*5.4. Market Quotation Information*

The information-disclosure mechanism is very important to increase market transparency, reduce transaction costs, improve market fairness, and enhance power-system security [44]. In the bidding process of the CBT, market participants can obtain not only their own private information, such as the submitted quantity price, but also the public information of the market, i.e., market information. The real-time display of market information is one of the major features of the CBT in the continuous bidding stage. Normative and transparent information disclosure can weaken information asymmetry. The market information of the CBT mainly includes top of book, market summary, last trades, and candlestick charts, which are e ffective guides for market participants to make bidding decisions. The market information is explained below, as shown in Figure 7.

**Figure 7.** Schematic diagram of market information.

Top of book: all the bids are queued in an order book. The five best bids or asking prices submitted to the market are listed on the publicly observable board. The buyers' bids in the order book are arranged in descending order of price, while the sellers' bids are sorted in ascending order (the higher the buyer's price is, the more likely the transaction will be caused by the lower the seller's price is).

Market summary: during the bidding process, the market summary can be disclosed based on the latest real-time market data, which include the number of orders received, the total trade volume, the average hammer price, etc.

Last trades: the last trades are listed in a contract table. The disclosure information is restricted to the paired transaction time, the clearing price, and the volume for each contract. The trades in the table are sorted in decreasing order of matching time.

Candlestick charts: as trade progresses, candlestick charts for different timescales, such as 5 min and 10 min, will be drawn based on cumulative data. In practice, the relationship among opening, high, low, and closing prices in a continuous trading process is utilized by proponents of candlestick charts, to predict price movements.

All the abovementioned information is updated when there is a new order or trade.

#### **6. Case Study and Application**

#### *6.1. Results and Analysis for CBT*

To illustrate the effect and characteristics of CBT, take the power energy trading market in November 2018 as an example. Its process and schedule are shown in Figure 8. The bidding period of the first stage was 09:00–09:20, and that of the second stage was 09:30–12:00. A total of 45 CONCOs and 34 GENCOs participated in this transaction. The bidding and clearing of this transaction are described below.

**Figure 8.** Process and schedule of the CBT in November 2018.

A total of 81 orders (including six orders withdrawn) were declared during the first stage, among which the CONCOs declared 101.929 GWh of electricity, with an average price of 0.19867 yuan/kWh. The declared electric quantity of GENCOs was 2774.818 GWh, with an average price of 0.22544 yuan/kWh. The market bidding data also reflected the characteristics of the YNEM, where supply far exceeds demand. However, during the process of the first stage, the market information was not transparent, which was precious for the bidding-strategy adjustment. The one-round auction made it difficult to realize full competition. As Figure 9 illustrates, buyers and sellers had grea<sup>t</sup> differences in price intention. On the demand side, the bid price changed greatly and dispersedly, the bid volume was relatively small, and the demand curve was steep. On the supply side, hydropower plants stored water to high reservoir levels after the wet season, and bid price was usually higher at this point. However, for such hydropower stations with poor regulating performance, they were willing to declare a lower price, to improve the probability of winning the bid, in order to increase its power generation and reduce the risk of water abandonment. The transaction volume of this stage was only 14 GWh, and the average transaction price was 0.21380 yuan/kWh. A total of 24 contracts were concluded, and the rate of trade for the users was only 13.74%.

**Figure 9.** Supply-and-demand curves of two stages of CBT.

Before the beginning of the second stage, the clearing price and contract volume of the first stage were announced, which played a certain price-guiding role in the subsequent bidding. During the second stage, a total of 248 orders (including 100 orders withdrawn) were declared, among which the cumulative e ffective order quantity of 124.05771 GWh was placed by CONCOs, and the average price was 0.21562 yuan/kWh. The e ffective order quantity of GENCOs was 83.444 GWh, with an average price of 0.21839 yuan/kWh. In contrast to the first stage, the market information was disclosed in real time in the continuous bidding process, and the market participants could directly respond to the price signals through their trading decisions. Figure 9 shows that the di fference in price intention between buyers and sellers was much smaller than previously. As opposed to the one-shot nature of the first stage's auction, the transaction intentions were reached gradually through the strategy adjustment of both sides in the process of constant gaming until market closure. As the deal closed, the accumulated transaction power reached 93.99562 GWh, the average transaction price was 0.21770 yuan/kWh, 83 contracts were reached, and the deal rate for the user was up to 75.77%.

Finally, all the two-stage unconstrained transaction results passed the security check by YNPG.

According to the two-stage clearing model of the CBT, the social welfare of the first stage and the second stage were 106,519 yuan and 14,880.08 yuan, respectively. Although the volume of matched deals in the second stage was far more than that in the first stage, the social welfare was less than that in the first stage. The reasons accounting for this phenomenon mainly included the following two aspects: (1) due to the high transparency of price in the second stages, the price willingness of both parties was gradually close to each other's, and the DOP of the two matched sides was small or even zero. (2) The second stage was a continuous transaction based on the principle with first come, first serve, which existed the loss of allocative efficiency. In other words, the previously matched result maybe lost the greater social welfare that was possibly brought by the subsequent bidding [45].

To further illustrate the dynamic bidding process in the continuous bidding stage, i.e., the second stage, a large user (user A) was taken as an example to describe the dynamic bidding and game process based on market information. Figure 10 shows the average transaction price during the period of the entire market used to indicate the market price signal. The bidding price and trading volume of user A, which are drawn in Figure 10, represent its bidding strategy and strategic results, respectively.

**Figure 10.** The dynamic bidding and game process of user A.

After the market was opened, the first hammer price of the whole market was approximately 0.218 yuan/kWh, and user A failed to win the bid for the first order, with a price of 0.2 yuan/kWh. In the early period of the market (approximately 09:30–10:00), the average transaction price fluctuated in the price range of 0.215–0.218 yuan/kWh; however, the highest bid of user A was only 0.205 yuan/kWh, so the contract was not reached. Then, the average price showed a large drop (approximately 10:10); at this time, user A had raised the price to 0.209 yuan/kWh but did not take the initiative and still did not win the bid. Subsequently, user A continued to adjust the bidding strategy but failed to satisfy the conditions for the conclusion of the business. The average transaction price was relatively stable at 0.217 yuan/kWh in the middle term of the market (approximately 10:25–11:00), but user A did not bid. As the market gradually entered the later period, the user significantly increased the order price to 0.217 yuan/kWh and won the bid for the first time, reaching a contract volume of nearly 2 GWh. Later, the quotation price of user A was steadily above 0.218 yuan/kWh, winning the bid many times, and the transaction contracts were concluded. With the deadline of the continuous bidding stage, the final cumulative transaction volume of user A was fixed at 46 GWh.

It can be seen from the above description that, in the whole continuous bidding stage, market participants can flexibly adjust their bidding strategies and conduct a dynamic game with the market itself by considering their own resource endowment, risk preference, and other factors to reach trading intention and maximize their own interests.

According to the above two aspects of the analysis, the combination of the two bidding stages accelerates the discovery of market prices. The real-time disclosure of market information weakens information asymmetry. Market players adjust their strategies based on market information, and dynamic games promote full competition in the market. Continuous clearing increases transaction efficiency.

## *6.2. Overall Application and Prospect*

The CETP has been in operation since the end of 2017, so it supports the effective implementation of the CBT. By December 2018, 12 CBTs were successfully launched in the YNEM. A total of 232 CONCOs participated in the bidding, with a total of 2178 Gigawatt hours (GWh) declared. A total of 167 GENCOs participated in the declaration, with a total of 32196 GWh declared [7]. In the whole year of 2018, a total of 5278 orders were processed, 5136 CBT contracts were concluded, and the total volume of the energy cleared was up to 1018 GWh.

Table 1 shows aggregated data from the CETP in 2018, including the total volume and average price of each transaction. Data analysis shows that the transaction size of the CBT varies from month to month, which is a ffected by the primary energy supply, market demand, and other trading varieties within a month. As a hydro-dominated electricity market, the YNEM has abundant incoming water in the wet season (from June to October in Yunnan), and the supply far exceeds the market demand. The typical buyer's market causes trading prices to be lowered to "the floor price", defined by trading rules (0.13 yuan/kWh) in many months of the wet season. In addition, compared with the dry season, most of the market demand has been met by other varieties (for example, the bilateral negotiated transaction) before the CBT is carried out in the same month during the wet season, leaving less demand space, resulting in a small amount of CBT contracts. In addition, from the point of the two-phase contrast of each month, the contract volume in the continuous bidding stage is more than that in the call auction stage. The longer trading period in the continuous bidding stage has a positive impact on the contract volume, but the more important reason is that the real-time disclosure of market information is conducive to market players tracking market trends, capturing price signals, and forming more contracts through more su fficient competition. Although the transaction amount in the call auction stage is small, it still has many functions: First, in bilateral transactions, many market players take the average transaction price of the first stage as the "benchmark" and determine the final transaction price of bilateral transaction electricity [9]; second, in the continuous bidding stag, market players usually bid or o ffer by reference to the average transaction price of the call auction stage [7]. More public information about the CBT can be found at www.kmpex.com.


**Table 1.** The market-clearing overview of the CBT in 2018 (data source: CETP).

Through the analysis of the application in the whole year, it shows that any mechanism can give full play to its advantages only under suitable conditions. The CBT needs to be further improved and adjusted based on the above practical operation analysis. Examples of improvements include the following:

