Reversing my strategy = Reversing the Results ?

[gums093]

Hi guys,

i would share some thoughts here,

i am currently testing an automated strategy in practice mode since one month, it's a Pre Race classic Swing trade strategy, when my signal is detected on the market guardian will trigger a Laying or a Backing bet depending wich trend is detecting (up or down) . I monitor the 3 fav and the signal is always the same for every horse. its really a simple swing trade detection triggering a back/lay bet.

For now, my bot traded almost 150 races. the results are horrifics and show up a very clear negative trend with almost every days finishing with losses. i use 10€ stakes and the results are are a -45€ Loss.

So as the trend is very strong i see it as an opportunity, it make sense to me that if i reverse my strategy, meaning i keep the same signal but i reverse the action triggered on that one (Backing instead of Laying and Laying instead of Backing) i would reverse my results trend as well.

are you agree with that theory or it is not as simple as that ?

[LinusP]

Reversing would assume you can back at what you lay and visa versa which therefore involves sitting in a queue.

[ruthlessimon]

It's tricky, because a negative edge is just as hard to find as positive one. The most likely scenario is that, as your sample increases, you'll tend to just under breakeven - although 150 is already good sample size. I remember a while back Peter was genuinely interested in hearing about consistently losing methods:

viewtopic.php?f=42&t=5322

[jimibt]

to add to what's been said earlier, switching to a lay from a losing back will not guarantee a change in fortunes. This is due to the fact that on a back based strategy, you are potentially placing bets some way above odds of 2. the £45 loss on backs will pale into insignificance when you realise that laying the same set of runners will incur a far higher liability. therefore, ironically, depending on how far out those odds are, you may find that a lay based strategy using the same signals will produce far worse losses than the back based approach... strange but true.

in order to make headway with this, you'll need to discover a quantifiable edge that is opinionated in terms of being back or lay based. you should then test that across at least a months worth of markets. even then, you should ensure that the data that you collate is seasonally relevant (i.e. no point having 2 weeks of jumps butting onto the start of the flat season etc).

fwiw - i reckon it's taken me the best part of 12 months to reconcile the fact that a finger in the air is no competition when comparing long term trends in the market that accommodate, season, distance, event type, field size, race makeup (in terms of short fave, contesting faves etc) and other metrics. as this is pre-race, your approach will differ significantly to sizing things up for IP - another ball game but one that also responds well to strong, consistent stats.

hope this pings a few twigs... i know it did with me on re-reading ahead of clicking submit.

[ShaunWhite]

You're only losing about 30c (3%) a race. The reason you can't flip it is the spread (diff betwen back and lay price).

Even if you could flip it and get the same prices, which you can't, then you'd have the commission to take off which spoils the party.
You'd need to start with a strategy much worse than the 'horrific' one you've found.

If I could flip every losing strategy I've found, I'd be typing this from my private jet. And I've had lots that are way way worse then -3%.

To be serious a while, I'd be looking to gently tweak a -3% strategy rather than flipping it. Take another look at jimibt's reply, and remember the saying that every market has a strategy and every strategy has a market. I don't have a whole heap of stats like jimi, linus and simon do, so I can't tell you where to look.

[gums093]

Thanks for the wise words. on testing purpose on that theory i made it a try today, i put the reverse strategy, and at the end of the day the result was.....positive with profit. not so strong as it was on negative results on the last days but in the regular range of the negative daily results.

But it was just a try for fun, not really significative at all.

what i am doing from now is implementing thes opposite strategies on 2 automations bots with 2 differents accounts on 2 differents machines, each one with each strategy on the same markets independently. And let's see what's happening there.