Gunbot Shop › Gunbot Forum › Gunbot Strategy Lab › GH – SG Mutation C1 Results
Hi everyone, I got GunBot last week (although I have about a year’s worth of bot-trading experience with other platforms) and Aitor and Javier have kindly provided me with a strategy to test. I will be testing this strategy and posting daily updates, regardless of whether the day’s results have been significant or not. Please let me know if you have any feedback or suggestions on my posting style and I can address it for future posts and strategy tests.
Strategy: SG Mutation C1
Running on: Windows VPS (4CPU, 8GB Ram)
Connection speed: 200mb/s
Number of pairs: 6
Pairs: USDT-BNB, USDT-THETA, USDT-TFUEL, USDT-MATIC, USDT-SOL, USDT-INJ
Trading Limit: 12
Start date: 28/06/21
Total profit: $4.37
Total running time: 24 hours (1 day)
Average profit per day: $4.37
Average profit per hour: $0.18
Average DU across all current open orders: 1.2
Max DU across all current open orders: 2
Capital currently utilised (Bags in Base) as a percentage of total funds assigned to the bot: 4.7%
For some reason, my GunBot PNL isn’t showing any profits even though the bot has made some sales. Therefore, I’ve included some screenshots from my Binance PNL account instead. I know it’s not as accurate as GunBots’ PNL as it also includes the asset’s value but it’s better than nothing.
1. The strategy has been great at capturing the bottom of some small moves. Previous platforms I used bought at fixed percentages (i.e. every 2% drop) and so weren’t efficient during downtrends. This strategy is more efficient so far as it waits until near the bottom of the downtrend before placing a buy order further decreasing the average entry price. Examples below:
2. On two occasions the bot didn’t sell even though the price went above the ‘SELL’ line. My bags were large enough to execute the sale and I couldn’t see any errors in the console, so I’m not sure why they didn’t sell. I’m not too concerned about this at the moment as the bot should continue to work the average entry prices down so I’m sure it will have another opportunity to sell them soon. Screenshot below:
Thanks for sharing your results GH. Regarding Binance PNL, I recommend you to take a look to this post:
Day 2 Update
Total profit: Unable to determine – GunBot PnL doesn’t appear to be working and Binance is showing a loss (-$33), however, this will be because of a drop in the asset valuation and not because the bot is selling at a loss.
Total running time: 47 hours (2 days)
Average profit per day: $4.37 (unchanged since yesterday as unable to determine)
Average profit per hour: $0.18 (unchanged since yesterday as unable to determine)
Average DU across all current open orders: 4.5
Max DU across all current open orders: 5
Capital currently utilised (Bags in Base) as a percentage of total funds assigned to the bot: 17.4%
Again, my GunBot PnL isn’t working (see screenshot below), so I currently have to use Binance’s PnL even though it’s not as accurate. Although Binance is showing a negative PnL balance, this is because of a drop in asset valuation and not because the bot is selling at a loss.
The bot has been great at timing the bottoms. Sometimes it buys midway through a downtrend rather than at the bottom of it, but this is probably because the price action increases above the trailing and triggers a buy before it continues back down. There’s not much we can do about this as the bot is performing exactly as it should.
Example of great buys and CT sells:
Examples of good buys and CT sells:
Hi, all the variants have:
1- Higher trailing than original version.
2- Increasing trailing when reaching to deeper DCAs.
In this case you are in the first DCAs, that is, the lower Trailme, but still higher than original version. Anyway, I can modify your version to increase trailme if you want and we can see how it performs. The variants with higher Trailme are running good, maybe it’s still possible to increase it to a higher number and still getting better profit. We are testing and this could be a good “experiment”.
Regarding Gunbot PNL: if you are running v21.6.2 get the upgrade to v21.8.9 or to v22.0.1 from SG Mutation group and you will get updated PNL results (if you stop seeing it updated, just stop the bot and start it again and you will see all the updated info).
Day 3 update
Total profit: $10.08
Total running time: 72 hours (3 days)
Average profit per day: $3.36 (-$1.01 compared to yesterday)
Average profit per hour: $0.14 (-$0.04 compared to yesterday)
Average DU across all current open orders: 4.6
Max DU across all current open orders: 6 (+1 compared to yesterday)
Capital currently utilised (Bags in Base) as a percentage of total funds assigned to the bot: 15.6% (-1.8% compared to yesterday)
Thanks to Aitor’s suggestion, I managed to update my GunBot version yesterday which fixed my issue with the PnL not working correctly.
@Aitor, thanks for the suggestion about increasing TrailMe. I think it would be a great experiment, but I’m using V2 of this mutation so it’s probably not worth it as it might have different results on the newer versions. I do think it would be interesting to try though because looking at the buys circled in green below, I think it would lead to more lower buys than higher buys, meaning that overall it would help lower the break-even price even faster. I did a quick manual experiment on TradingView with the percentage tool, and it looks like increasing the TrailMe by 25% of its current limit could be beneficial, but it would need testing of course.
On another note, in the past when I used a different bot I found that starting my DU as 2 times my TL helped in closing deals much faster (i.e. 2-3 hours instead of 1 day). For example, assuming a 1.05% TLM and a TL of $10, it would BUY at $10, then DU at $20.00, $21.00, $22.05, $23.15, $24.30 etc. Because the first DU is double the BUY, it would bring the break-even down by 2/3 instead of only 1/2. It does increase the funds needed per each pair, but overall it yielded higher profits compared to running more pairs with less funds each. I just thought I would share this as it could be a useful experiment to run.
Increasing trailme could lead to lower buy points… Also note that the easiest and most direct way to get lower buy points is to increase Grid Multiplier parameter.