Trading On Auto

[PeterLe]

Ive been busy working today, so i left BA running on auto. Ive just checked the P&L and logs etc and just surprised at how little they have produced.
I only have my own records to compare against, but despite creating more sophisticated strategies, it just seems to be getting harder and harder. (note: I know its only Feb, but my best month ever was in a February)
(incidentally I have also tested old strategies that have made BIG profits over the years only to find they no longer work, or trigger so infrequently that they are ineffective).

Please dont get me wrong; Im very grateful for what I have taken over the years and still doing OK, but Im wondering if now I should be spending more time on the manual trading. (Ive been tending to do this more and more over the last twelve months anyway).
I would be interested (and grateful) for feedback from those who trade on auto.
I dont mean this the wrong way, but Im interested in those achieving greater than say £1k/Week as it can distort the findings if someone say "My profits are going up and up" Yet they are only doing £10/Day
I know there are some good Auto traders on here; Liam; Gazuty, IPC, amongst others (inc Peter of course!)

I would never ask nor expect any details on strategies (thats a given), but just interested in knowing - is anyone doing well on auto! Looking for inspiration please!

Thanks in advance
regards
Peter

[Scalping2go]

Great question peterle. It appears to me – no proof – that when compared to a few years ago uncertainty has entered the market with the result being more choppy market prices. Therefore less defined trends. Lack of trends may be more prevalent in some markets rather than others – for example the Australian horse racing has better trends BUT less liquidity. Maybe a factor when determining a market to scalp is the total liquidity. Looking at this aspect

[PeterLe]

Thanks Scalping2go. I didn't mention it in my post, but I was talking primarily about In running (horses).
I always hoped that the Aus markets would develop better in play than they have. The sheer number of them would make it worthwhile. For me personally, Ive never been able to make them pay.

[gazuty]

Fortunately everything still going great for me on auto. I must load the markets each day and apply the relevant automation.

Returns are stable, if not slightly improving.

My only down day this month involved a manual trade on a group 1 race I forgot to close.

A few tweaks and I'm sure you'll be back in town.

---

Just checked and today thus far, 32 traded, 28 positive, 4 negative.

[gazuty]

Looking for inspiration please!

Thanks in advance
regards
Peter

Ok end of my day. Here should be some inspiration for you.

50 markets traded. 41 positive. 9 negative.

Overall fantastic day and almost the equal of a Saturday. All auto.

[Euler]

I think there are two contexts in which to look at this.

If you are on the up curve it seems easy compared to somebody who has been in the market ages. To the older participant they have probably explored and refined many times and are at the top end so it's harder to make progress.

Then, you also have the race to the bottom that occurs in any market, whether it's Betfair, banana imports from Africa, widget making in china. Everybody is seeking an edge and when you find one, somebody else is probably refining it elsewhere. To get ahead they will trim margins a little and that forces you to move, so you probably nip in front of them. So they nip in front of you, you nip in front of them etc etc until both reach the minimum point of return. Everybody has a differing point they are willing to accept.

But also you have to consider the wider dynamics of the market. If it's growing fast there is probably less competition as growth offsets declining returns. But if growth stalls or declines then competition will become more intense.

To keep going for 14 years I have constantly looked to explore and expand the scope and scale of what I do, so thats kept new ideas constantly flowing. As one strategy declines there is usually other coming in somewhere behind. But even then it can only cope with a certain level of participation and also scale has become an issue for me, so that means I have to spend more time elsewhere to get a decent return on all my capital.

Anyhow, I've rambled a bit.

[kelpie]

50 markets traded. 41 positive. 9 negative.

gazuty - sounds good.

For interest and this may further encourage everyone, how many winning markets do you need per losing market?

I would guess no more than two... but maybe yours is different?

[xitian]

I trade both pre-race and in-play on auto, and have been doing so for nearly 4 years now.

I think PeterW summarised it all very well.

My personal profitability has had its ups and downs, both pre-race and in-play. Interestingly, I find when my in-play is bad, my pre-race is better and vice versa. Perhaps that means there's a fixed amount of profit or "inefficiency" in the strategies I employ (since my pre-race and in-play are really derivations of each other), and so if I don't extract the inefficiency in one place I just end up extracting it in the other. That's total speculation though. But in a similar vein, I find days with less races aren't necessarily less profitable than days with lots of races in terms of final absolute profit.

Normally my cycles of ups and downs evolve over periods of weeks or months, although I've not particularly noticed any seasonal trends. There has only been one sudden drop for me which relates to pre-race which was sometime mid-2013, I can't remember the exact time. Otherwise overall (after accounting for PC) my gross profits are still down from my peak in 2011, however that could be because I'm no longer developing my strategies full time (yes, yes, I should have posted in the full-timers thread).

I find that there are two situations which create enough inspiration for me to innovate on my strategies. Firstly, if I'm going through a particularly bad patch, I'll work harder to think of new ideas and try to understand why my existing strategies don't work and try and take advantage of the new market shift. And secondly, if I've taken a period of time away from Betfair trading, I'll come back with a fresh mind and look at things in perspectives I've not considered before.

Back to the original question - Yes, I'm doing reasonably well trading on auto at the moment. Mainly because of work I put in in January (an example of the second situation I mentioned above). I'm on track to have my best month in 2 years, mainly from in-play. My pre-race is so crap at the moment. However this is in part due to a couple of lucky big wins, so it remains to be seen whether it'll remain consistent over the next few months. I want to concentrate on my pre-race now, or start researching football, but I don't have the time.

Back to the full-timer's thing though - I used to be full-time, then went back to the office (I wanted to buy a property), and now I'm planning to go full-time again because I hate the office.

[gazuty]

chart_1.png

gazuty - sounds good.

For interest and this may further encourage everyone, how many winning markets do you need per losing market?

I would guess no more than two... but maybe yours is different?

Losing markets and winning markets for me are graphed above for the last 131 markets I participated in starting at a bank 1000.

Given the fact things are happening close to the off i get three or four a day that go in play without closing. That simply evens out over the long run, but you see the odd large drawdown or spike up above results from going in play. In fact it more than evens out and is running in my favour across about 20,000 events (for some reason, although over the very long term it should be flat).

131 markets traded. 100 positive. 31 negative.

Hope this sparks ideas or inspires other people. Developing a winning automation is one of the most deeply satisfying intellectual pursuits I have experienced.

[PeterLe]

Thanks gents
Some good info and thought provoking stuff. Thanks for the replies

Otherwise overall (after accounting for PC) my gross profits are still down from my peak in 2011..

After Cross matching (XM) was introduced (23rd Nov 2010?)...the next six months produced some massive months...(Notably Jan/Feb 2011), then the profits started to drop back to normal over the following six months. I was surprised (but delighted!) it lasted as long as it did.
Im not sure who was losing, but they must have lost a LOT.
There are times when the market goes haywire for want of a better word. A couple of days a go there were a couple of races at Wolv that seemed to have XM switched off. Having looked at the results Im not so sure. One was a dead heat which is the likely cause.
in that particular race the winner touched 1000 and one touched <1.05. If you do the math, in theory XM shouldn't work above 100 but its another thing in practice.
During the 'boom' months I developed my own version of the kelly criteria that calculated the stake on the fly as the race was in play. It took into consideration the profit I already had greened up plus the what I considered the true odds to be at that split second. (For example if a horse had an avail back price of 30 but was then trading at 1000 with two miles to go, could it really be 1000 to 1? Not in my book. Equally, if two horses are neck and neck with 10 yards to go, would laying at say 1.05 be an appropriate bet? Yes it would, The stake was adjusted and ploughed back in for the next trade within the same race, often placing dozen of bets in seconds etc. I called it "Ratcheting". This was a real defining moment for me.
On one particular profitable system (my sister was running), the stakes ratcheted to over 35 times the initial stake on certain races. NB: Not every race, but when they hit, they hit big. Its not always the most profitable to design a single algo and then apply it to each and ever race, if it is felexible and can adapt to the race in play, then this can add to the bottom line.
From 2012 onwards the frequency of these races where it was possible to Ratchet dropped off.

I too have a pre race trading BOT (its semi automatic rather than fully automatic), Ive had a great deal of satisfaction building that, still work in progress.

Interesting comment from titian about going back to the office too...

Gazuty - Great work. You have a great strike rate and I like the ratio of average win V average loss. Long may it continue for you

What I have found over the years, is that once you develop a profitable initial idea, it leads to further systems. As in the car industry, Its almost like building a chassis and then placing different body types on top, if that makes sense?
At the end of the day , there are only so many variations:-
Back/lay
Lay/Back
A form of bookmaking
A Form of Dutching
This is going to sound a bit obvious, but if you program something to only ever back higher than your lay, no stops and dont trade out for a loss, over a period of time, you will never lose much. from that point its merely a case of refine and repeat.
However there are thousands of variations of the above four. I was running a pre race system yesterday in test mode using a £1000 stake. Out of 18 races, it won 14 and lost on 4. In 3 of the 4 losing races, all three were < £10.
My live in play system still does OK each month. Interesting enough, Ive found my profits have started to increase over the last couple of months as I dont always close the trade now. (ie I enter the market and then dont green up, alternatively, I set the offset to match my expectation of a green (not whats available). If it gets matched, fine, if not, then as gazuty says, it evens itself out over the course of the month). Ive mentioned in other posts that quite often the price is held up in play and only ever comes down if there is the remotes chance of a win. It never used to be like that (Steven 1976 has mentioned this before too). Once you see the pattern, you make a informed decision, easy programmed into the algo

I couldnt agree more than the statement by gazuty - "Developing a winning automation is one of the most deeply satisfying intellectual pursuits I have experienced". I think this is the hook for me. I love the challenge and even if i wasn't making any money from it, I wouldnt be able to walk away from it.

Overall i am surprised that there were only a few who replied. Looking on other forums that use automated betting, its a similar story.
However, onwards and upwards!

[LeTiss]

Overall i am surprised that there were only a few who replied. Looking on other forums that use automated betting, its a similar story

I have lots of different strategies through Excel & BA Automation for identifying opportunities, but then tend to trade them manually after it triggers a nominal bet (outside the range with a kill of 30 secs), simply to alert me to the opportunity.

I became nervous with auto trading after Excel lost me money - this was because Excel has a nasty habit of sticking when there's multiple markets, and it left me with my pants down on a couple of occasions.

I need to actively pursue automating my bets though, as I'm trying to free up my time more

[freddy]

I use automation through excel allot nowadays.

One thing i have learned over the years is that you can not and will not 100% cut out the screw ups when relying on the stupidity of computers ,

so for me rather than wasting time, my strategy's needed to be able to absorb the cock ups.

for example Trading pre race was always difficult for me on auto, if your using 1,000 stakes with say a 10% stop loss, then your risk should be a max of 100-00, but trouble is i 100% guarantee a time will come where a bet fails to fire and you lose the full 1000-00.

So what im saying is rather than stress over this, i tried to embrace it and make it work with that in mind.

[LinusP]

Sorry I missed this thread, I am a big fan of automation! It lets me set up everything in the morning and then let it do its thing. I use very small banks so I don't worry about betfair going down but profits are more than enough to keep me going at the moment.

The thing I have found is that it is very very difficult to get something that works on every race or market. I only have one strategy that gives me a straight line in p/l and that only works on jumps racing using minimum stakes. Any increase past £2 and it all goes wrong, as I only get matched on losers.

Peter, I have never heard of 'Ratcheting', but I think I might look into implementing it on a few strategies. I am having problems with scalability at the moment, any increase in stakes and my p&l nosedives. I therefore focus on achieving a high ROI, most of my strategies get about 10%+ but I turnover under 1k a day.