Barjon's Money Machine

[Soloquan]

Hey solo

Paying 1 point/tick and shooting for 10 is way too much for me if using a bucketshop. You would have to have a massive edge to overcome that spread. For me I would only use a bucketshop for 30 point target with 1 point spread, that's just me others may disagree. If you are trading the futures you can try and buy at the bid sell ask but the spread may move out of line waiting for a fill, then if you do get a fill you have to manually 'leg' into the other side. You could use a spreader like xtrader/cqg but you are probably looking at £500 to £800 pm software costs.

To illustrate the effect of the cost of execution (spread, commission, exchange fees) I know of a ES trader who does 10,000 lots per day. sounds a lot and it is but he trades with a 50 lot and goes in and out of the market 100 times per day. Mean reverts and averages in once, twice but never thrice. his winners are 2-6 ticks, losers 1 to 5 ticks he makes an average of $1 per contract traded. Tick size is $12.50. So he makes like 8% of one tick per trade! with a retail round turn cost of $2-$4 you can see how he would get killed by execution costs. Due to volume and trading through an exchange clearing prop firm his all in round turn cost is around 40c. I say he makes $1 per contract, he is salaried plus small % of performance.

I know this is an extreme example but it has 2 'lessons'

1. Best not to trade where total execution cost relative to your target is large, it's a personal judgement, don't underestimate the total cost of execution be it bucketshop spread, futures commission, exchange fees etc.

2. More importantly there ARE pro traders out there who are making money from people not paying attention to point 1 as shown by my example above of the guy doing 10,000 per day ES volume. they make money from what they call 'liquidity monkeys' who trade with retail commissions and tight stops.

Sorry for the ramble that's what happens when you get stuck on a train!

Good trading everyone.

Thanks for the reply CD and yes I completely understand where you are coming from.

However to quote you: "Paying 1 point/tick and shooting for 10 is way too much for me if using a bucketshop. You would have to have a massive edge to overcome that spread."

I think far too much emphasis is placed on this word 'edge'. This is a complete construct (i.e subjective and not based on empirical evidence) that has been fabricated by people looking for answers as to why some people can consistently make money and some lose.

In my opinion and I don't use that word lightly I genuinely mean it is just my opinion, in my opinion I haven't got a "massive edge" as you put it. In fact I don't even have an edge. What I do have is a knowledge of the instrument I'm trading simple as that.

If my bucket shop account is consistently growing in the right direction then all is well. Paying a 1 point spread to make a minimum of 9 is okay with me.

 

[scose-no-doubt]

Interesting thread and strategy and something I've looked at myself. The problem I had with it is that I'm still struggling to see what exactly it is you're being compensated for as I'm assuming a fundamental shift factor in the regressed values (i.e. a decoupling in the historic UK/US economic relationship) is not a common enough occurrence to warrant a quasi-continuous reward for acceptance of that particular risk. I'd assume there are some FX considerations too but overall I doubt there is a significant and inherent mis-pricing which would continually provide opportunity. Given what Martinghoul said about edges yesterday, it could just be that offering liquidity outside of normal hours or at times when it's not just about... I'm stumped and I'd never trade something I didn't understand no matter how loud the numbers scream it at me. I'm thinking it;s related to liquidity premiums but I dunno. Do you (or anyone else) have any idea why you're being paid, Jon.

 

[ChocolateDigestive]

I think far too much emphasis is placed on this word 'edge'. This is a complete construct (i.e subjective and not based on empirical evidence) that has been fabricated by people looking for answers as to why some people can consistently make money and some lose.

Hi Solo

That's interesting your thoughts on the word 'edge' my own thinking is the exact opposite in that an edge for me is defined by objective empirical evidence over a large sample size.

What matters at the end of the day is that you are making money, so if you are keep on doing that. If you are targeting 9 pips/points/ticks and paying 1 pip/point/tick spread then imo you must have a big edge however you are defining it. just my 2c.

by the way in your signature what does JSA stand for? I am hoping that it doesn't stand for job seekers allowance, apologies in advance if it isn't, no problem if you are genuinely entitled but if you are milking it for the sake of it that is bad cricket.

 

[ChocolateDigestive]

Interesting thread and strategy and something I've looked at myself. The problem I had with it is that I'm still struggling to see what exactly it is you're being compensated for as I'm assuming a fundamental shift factor in the regressed values (i.e. a decoupling in the historic UK/US economic relationship) is not a common enough occurrence to warrant a quasi-continuous reward for acceptance of that particular risk. I'd assume there are some FX considerations too but overall I doubt there is a significant and inherent mis-pricing which would continually provide opportunity. Given what Martinghoul said about edges yesterday, it could just be that offering liquidity outside of normal hours or at times when it's not just about... I'm stumped and I'd never trade something I didn't understand no matter how loud the numbers scream it at me. I'm thinking it;s related to liquidity premiums but I dunno. Do you (or anyone else) have any idea why you're being paid, Jon.

you are clearly searching for a reason why this method could work. stat arb traders looking at stat relationships so they are less concerned about the why. If you know the x-y spread as crossed the centre line z times (mean reverted) in the period p and you know the standard deviation is sigma then all you are playing it the probability of that event occurring again. It's not what you want to hear I know but sometimes there is not a why, there is no reason.

This is not the case with all spreads, if you look at 'new crop old crop' or calender spreads various price action can be explained / part-explained by various fundamentals i.e. supply imbalances, carry costs etc.

 

[barjon]

Interesting thread and strategy and something I've looked at myself. The problem I had with it is that I'm still struggling to see what exactly it is you're being compensated for as I'm assuming a fundamental shift factor in the regressed values (i.e. a decoupling in the historic UK/US economic relationship) is not a common enough occurrence to warrant a quasi-continuous reward for acceptance of that particular risk. I'd assume there are some FX considerations too but overall I doubt there is a significant and inherent mis-pricing which would continually provide opportunity. Given what Martinghoul said about edges yesterday, it could just be that offering liquidity outside of normal hours or at times when it's not just about... I'm stumped and I'd never trade something I didn't understand no matter how loud the numbers scream it at me. I'm thinking it;s related to liquidity premiums but I dunno. Do you (or anyone else) have any idea why you're being paid, Jon.

scose

I don't think it's a question of mispricing - as is the case in a true arb - but more a case of DOW leading the way as far as FTSE is concerned in terms of general sentiment. Conditions in UK cause the FTSE to deviate from that relationship for periods but the pull of the DOW brings it back towards the mean even if it doesn't make it the whole way.

 

[counter_violent]

scose

I don't think it's a question of mispricing - as is the case in a true arb - but more a case of DOW leading the way as far as FTSE is concerned in terms of general sentiment. Conditions in UK cause the FTSE to deviate from that relationship for periods but the pull of the DOW brings it back towards the mean even if it doesn't make it the whole way.

Yeah, it's just normal intra-day /intra-week (however long it takes) variance, and a damm site safer strategy than most employ.

 

[DionysusToast]

No it doesn't matter if you are holding for hours/days. The spread at a bucketshop will be tiny compared to the trade result. For the intraday index arb guys they are mainly automated on the lower timeframes the fills are important in that space. the majority of HFT arb index guys are flat end of day for reason I won't go into.

Well if it's an index arb that you can hold for days...

Then he's found the pot of Gold @ the end of the rainbow....

As for why index arb guys are flat @ the end of the day, it's because the opportunities are fleeting in markets like S&P Futures vs Cash....

 

[ChocolateDigestive]

Yeah, it's just normal intra-day /intra-week (however long it takes) variance, and a damm site safer strategy than most employ.

aye aye. nothing wrong with 'however long it takes' as long as the trader is aware of the psychological trait of some traders which is 'inability to take controlled losses' something I have suffered from in the past. There is a danger with spreads that a mindset of 'it has to come back' can set in and the high hit rate% of these methods causes this trait to be tested. Strict rules on averaging down (amounts, how many times (if any)) need to be employed as well as the ability to take controlled losers, plenty of pro traders have average loser greater than average winner contrary to popular belief. otherwise things could go pop.

 

[Soloquan]

scose

I don't think it's a question of mispricing - as is the case in a true arb - but more a case of DOW leading the way as far as FTSE is concerned in terms of general sentiment. Conditions in UK cause the FTSE to deviate from that relationship for periods but the pull of the DOW brings it back towards the mean even if it doesn't make it the whole way.

For instance... ftse just rose from 5626 to 5638 (a gain of 11 points in a matter of about a minute) however DOW only rose from 12754 to 12768 (a gain of ONLY 14 in the same space of time) As Barjon stated DOW = 2 to ftse so in reality ftse should have seen a bigger gain yet didn't as the DOW pulled ftse back to it's average.

I think on a larger scale it is exploiting this 'anomaly'... well that's the wrong word but I can't think of the word I'm looking for right now, but you know what I mean.

 

[scose-no-doubt]

you are clearly searching for a reason why this method could work. stat arb traders looking at stat relationships so they are less concerned about the why. If you know the x-y spread as crossed the centre line z times (mean reverted) in the period p and you know the standard deviation is sigma then all you are playing it the probability of that event occurring again. It's not what you want to hear I know but sometimes there is not a why, there is no reason.

No! I refuse to believe it

I think I know what it is now anyway...

Dunno if anyone wants this spreadsheet I made whilst bored. Hope BJ doesn't mind

 

[DionysusToast]

Yeah, it's just normal intra-day /intra-week (however long it takes) variance, and a damm site safer strategy than most employ.

Why is it safer, exactly? Why is it not just another Howard Cohodas/Spanish89 waiting to happen?

The reason I ask is that arb opportunities tend to exist for very short periods.

Look @ the title of the thread "Barjons Money Machine". Whilst tongue-in-cheek, that is exactly what this looks like.

As did Howards option strategy to many. I got shouted at for being one of the first to point that out....

Jon - do you have any numbers on this in terms of what you are potentially risking on a trade, what your hit rate is, when you dump a trade and what a winner would yield?

 

[scose-no-doubt]

It would take a massive event that caused a decoupling of the US/UK economies to make this kind of strategy approach HoCo risk profile, wouldn't it. Ain't as many historical black swans there afaik cos they've normally got the same vested interests. Plus we have this issue of of asset beta, ADR pricing and FX arb and EMH that I've become obsessed with lately

As long as you're safely leveraged of course. I'd think diminished relative returns would be a foregone conclusion.

 

[ChocolateDigestive]

As for why index arb guys are flat @ the end of the day, it's because the opportunities are fleeting in markets like S&P Futures vs Cash....

yes that is one reason. In the automated HFT algo space you either need to be first or the fastest (or both). another reason is they focus on high value sharpe ratio returns and the preference is for a high number of intraday trades. another reason is the risk is much higher holding overnight and there clearing partner will place many more constraints on them for overnights.

there is some really good stuff in this up to date book on HFT - see below link. I dont trade automated HFT but it is def important to stay abreast of HFT so you know what is going on in your chosen market.

Book « The Speed Traders, The Newest High-Frequency Trading Book by Edgar Perez

 

[DionysusToast]

you are clearly searching for a reason why this method could work. stat arb traders looking at stat relationships so they are less concerned about the why.

So you are saying that institutional stat arb traders pay no consideration to WHEN they should trade & when they shouldn't?

Interesting.

 

[DionysusToast]

scose

I don't think it's a question of mispricing - as is the case in a true arb - but more a case of DOW leading the way as far as FTSE is concerned in terms of general sentiment. Conditions in UK cause the FTSE to deviate from that relationship for periods but the pull of the DOW brings it back towards the mean even if it doesn't make it the whole way.

So the DOW is the dog & the FTSE is the tail?

How long have you been running this Jon?

Have you considered the possibility that what you have actually manufactured here is a synthetic USD/GBP directional trade?

 

[DionysusToast]

yes that is one reason. In the automated HFT algo space you either need to be first or the fastest (or both).

And cheapest of course.

I think 'first' encapsulates this but it's worth pointing out that being willing to accept a smaller spread (which really requires lower costs) is what gets you in first.

Hence, large obvious arb opportunities being as rare as a Howard Cohodas losing trade!

 

[ChocolateDigestive]

Why is it safer, exactly? Why is it not just another Howard Cohodas/Spanish89 waiting to happen?

The reason I ask is that arb opportunities tend to exist for very short periods.

Look @ the title of the thread "Barjons Money Machine". Whilst tongue-in-cheek, that is exactly what this looks like.

As did Howards option strategy to many. I got shouted at for being one of the first to point that out....

Jon - do you have any numbers on this in terms of what you are potentially risking on a trade, what your hit rate is, when you dump a trade and what a winner would yield?

well I would hardly call what Barjon is doing an arb trade, certainly not a true arb. more accurately it is a spread trade.

why is it safer? because you are somewhat insulated from extreme events. lets say someone launches a nuclear missile at new york later today then both components are going to tank however the spread is likely to tank less relatively speaking. CV is not the only person to think they are safer, if you take a look at the trading margins required for spreads versus single instuments you will see they are way lower.

check it Performance Bonds | Margins Requirements for CME Group Products

 

[DionysusToast]

It would take a massive event that caused a decoupling of the US/UK economies to make this kind of strategy approach HoCo risk profile, wouldn't it.

Well yes - if the indices were the same currency & in 2 stable untroubled economies....

You have both currency risk & risk that the US & UK economies are not hard linked.

A devaluation of the FTSE is hardly going to bring down the US for instance.

Your original question of 'why' is an important one.

 

[scose-no-doubt]

Have you considered the possibility that what you have actually manufactured here is a synthetic USD/GBP directional trade?

US GDP is comprised mainly of domestic demand whereas Uk is highly reliant on EU exports so if anything I think the first line would be USD/EUR rather than USD/GBP. Oviously there's be more too it than that but given the majority of currency transactions are USD JPY EUR GBP I think that cross would be more relevant.
Personally, as we're talking about globally diversified companies, I'd guess we're looking at an the aggregate beneficial/adverse exposure of the index constituents to global growth trends rather than being in an eggs in a single basket scenario.

A devaluation of the FTSE is hardly going to bring down the US for instance.

Your original question of 'why' is an important one.

But I think that a devaluation of FTSE would (could) not occur independently to the Dow due to globalisation and the general nature of modern corporate structure and interests.

 

[ChocolateDigestive]

So you are saying that institutional stat arb traders pay no consideration to WHEN they should trade & when they shouldn't?

Interesting.

not sure what you been smoking toasty. I said 'stat arb traders looking at stat relationships so they are less concerned about the why.

this means they are less concerned with the fundamentals as other team members will look at that. often maths and physics phd's are not the best to consider fundamentals anyway they do the maths.

Institutional HFT algo traders work in teams, programers, traders, stat arb guys.

not sure where you got the WHEN they should trade, that is the exact opposite of what they do.

vendors should be banned from posting lololol