About the option: ES FOP 20220729 2800 P GLOBEX 50 E-mini S&P 500 [EWN2 P2800, 542609194, mult: 50] which we observed to have a spread worth $2.5K, you can see now that approaching RTH, the spread becomes more acceptable ($625) and even the layer PNL turns positive: Clearly, from the point of view of automation, unpredictable spread and wild quotes pose a significant challenge and a good number of numeric checks are needed before taking any action and avoiding being filled at some wild unfavourable price. We always only use LMT orders.
Finally ending the week with ES rising and a decent PNL: 92K.3K, which is another high-water mark. We have been trading for almost 117 solar days. Total commissions: 10.3K. Currently, we have 240 positions open with a fund usage of 66.72%. As said, we keep around 50%, to be able to ride and make $$$ on the corrections. ES FUT 20220617 GLOBEX 50 E-mini S&P 500 [ESM2, 477836957, mult: 50] Two options are expiring in 4 days (both negative PNL), and these are the last ones we bring to expiration. From now on, I will take minimum 50-60 gg or more (>120) and, before $4 or higher, I am going to "transfer" their game to another layer (rollover). I have been wasting enough time and resources on these "penny options". Global situation of the folio: Even though we are trading a "unique" type of instrument (ES options), which "on the field" proved to be the most "suitable" for this algorithmic game, we can still achieve some "diversification" through varied timing and a variety of choices within the options matrix. One clearly needs to let the market move and not jam all layers at the same time/situation I know you are curious about the luck of the option we have seen in the previous post. Here it is: ES FOP 20220729 2800 P GLOBEX 50 E-mini S&P 500 [EWN2 P2800, 542609194, mult: 50] Transfer: ES FOP 20220617 2400 P GLOBEX 50 → ES FOP 20220729 2800 P GLOBEX 50 You see here another example where we turned a "loss" into a profit through "information transfer" (or rollover, if you prefer.) Recall that this one was almost -2K in the red at expiration. So this is another example of a "stop-loss recovery" mechanism (we have seen several others previously), which is obviously essential to obtaining long term profit in an algorithmic context (otherwise, in general, the stop-loss orders will practically kill most of the gains).
Not much happening after the mkt holiday. ES is still around 4130, and currently seems to be heading down, but not with much conviction: PNL situation at the moment around 96K with a maximum of 98.8K reached yesterday. 121 days elapsed sine start and 10.5K gone in commissions. Two layers just expired. One with negative PNL was taken care of by "information transfer" (rollover) to recover stops. Less than 50% of funds are used in margins.
Sure dorietrading, about that. More or less in line with what was expected. Everyone else doing this approach with my algo is doing almost exactly the same, even with completely different and personal choices of specific instruments. After all, it's pretty "mechanical", as it's more or less the decay plus the result of the scalping/hedging action and stop-loss recovery. No "prediction" or "signals" whatsoever are involved here. We use most of the time around 50% of funds, in order to be able to trade the dips and corrections and make more $$$ on those. (The difficult part is not to yield to the greed and temptation to load more than it is safe.) Keeping a good reserve of margins takes care of riding possible "storms". Starting layers at different times and price situations, like the "stages" of a missile, helps "diversify" and "distribute" risk. Pretty viable, imho.
To reinforce the concept, the "mechanical" nature of this approach makes it quite robust. This is the opposite of prediction or "signal"/"pattern" or "learning" based strategies, which instead live under the delusional assumption that the market is somehow putting out somewhere information useful to obtain locally a statistical advantage in terms of take-profit / stop-loss avg balance. Which of course is an underlying assumption that nobody has ever been able to prove. Because in fact, it is not true, and it is an "invention" of people who are statistically illiterate, often fallaciously "supported" by conceptually incorrect application of statistical validation methodologies, such as so-called "backtesting", that is, in practice, computer-assisted generation of useless and misleading "45-degree lines" : (https://jpm.pm-research.com/content/early/2019/12/24/jpm.2019.1.123) instead of proper and thorough "simulation studies" with meaningful simulation methods carried out on the entire sample space, as it happens for any legit statistical methodology). The fact that the "hoped" advantage does not really exist and the fact that they are not able to recover the stop loss, it's often the ultimate cause of long-term unprofitability. Also, note that it is not even sufficient that an approach is conceptually valid. It is also fundamental how you "deliver" it. That is aspects like implementation, execution, and risk management are crucial. And, even the best approach in the world, can and will fail if also all these other aspects are not aligned. Daily practice, with the consequent experience, in this regard is certainly crucial and one never really stops improving. Enough now with "philosophy"! Let's go back to the results. Just after a made the celebration post, the mkt thought well I needed a beat-up, and actually, after a few minutes, I was under about 15K But that is the way it is, and that is why one cannot rely on guessing but have a strategy in place plan to deal with anything, especially hedging measures to fight dd, obviously. ES FUT 20220617 GLOBEX 50 E-mini S&P 500 [ESM2, 477836957, mult: 50] Anyway, we are currently around 104K with a maximum reached PNL seen of 109.9K. in 123 solar days trading. With a "precautional" average use of less than 50% of the funds (2M). Currently 34%. Let me show again the point about the "stop-loss recovery mechanism", which is obviously a crucial point, and which actually makes this approach work better than others which don't implement that. The recovery is based on the capability to move all the trading information across layers, and it is also the reason why this cannot be done manually, or "copied" (trade copying). I am contacted on a daily basis by all kinds of "postulants", and many also ask about "copy trading". This approach is purely algorithmic and cannot really be used for "copy trading" because it requires additional information which is stored in the layers and moved across. Copying trades would let you with "orphan" orders on single layers, with possibly catastrophic results. I hope this would help avoid more people coming to me and asking about this in the future. Further, the bot works at tick level (just to attempt a single entry it makes more than 15 different checks on quotes and spread) and is able to pick quotes that a human and not even another machine (if not adequately programmed) can pick. And in any case, the delay would not really allow performing the same order. That means a single copied execution may give such unfavourable entries that would make you cry (like -1000$ or more), especially with options of course and their crazy quotes or you may not be able to fill at all. Not to mention the margins aspect so that what can execute on one account can either cause liquidations on another one or be an unbalanced position. Further, there is a possible problem with loss of sync with the underlying account (skipped event, connection lost, etc.), that the bot has the proper built-in means to solve, while a copy would be lost in a nightmare of confusion. Here are 2 examples of layers where the stops did not allow us to make a profit, but that -as always - we are able to make profits through information transfer and stop loss recovery: ES FOP 20220729 2800 P GLOBEX 50 E-mini S&P 500 [EWN2 P2800, 542609194, mult: 50] ES FOP 20220729 2930 P GLOBEX 50 E-mini S&P 500 [EWN2 P2930, 563754452, mult: 50] to make more evident where the "information transfer" happened I just coded that "green arrow" you can see now on the screen, with the name of the origin and target layer. And this can go on forever, thus riding a continuous curve. There are no "gambles" here. "Win or lose". It's "mechanical" and we stop the layer only in profit. That is why ALL my layers will always - sooner or later - turn "green":
Yesterday I did not make it, as I was too busy. Let's see what's happening today. We left the S&P futures at level 4161, and right now we are just a bit below (4150): ES FUT 20220617 GLOBEX 50 E-mini S&P 500 [ESM2, 477836957, mult: 50] I have started another layer (pretty safe as you can see on our "riskiness" indicator): ES FOP 20221031 2850 P GLOBEX 50 E-mini S&P 500 [EWV2 P2850, 559670304, mult: 50] because funds usage was too low. Currently, we are using about 46% of our funds (see margins gauge): The current profit is about 116 K, with a maximum of 118 K (127 days trading) and almost 11 K gone in commissions. The layer which is closer to being "transferred" (or just closed) is the following, as the price is getting too low: ES FOP 20220729 2850 P GLOBEX 50 E-mini S&P 500 [EWN2 P2850, 542609085, mult: 50] Let's wait for a smaller spread though (this spread is worth $ 525 to which we need to add $ 1.42 x 2 commission).
Touching today 120K, which is good news. 11K in commissions, even better news for the broker. 129 days so far. Current position 180. Margin usage is 55.8% (1.2M). We are slowly turning all layers positive.
Here we have another practical example of how we turn a negative layer into a positive one by "information transfer" and "stop-loss recovery" mechanisms: Transfer: ES FOP 20220624 2800 P GLOBEX 50 → ES FOP 20220819 2750 P GLOBEX 50 (This was closing negative and now is 7.3K.) The green arrow in the screenshot, where there is the price jump indicates the instant of "transfer". The possibility of moving freely across different levels of sensitivity and volatility (Delta, Gamma, Theta and other greeks) in the options matrix, allows us to modulate risk, which would not be possible working directly with the underlying futures (or it would be possible only manipulating sizes and taking "losses" due to hedging: buy high, sell low).