health of the market to make it less likely someone will accidentally nuke their entire supply ot 20,000 hydrogen entirely into the marketmaker. Or vice versa buying all available supply of Carbon.

Fat-finger protection is a real-world item and it’s very much welcome here. Even though it’s magic-space money, it’s still not fun to accidentally in a big way like that.

Perhaps tick sizes can be seen in the same light.

1 Like

I would also help to keep a healthy market if more people flipped items, will drastically close the game as people keep under cutting them.