Robot versus retail on the ASX

[Gringotts Bank]

So you are saying you can see how 'someone' is manipulating a/few/many stocks and making money at will and you are not going to copy it?

Not 'someone'... me. I put a buy order in, the price moves up by 2-3 ticks and ahead of me in the queue is a bunch of orders totaling less than 1/5th of the volume I want to offload.

You obviously don't trade the ASX. Today that same stock has a tighter spread with reasonable volume. Why? Because we're not at a point where MR traders want to get in on the action.

 

[Trembling Hand]

GB what is the stock. If you can do something and know what will be the reaction it's a guaranteed road to riches.

 

[skc]

Someone kindly lent me the trades that went through around that time, with the broker codes.

Of note:

1- Many different instos buying/selling. Competition is good, since it reduces the chance of manipulation significantly.

2- All insto broker codes appear as both buyers and sellers. I don't like that one bit, but it could be legit. We'll never know.

3- My little retail trade was bigger in size than 95% of insto trades that went through. Hard to believe, but this accounts for the fact that the bots see it and always move the price away. It probably registers as 'big buying interest', which is hilarious considering the amount involved.

4- The number of retail trades is absolutely miniscule compared to insto. It's basically all insto.

Having access to the same data... here are my findings. Specifically, I am looking at the 26 minutes and 6 seconds between when you placed your order in the bid and your order's completion (after you were manipulated/coerced/enticed to lift your bid price).

The following matrix shows buying broker in rows and selling broker in columns, and how much volume each broker transacted with one another.



1. 16 different brokers were involved, in which 7 of those appeared as both buyers and sellers (light green shades).

2. Of the 7 brokers who appeared as both buyers and sellers, only 3 of them appeared as both buyers and sellers on the same transaction. These transactions (of which there were 7 instances), totaled only 699 shares (dark green shades). This represent on 2.3% of the total volume traded through the period. All 7 transactions were clearly marked XT which is information available to all market participants in real time (i.e. not delayed or insto only).

3. The total volume traded thru the period was 30,174 shares. You know which one is your order so you can conclude what % of total volume that represents. Remember that the actual volume of flow was 30,174 x 2, as there are equal number of buyers and sellers. So consider if it is reasonable for 16 different brokers to collate and manipulate your order by moving that amount of share, relative to the size you tried to trade.

4. The total volume of the 4 retail brokers (yellow highlight) during this period was 11,692 (6,256 bought and 5,436 sold). So retail transactions represented 20.7% of buying volume and 18% of selling volume, which is not insignificant - and certainly not minuscule. I think you may have mis-interpreted the course-of-sale data that small trade volume = insto activity. That is true only for one half of the transaction... the other half could be insto or retail.

5. Look at the course of sale and see who you actually traded against... another retail broker was the seller to >60% of your size. So it's not like an insto front ran you and sold it back to you a few cents higher. You actually bid into a seller who's order had been there for over 3 hours.

6. The difference in cost between your actual traded price and your intended trade price (which you queued in the bid) was $162. But there was never any guarantee that a bid would be filled, regardless of how large/small/chopped up your order may be. At the time you placed your order, there was sufficient liquidity to fill your volume at the ask, 1c above where your order was. Looking at the course of sale, 90% of this ask volume was indeed filled at that price over the next ~6 minutes - meaning they were not fake phantom orders that disappears in micro seconds (i.e. not high frequency trading). There was a further 1.5x your volume at the next ask level, which again were mostly filled at that price over the next 4 minutes or so. So on hindsight, a straight buy the ask would get you mostly filled and a limit order 2c above would most likely see you fully filled.

7. Looking closely at the matrix again... the biggest seller was Citi with just under 1/3 (9,353 shares out of 30,174) of the selling - could it be the manipulator?. Yet in the same period, they only bought 205 shares (or 0.68% of the buying). So it seems unlikely that they would have been able to manipulate the market given the small amount of buying. Citi's 205 bought shares were transacted at close to the end of the 26minute period, so it'd have zero impact on you.

8. Citi's selling coincided with what you considered a price rise after your order was placed. So if indeed some party other than Citi tried to manipulate the price to get you to lift your order, they were met with another insto who was a keen seller. As long as there is more than one insto in a stock, manipulation would simply be met with a "Thank you for a better price" from another insto who wanted to trade in the opposite direction as the manipulator.

In summary, there is no evidence to suggest that one party was specifically targeting your order in the queue in order to benefit. The probability of all brokers colluding to manipulate the price against a total prize of ~$162 (the amount you think you got fleeced) seems rather low. So I am of the opinion that it just happened that way, and you'd be better off either crossing the spread the moment you wish to trade, or not chase the price up unnecessarily.

 

[Gringotts Bank]

skc, that's impressive. A bot can still see what appears to be buying pressure in the depth (a larger than usual single order in the top few lines), and move both the bid and offer upwards just one tick with small volumes. Like moving your beach towel up a bit in anticipation of the tide coming in. And as soon as one person moves his towel, others follow... before you know it you're paying $160 more to enter the trade. That's not manipulation as such, but it does reinforce the importance of not sitting a bid in the queue.

I maintain that larger than usual buy/sell volumes in the top line will move the price up/down away from the order. I've seen it too many times.

 

[DeepState]

A bot can recognise what you can...that the instos are trading in small volumes that change all the time in terms of levels and queue priority. Retail looks just like what you say it does. A blob on the screen. If you are trading in size, you don't sit like a blob on the screen. It demonstrates that you have no capability to trade for low impact or are asking to be hit. It's not a clear indicator of major size coming. If trading in size, you don't do what you are doing.

If the bots jumped ahead of you on the queue, it's because someone/group wants the volume more than you do at that time. How is that weird? If the stock was good enough for you at that price, why not someone else at a fraction higher? If you think that they are head-fakes that are designed to draw you to a higher price, then being a blob is its own protection.

Brokers trade for both buyers and sellers in the same lines all the time. It's not like they say "Sorry Mr State, we are acting for Mr Banks today and it would be a conflict of interest for us to accept your business, good day" when I call in. The algos applicable to the different participants and the different sides can vary with the clients' discretion...and they do not know who is trading with or against them with the same broker.

Lengthening your timeframes is a good place to start....if your insight is flexible enough to do it. Or just cross the spread when you find the level you want.

 

[Gringotts Bank]

If the bots jumped ahead of you on the queue, it's because someone/group wants the volume more than you do at that time. How is that weird?

It's only annoying when the price is sitting doing nothing for ages then decides to move as soon as your bid hits the screens. It's an obvious reaction to my bid appearing. Breaking a trade up into small bits is not an option for me - I'd rather just pay a fraction more and lift the offer. But these are not big trades, as I've said. They are retail size. They are just big compared to bot trades which tend to 'crowd the plate'.

 

[DeepState]

It's only annoying when the price is sitting doing nothing for ages then decides to move as soon as your bid hits the screens. It's an obvious reaction to my bid appearing. Breaking a trade up into small bits is not an option for me - I'd rather just pay a fraction more and lift the offer. But these are not big trades, as I've said. They are retail size. They are just big compared to bot trades which tend to 'crowd the plate'.

I am wondering what you expect. When you appear, it becomes apparent to all that more demand appeared. Somewhat like the sack from Schrodinger's cat, instead of guessing, it's opened, and we can now see there is demand which has materialised. If the bots are doing nothing all day, and this occurred, the demand curve moved and some trickling of the order book will occur to allow for that. Your volume is small, the volumes involved in the trickling are even smaller. So who is manipulating who if you sit on the blob/bid and just wait. The change in marginal value of trade is negligible. There was effectively no volume surrounding the prevailing price before you arrived...and it didn't really change afterwards.

Hit the offer and/or be patient. It's the same as it was before the bots. Be the blob or hit the offer.

It's interesting that you might choose to complain about this. The general finding in the markets about the impact of bot trading has been that the top of book liquidity has become much better but the depth of book has now suffered.

Translation: Bots are good for GB. Smile and dial.

 

[captain black]

Captain black knows about this stuff.... Cap been awol for a while.

Thanks for the mention GB. I don't want to derail your thread or make a song and dance about leaving so really just want to say that I agree 100% with what McLovin has said at the link below:

The Five Commandments of Aussie Stock Forums

Thanks for all your interesting posts over the years GB. All the best with your trading.

 

[Gringotts Bank]

Thanks for the mention GB. I don't want to derail your thread or make a song and dance about leaving so really just want to say that I agree 100% with what McLovin has said at the link below:

The Five Commandments of Aussie Stock Forums

Thanks for all your interesting posts over the years GB. All the best with your trading.

Thanks to you too Cap. Let us know if you decide to post elsewhere online. This was one of your best, imo: https://www.aussiestockforums.com/threads/captain-blacks-day-trading-futures-thread.32793/ I never quite cracked the code, and was hoping for more tidbits.

 

[skc]

A bot can still see what appears to be buying pressure in the depth (a larger than usual single order in the top few lines), and move both the bid and offer upwards just one tick with small volumes.

Yes a bot can see it and so can everyone. So it doesn't really give bots a massive advantage over retail. Again, in your example, the prices were essentially unchanged for 6 minutes after you entered your order. That is a long enough time for bot or human to react. You can't expect prices to stay the same forever until you get your fill.

I maintain that larger than usual buy/sell volumes in the top line will move the price up/down away from the order.

So here you complained about your order in the queue not getting hit.

If you're in the queue and your order goes through, you've usually made a mistake, because the price is continuing up (if you've been bought) or down (if you've been filled).

And here you complained about your order in the queue getting hit.

Sure in an ideal world we all like to get a fill and have the price move in our favour straight away and never experience a single cent of drawdown... but short term price actions are random enough that any of the above situation can happen. Every trade that occurs on the market involves one side getting filled (who was sitting in the queue) and another side crossing the spread. So there's no inherent right or wrong in joining the queue. Again, using your trade as an example... your selling counterparty managed to sit in the queue with a reasonable sized retail order AND get a fill AND not have the price immediately go against him. So every problem you perceived did not happen to your counterparty... and worked in his favour instead.

If I were you, my takeaway from everything here would be:
- Am I trading too impatiently when I moved my order to chase the price?
- Am I being a dick for a tick?
- Are my stop too close / position size too large so I overreact to noise that is perfectly normal intraday?
- Am I reading the price momentum correctly?
- Am I blaming my lack of trading success on others?
- I need to learn not to jump to conclusion when I haven't done the proper fact check and analysis.

 

[traderxxx]

Hi GB
if you want instant fills with lots of liquidity
and rarley any slippage have a look at trading
the spi futures.