[Trader's Handbook] Draft articles, "out of pocket" discussions - page 26

[---]

GaryKa:
It would be a good idea to add a post about ticks and the main types of trade matching algorithms

I' ve been asked by sanyooooook to do this in the thread, but I don't think there is any information available.

[hrenfx]

GaryKa:
It would be a good idea to add a post on ticks and the main types of trade matching algorithms.

I don't get it.

As for the topics you've identified, write down what's in the course and what's not. Then, perhaps, I will understand which gap to fill.

[GaryKa]

hrenfx: I don't get it. Regarding the topics you've identified, write down what is now in liquor and what is not. Then maybe I'll understand which gap to fill.

• What is a tick for a trader? A change in at least one of the bands (price or volume) in the stack? A change of at least how much? How do brokers provide ticks to clients (connection to onTick event, grouping, filtering)? The difference between STP, ECN, ECN/STP ticks? You could just add this post.
• There are somealgorithms like this. If there are some implementations, maybe there will be arbitrage skills )
  

[GaryKa]

papaklass:

1. Tick is not a price, it is information about a price change. I agree, this event Tick does not show how many points the price has changed, it informs about the price change. Therefore, we cannot measure the price difference between Level2[0] and Level2[1] in ticks. I don't understand why not. We had a tick, it changed, then it changed again - 2 ticks have passed. One change of a glass is 1 tick. In the phrase "Wecompare Level2[0] and Level2[1].If Bid[0] >= Ask[1] " Bid[0] and Ask[1] are best prices on neighbouring ticks.

The author is here and he can explain more precisely what he meant. I personally understood it this way.

[hrenfx]

GaryKa:

• What is a tick for a trader? A change in at least one of the bands (price or volume) in the stack? A change of at least how much? How do brokers provide ticks to clients (connection to onTick event, grouping, filtering)? The difference between STP, ECN, ECN/STP ticks? You could just add this post.

You seem to have overdone it here. All in one pile. Either I didn't understand you again.

• Thetrading algorithms are something like this. If there are some implementations, perhaps there will be some arbitrage masters.)
  

I didn't understand what to add here. The links give some reports/presentations that can be fit into just a few simple paragraphs. And those paragraphs are available in the libs.

Exchange:

Forum on trading, automated trading systems and testing trading strategies

Trader's guide: Orders, Prices, Brokerage, Funds, Currencies

hrenfx, 2013.05.30 12:45

Execution of Limit orders on exchange

The correct exchange algorithm does not allow for the Bid >= Ask situation in public pricing. In the algorithm itself, as orders are received at the initial stage, a cup is formed in which the Bid >= Ask situation often occurs. In such a situation, the executive part of the exchange algorithm is activated, whose task is to resolve the situation to the state Ask > Bid. And only after breaking it up, the formed market stack with correspondingly formed Last-data becomes public - available to all.

SellLimit is always executed at the Bid price, BuyLimit is always executed at the Ask price.
But only these Bid and Ask are not public prices of the formed at the initial stage of the cup as described above.

Forex:

Forum on trading, automated trading systems and trading strategy testing

Traders guide: orders, prices, charts, funds, currencies

hrenfx, 2013.06.10 11:18

Let's analyze the algorithm of execution using the example above:

At the moment Bid >= SellLimit, the virtual SellLimit is frozen (removed from the accounting by the execution system, the client cannot do anything with it) and the same SellLimit is sent to the corresponding LP. All these actions have no effect on pricing. That is, a little later than freezing (for example, by 1 ms) a better Bid can be formed (from other LPs), but it almost (there are nuances) will not participate in execution of our client's SellLimit.

The LP where the SellLimit was sent replies that it executed one part of the SellLimit and did not execute the remainder for various reasons - a reject. After that the client receives a high-level open SELL position for the executed volume and an unfrozen SellLimit for the remaining volume.

The speed and quality of LP response depends on many factors. The response time may be up to several seconds. Situations where there is no response from LP are not excluded.

Note, (with this implementation of STP) aggregator clients' limits do not slide into the negative zone. Moreover, there are frequent positive slippages, which cover a significant part of the fixed trading costs - commission.

Here you can see how much the execution and pricing differ between exchanges and FOREX.

[GaryKa]

hrenfx: ... I do not understand what to add here. The links give some reports/presentations, which can be fit into just a few simple paragraphs. And these paragraphs are available in the libs.

Rather a slightly ambiguous title (the other being trade allocation algorithms)
Recently came across it, googled it, found it interesting, asked.

I will try to explain in my own words - algorithms that determine which orders and in what volume a market order will cover on a certain gang, if the remaining unfulfilled volume of the market order is less than the total volume on the given gang. For me personally it was interesting to learn that in some cases (depending on which of these algorithms is used) the trader does not have to "get into a band" to obtain liquidity.

The trailer, although a bit old, is an adequate document that spells everything out. In the link above there were just easy to understand drawings.
But it is not for me, a layman, to decide what should be in the lyknobes. I totally agree.

In a good way, it is better not to copy theorists. And ask practitioners to write. With all due respect to forum users, there are very few hardened practitioners. It would be better if they would write. And incomprehension develop in the form of questions, not arguments. For a known strong practitioner knows a lot more and deeper than what is written in most literature, which is where most of the knowledge is drawn from.

P.S. About the ticks I'll answer later.

[hrenfx]

GaryKa:
Probably a bit ambiguous name (the other is trade allocation algorithms)

Wrote. Now we can make a little observation:

Forum on trading, automated trading systems and testing trading strategies

Trader's Guide: Orders, Prices, Brokerage, Funds, Currencies

hrenfx, 2013.06.10 11:18

Nuances of execution.

It is very difficult to describe the many subtleties that affect the quality of execution. The simplest thing that the aggregators do one of the first is to improve the communication channels between LPs. however, what determines the execution quality (FillRate) are the algorithms of the STP-aggregator execution. These are a whole class of constantly improving (not by everyone) interesting algorithms, some of which even affect STP aggregator pricing. They are largely unaffiliated competitive solutions and very specific to the trader's handbook, so I probably won't tell you about them.

The fact is that two STP aggregators on an equal footing can differ significantly in FillRate, due to the use of different execution algorithms. This can be especially noticeable when toxic. I.e. the performance of a trading strategy can be very much dependent on the aggregator used, especially when trading volumes are increasing.

If the trade values of two identical TS, running under equal conditions on the same aggregator, are the same, it is an indirect evidence of a very high quality of the execution algorithms of that aggregator.

The underlined suggestion implies, in particular, Pro-Rata allocation.

[GaryKa]

papaklass:

I get the impression that you are trying to get the point across that execution of trades does not necessarily change best prices (I know that). And so the "inter-tick" analysis of Bid and Ask is inadequate.

Now, I ask you to pay attention to the title of the post"Rough algorithm for getting T&S data from Level2" and to the phrase below

This algorithm is easy to check on the linkage to the official T&S data provided by the exchanges. Then make some improvements through unequal gang distributions etc.

What does it mean? That T&S (or rather its estimate) caught by this algorithm is roughly proportional to the true T&S at the given site. In other words - yes we cannot catch all of the execution (for example deals in your examples), but what we will find (by the analysis of best prices at adjacent ticks) can tell us about the nature of the execution. For example using this algorithm, we notice that T&S estimated using this algorithm in the middle of the trading session is more than before its closing. It means that the real T&S behaves the same way.

P.S. For my part, I'm going to quit the discussion on this subject until the author gives his comments. It is quite possible that hrenfx wrote about one thing, you argue about another, while I speak for the third (it is not clear to me everything with STP too, so it is quite possible that I misunderstood with ECN too).

[hrenfx]

Honestly, I didn't set out to figure out who understood the post about Level2 -> T&S. Therefore, not on the topic of the argument.

Perhaps you should read the post again carefully.

Or maybe I just don't see possible ambiguous interpretations in some places there. Then it is worth pointing out those places to make clarifications.

[hrenfx]

Offtop, but a clear example that we don't ask ourselves much of the childish question "why?"

Forum on trading, automated trading systems and trading strategy testing

Market Laws

hrenfx, 2013.07.15 13:37

Recently explained multiplication in an accessible way to a child. And he showed the proof that 2 * 2 = 4:
Multiplication is an asterisk (*) operation with the following properties:

1. 1 * a = a.
2. a * b = b * a.
3. (a + b) * c = a * c + b * c.

Let us prove the statement that 2 * 2 = 4:

2 * 2 =
(1 + 1) * 2 = apply item 3
= 1 * 2 + 1 * 2 = apply item 1
= 2 + 2 = 4.