Strategy Development: Modelling

[spreadbetting]

Here was my reasoning:

If there are thousands of people walking down a path, & there's £20 lying on the floor - it took no skill to get that £20, & was simply a case of you being the first to spot it. (That's where I feel I am - deploy a random strategy, find something half decent, now what ). I don't think it's sustainable - I wish it was.

On Betfair it's not neccessarily even being the first to spot it, just being the first to bend down and pick it up usually works fine.

I've always thought this Simpsons clip is quite apt when it comes to Betfair

https://www.youtube.com/watch?v=yR1QL7CJltQ

[northbound]

On Betfair it's not neccessarily even being the first to spot it, just being the first to bend down and pick it up usually works fine.

Perhaps once I’ll manage to get out of this “break even grind of death”, I’ll agree with you.

So far, almost all of my strategies tend to revert to the mean in the long term.

[spreadbetting]

If you've managed to break even after commission you must be doing something right, all you need to do now is figure what you're doing that's making money and the reason why it's making you money.

All I was trying to imply with the simpsons video is that the "sharp minds" tagline is just a load of spin. There may well be a lot of "sharp minds" on Betfair but thankfully there's plenty more thick minds who continually litter the site with more money than the sharp minds can pick up, especially if they're waiting for another 5 conditions to be met before going in.

[northbound]

If you've managed to break even after commission you must be doing something right, all you need to do now is figure what you're doing that's making money and the reason why it's making you money.

That's the tough part.

If I may ask, how often do you have to ditch winning strategies (proven over hundreds of trades) because they start reverting to the mean (over the next few hundred of trades)?

[Euler]

(sorry, I'm in shed mode, wife's away till Sat so i'm funning myself to death with non-stop Excel, metalwork & pizzas )

Comment of the week

[ruthlessimon]

There maybe many possible ways to retrieve the £20, the most elaborate may be recruiting a team of engineers; roadworkers and JCB's. Surrounding the area off (with a statement of work and safety assessment in hand). Digging the area and carefully collecting the money.
A simpler method may involve a large bamboo stick, with a piece of chewing gum on the end and 20 seconds work, quickly moving on to the next manhole.

That's more than just a decent analogy, next time I see some money down an actual manhole I'll probably use your chewing gum method

But yeah, I'd call that genius, flair, talent - to be able to spot ways to "remove the red tape" of a strategy/process. Having that sort of initiative is something I try to improve daily

[Euler]

Got to say, this is a great thread and a lot of the comments from a number of posters echo my thoughts exactly. You would be surprised how reckless the market is at leaving money on the table. I quite like chaotic markets because when people are running around like headless chickens I'm just sat there dodging the oncoming traffic and picking up the litter they leave behind.

[ShaunWhite]

I'll file that in the If you can keep your head when all about you are losing theirs.... chapter.

Thanks for contributing, especially while you're busier than a five dollar hooker. Every sentence really does get all the attention of another fragment of Rosetta stone by your tifosi as you can probably tell.

[spreadbetting]

If you've managed to break even after commission you must be doing something right, all you need to do now is figure what you're doing that's making money and the reason why it's making you money.

That's the tough part.

If I may ask, how often do you have to ditch winning strategies (proven over hundreds of trades) because they start reverting to the mean (over the next few hundred of trades)?

I'm the wrong person to ask , I wouldn't have a clue where my mean even started from to be honest. My approach is more based on looking for value than looking for patterns. Like I said earlier in the thread, it's easy to backfit winning strategies by looking thru old data but they rarely hold up in real life though.

[stueytrader]

In my past life (mostly as a punter not trader) I have used regression models for sports such as football goals markets.

Exactly the point above though: the past does not always (enough) predict the future, that's the massive issue. Plus, it's already factored by others who did the same analysis as you.

Seems the secret is to only analyse those factors or areas that others have less interest in, but then they can become weak factors to use.....

[stueytrader]

However, I do think that trading is rather different, because we're mostly trying to predict movement not outcomes, and this can (IMO) be very specific to each specific selection that you are trading. More room for factors to go unnoticed by others looking at the same market, and of course more choice of which/when/how to trade it around, given the continuous flow of prices and changes. That's got to be good news.

[ruthlessimon]

By having extra layers of complexity hedges that risk

I find the opposite is true.

Yep agree - check out https://en.wikipedia.org/wiki/Occam%27s_razor

The criteria of this strategy ("optimised" - found randomly via excel) literally could be written down in 3 words. (April & May sample)

My issue is, it's so simple anyone could find this.



Anyone could make a bamboo stick with chewing gum after seeing the first guy. However, if my saliva made the chewing gum stay sticky far longer than average Joe, suddenly the edge is inherently linked to myself - & impossible to replicate. I feel that's what adding complications will bring - assuming the above is an edge Delaying the inevitable erosion of the edge, & increasing profitability.

[ShaunWhite]

My issue is, it's so simple anyone could find this.

Do you want to spill the beans ? Then we might understand more about the example and if what you're saying is relevent.

[ShaunWhite]

You would be surprised how reckless the market is at leaving money on the table. I quite like chaotic markets because when people are running around like headless chickens I'm just sat there dodging the oncoming traffic and picking up the litter they leave behind.

This an area where sports markets provide opportunity where the financials don't. All significant money in financials is smart money, smart enough anyway not to leave scraps on the table. Sports markets have a much higher ratio of less smart money.

It's tough competeition to get it though with preditors like Mr Webb around. Killing seems easy when you're a tiger, not so easy when you're a cub.

[ruthlessimon]

Do you want to spill the beans ? Then we might understand more about the example and if what you're saying is relevent.

What's interesting is; removing the bias; & trading every other market using the criteria; kills it (looks random imo)



Ironically it's been slightly boosted by including every market (May only)

However, if I wasn't aware of the bias & had it merged it looks like this - most people would probably call that an edge, & stop:



But "increasing the complexity" - highlights that there was a massive driver:



Perhaps I've come to this conclusion backwards. & if I had looked at the losers, the spotting the potential for refinement would have just appeared?