Selectivity vs turnover

[Derek27]

It all depends on whether you have a strategy that works best on certain markets, whether it's applicable to any market, or if you can adapt your strategy (or have a different strategy) for any market.

[ShaunWhite]

Two types imo
1. Execution psychology (saturated)
2. Analysis psychology (bone dry – because the answers are hard!)

I get it. I should have realised the distinction.

Speaking of anaysis and analysis psychology, I've just been running some what-ifs on a project that I must have spent >90 hrs on....

Collecting data in a whole new way. Adding additional non-obvious content. Writing a new test rig with everything parameterised incl independant entry and exit triggers and including different trading styles (offsets, reverse prices, stops etc). Running tests to validate my data. New live monitoring and automatic bet placement engine, the whole 9 yards and then another 9.

My machine is just cooling down and the end results is.....the usual gentle slope downwards Unless you assume hypothetical matches at reverse prices on entry and/or exit, those always look much better on paper.

The initial charts looked so promising, the concept was given the wink by some serious guys, I've been excited about it for about a month....Now it feels like an absolute *&%^*!&" waste of time.

I 100% agree that info on this whole area, technique and interpretation, is very hard to find.

I'll keep looking, there must be something. You can't just stop can you

[ShaunWhite]

This what I always disliked about coding, switch off and there's nothing to see, and sometimes nothing worth keeping. At least when I spend 90hrs in the shed I have something nice to show for it. Is there any wonder it's a 'kin struggle booting up somedays.

[ShaunWhite]

lol sorry again stuey. Personal winge finished.

[Euler]

I think the answer is quite simple really.

When I first started trading I went for the one fits all approached and just tried a generic strategy across all markets. I then started cutting out errors, basically where my biggest losses were, and that led me to a selective approach for that strategy.

I would then try another strategy on just the markets that didn't seem to work and repeat the process until solved.

I now have hundreds of variations of individual strategies aligned to specific markets and scenarios.

I would recommend focusing on one problem until you solve it then branch that to new variations.

[spreadbetting]

I would recommend focusing on one problem until you solve it then branch that to new variations.

I think that's that's a very good point especially for people starting out. Usually they're trying far too many things at once and listening to well meaning but useless advice on the whole. So many times people ditch working strategies just because they're hidden amongst the poor trades they're also doing but they don't have the patience to separate what they're doing that works from what doesn't.

I'm sure we've all been there at the start with far too many ideas bussing around but you really do need to take things a step at a time to find what's working and just as importantly why you think it's working. Once you start to figure out how and why markets move it's much easier to find other things that'll work for you also.

[ShaunWhite]

I then started cutting out errors, basically where my biggest losses were, and that led me to a selective approach for that strategy.

This approach only works when the initial 'random' strategy is profitable on a non random set of data.

If you filter the results by all possible combinations of volumes, market profile (prices), time ranges, number of runners, distances, types etc etc, and no subset is significantly better than any other, where do you go from there? And where next when the best or worst performers don't have anything in common?

I'm certainly not doubting your approach, it's clearly extremely effective and without it I'll struggle, but it gives the impression that almost any strategy fits a certain set of markets that can be singled out in a logical way with the data we have at our disposal.

I'm so convinced that my approach is a logical conclusion to real world events that I don't want to share it openly. But unless I can find someone who'll take a look at it and give me some guidance, I'm snookered.

[ruthlessimon]

If you filter the results by all possible combinations of volume..... and no subset is significantly better than any other, where do you go from there? And where next when the best or worst performers don't have anything in common?

I'm certainly not doubting your approach, it's clearly extremely effective and without it I'll struggle, but it gives the impression that almost any strategy fits a certain set of markets that can be singled out in a logical way with the data we have at our disposal.

That's a really good post & exactly why I'm delving into "Target shuffling" etc.

Let's assume hypothetically, that I was working on a pre-of trend based lay strategy.

1. The losers median entry time was 215, & the median total market volume on entry was £126k
2. The winners median entry time was 199, & the median total market volume on entry was £154k

Is that then enough for a trader to say, "when the volume is higher, that predicts trend-based strategies will perform better?" - "I'm only going to trade this strategy when volume is at least £150k?"

A trader could keep on adding & adding variables (does race type influence this, does day of the week influence this), & eventually, something (by sheer luck) will look amazing - but still provides little real predictive value - & that's my issue

[ruthlessimon]

I think the answer is quite simple really.

When I first started trading I went for the one fits all approached and just tried a generic strategy across all markets. I then started cutting out errors, basically where my biggest losses were, and that led me to a selective approach for that strategy.

You didn't look at the "niggly losses"?

Here was a basic strategy I employed through May - I thought it was interesting that summing the "minor losses" came to -44 ticks nearly double the worst loser. Therefore cutting out trading the "50/50 markets" could boost that strategy just as much as the huge losers - possibly more. Or rather than improve the strategy, just reverse it?

[Greco]

Euler has spoken of the Pareto principle before and it is something that is illustrated by your point Simon. I use a price range filter for my trading. Yes I have profitable trades outside that range but I also have a lot of small losers as well. Therefore I have simply removed them and concentrate harder on the ones that make me the most money.

[FrogThimble]

I would recommend focusing on one problem until you solve it then branch that to new variations.

I wish I'd realised that 7 months ago when I started.

[ShaunWhite]

I would recommend focusing on one problem until you solve it then branch that to new variations.

I wish I'd realised that 7 months ago when I started.

Unless it's unsolvable, in which case you never move onto a second problem.

[FrogThimble]

Unless it's unsolvable, in which case you never move onto a second problem.

Absolutely! That's certainly a factor that held me back for a long time. Sometimes it's important to realise that a problem has no solution and just move on. Or totally change direction. I've made more money here in the last 6 days than in the entire previous 6 months - all because I finally stopped fooling myself about which problems I could solve.

[ShaunWhite]

Sometimes it's important to realise that a problem has no solution and just move on.

That's another tricky part. Especially if you're the sort of person who won't be beaten just because something's difficult, and when you're told that almost every strategy can be refined by discarding the negatives and improving the positives.

[Euler]

Every market has an element of entropy, but you can 'solve' that as well.