Support Board
Date/Time: Wed, 27 Nov 2024 05:53:22 +0000
Post From: Sierra Chart sim trading exiting positions at impossible prices
| [2023-09-24 13:02:51] |
| sgne - Posts: 105 |
|
sc.Engr, I will show you what I mean. NQZ23, Wednesday, 20-09-2023, 1p CT, 2p ET, Fed announcement. Short position. Market moves down at 14:00:01.002, when there's a large sweeping sell order. Here's an image showing a position of -1, and then a flatten and cancel order once the new range bar begins, at much higher than the high of the bar. The spreadsheet issued the market F&C order at the beginning of the new bar, which opened at 15336.50, not during the previous bar. And prices continued to flash down to 15319.50, 4 milliseconds later, before rebounding, see below. https://www.sierrachart.com/image.php?Image=1695503744422.png To reflect reality, the Sim1 market buy fill should be no higher than 15339, which is the high of the range bar when the F&C order was issued. So if SCID Bid/Ask values come from the exchange and can't be changed, then SC's fill methodology could be adjusted by having different Sim Bid/Ask values. Here are 3 lines of the Sim1 trade activity log: Buy 1 Sim1 Market Order Sent 2:00:01 PM [Sim]NQZ23_FUT_CME Auto-trade: Replay 1.00X: NQZ23_FUT_CME [CB][M] 49.75 Range #1 | nqSpread | Source: Flatten & Cancel cell | Bar start date-time: 2023-09-20 14:00:01.002. Flatten&Cancel | Last: 15336.5. Current Position quantity: -1 | AO=0 Buy 1 Sim1 Market Open 2:00:01 PM [Sim]NQZ23_FUT_CME Simulated order accepted Buy 1 15371.25 1 Sim1 Market Filled 2:00:01 PM [Sim]NQZ23_FUT_CME Trade simulation fill. Bid: 15368.25 Ask: 15371.25 Last: 15336.50 The last line specifies the problem, which is the Bid and Ask values in the SCID file. Bid and Ask prices stayed high at Bid: 15368.25 Ask: 15371.25, even though the close of the bar was 32 points below the bid, and trade prices went lower by 17 more points in the next 4 milliseconds. Only when the milliseconds changed from 002 to 006 did the Bid and Ask change in the SCID file. Here's a snapshot of trade prices, from the SCID file, with many intermediate trade prices deleted: NQZ23 ===== 20230920 140000.005 15368.25 20230920 140000.152 15368.25 20230920 140000.371 15368.25 20230920 140000.469 15371.00 20230920 140000.843 15371.25 * prices drop continuously at this millisecond * 20230920 140001.002 15368.25 ———— 20230920 140001.002 15337.00 * new range bar begins * 20230920 140001.002 15336.50 ———— 20230920 140001.002 15331.25 * new millisecond * 20230920 140001.006 15331.75 ———— 20230920 140001.006 15319.25 * end of fast price drop * Here's some more extracted SCID data with Trade, Ask, Bid prices, in that order: 1 20230920 140000.843 15371.25 15371.25 15368.25 2 20230920 140001.002 15368.25 15371.25 15368.25 3 20230920 140001.002 15368.25 15371.25 15368.25 4 20230920 140001.002 15368.25 15371.25 15368.25 5 20230920 140001.002 15368.25 15371.25 15368.25 6 20230920 140001.002 15368.25 15371.25 15368.25 7 20230920 140001.002 15368.0 15371.25 15368.25 8 20230920 140001.002 15368.0 15371.25 15368.25 9 20230920 140001.002 15368.0 15371.25 15368.25 10 20230920 140001.002 15367.75 15371.25 15368.25 Ask and Bid are fine up to line 6. But for line 7, there really wasn't any bid at 15368.25. Theoretically, the Bid was 15368. And since the line 8 trade was no higher than the line 7 trade, the Ask on line 7 could be taken to be the price of the previous higher trade, on line 6. And so forth. (And I have had real, live fills, but which later in a simulation did not fill, even though trades went off through and beyond limit order prices; I have mentioned this before; this is because of artificial SCID Bid/Ask values.) Date Time Of Last Edit: 2023-09-24 13:04:37
|