SHOULD YOU GET OUT WHILE YOU CAN

It depends what you do, are you methodical or random? There's often no profit advantage to greening up over the long haul, but if you want to reduce volatility (the length of losing streaks) then greening up at a consistent point is one way of doing that. Some punters then increase stakes to return double the profit but with the same original volatility.

Assuming you have not over-staked, you should only trade out if it is value to do so, not simply because you can. This is the biggest single reason why traders fail - they cut winning positions short, but let the losing positions run.

thats like saying you should only place a bet on something that wins!

Er, hardly.

cpfc i'm not so sure 'bout that, cutting losers short and letting winners ride...sometimes you might back something, say at 1.5 (at value), it drifts to 1.6...you cut your losses..it wins. Therefore why should it matter, surely all that matters is price relative to chance of the even occuring?. Getting value and then trading out at "correct" price might be a different argument as regards commission.

If you back at 1.5, it should be because the probability of the selection winning is > 0.667. If the price drifts to 1.6, then if the probability is now less than 0.625, you should cut your losses. If it is still value to back, then you should top up.

The point is that you enter and exit positions only when it is value to do so, not simply because you can. Too many traders enter a position, it moves a tick or two in their favour, and they trade out - not because the value is in trading out, but because they can't resist taking that profit. The end result is that you have several small wins, but when a trade goes against you, you watch it drift away into the sunset and it wipes away all the small wins and more.

I agree, i meant in running if you backed at value of 1.5 (say you thought the correct price was 1.47)  and the 1.6 was the current price in running that you also deemed to be correct at that time.

if you green up (or red out) by taking existing offers, you're losing out on the spread twice in the same market.

If it's horses you are talking about , any race under 2 miles get out before the off , if the liability is a front runner get out before the off , otherwise let it go IR and proceed with caution.

Only take value bets. QED.

There is only one correct response to the question "should you get out while you can when trading?" and that is Yes.

The only time you can't get out (apart from when the site goes down) is when the event is no longer open. More often than not this is when the event is over or when it goes in-play and in-play trading is not available. Anyway if you don't get out when you can you are not trading you've just had a bet.

There seems to be an implication in the question that you can't get out of a trade if it means you have to accept a loss no matter what, I find that a dangerous way to trade.

There is only one correct response to the question "should you get out while you can when trading?" and it is not Yes. It is not No either. It is get out if it is value to do so and if it not value then run it.

Hi Alex, apart from the technicality that if you don't get out of a bet then you haven't traded it, there is a problem with your "get out if it is value to so and if it is not value then run it". An example of this could be trading pre-kick-off for football. If you trade then you would trade with much higher stakes than a bet.

Say Man U and playing Chelsea and you think the correct price for Man U pre kick-off is 2.3. You lay them at 2.2 and they drift to 2.24 just at kick-off. If you have a bank of £10K you could easily have a 2K position on this prior to kick-off. Once the game starts are you saying that you would leave that £2k on because you don't think it's value. I certainly wouldn't. Keep trading and betting separate.

I am not sure I can put it more succinctly than I did before UA.

I would trade when I can put up value or take value and that is it.  In your example you seem to be saying that you have over staked and that is why you need to take poor value to bring your stake back to a comfortable position. I would not have over staked in the first place and yes I would run that bet, indeed I might even top it up if the value was sufficient.

Look at each and every transaction rationally as if you had no pre exisitng bets is the logical strategy, imo.

In the example given, as long as you still feel the correct price is 2.30 and there was enough liquidity in the market to lay as much as you wanted when it was 2.20, then you should be closing part of your bet out by backing at 2.24.

Why?  You have just taken a bet with a 0.06 poorer price than you assess the value.  Surely by consistently doing that you must lose in the long run or at the very least dilute profits.

Because it will mathematically increase the long-term growth of your bank by doing so.

Your price at the beginning is 2.30, therefore you would find value in either laying 2.28 or backing 2.32 and there is an optimal amount for each. Once you have laid your optimal amount at 2.20 (which by Kelly is laying just over 3.6% of your bank), you would now find value in either laying a bit extra at 2.18 or backing a bit back at 2.22, amounts of which are trivial to calculate and will each have a positive effect on long-term growth.

Have I got this right?  You advocate taking a bet at a price that represents odds that are poorer than your assessment of the probabilities?

If you have already got your fill at 2.20, then absolutely, yes.

So by taking poor value how does that mathematically increase the long-term growth of your bank? I don't get it.

Hi Alex, I think basically the problem is in the understanding of a bet and a trade.

When betting you are absolutely spot on, in that it is all about value. Everything you do is about finding value bets.
When trading it is all about which way you think the prices are going to go. I read somewhere and I agree with someone who said that they would back an outcome of a coin toss at 1.9 if they truely believed that they could lay out at say 1.8. The only affect that value has when trading is the impact it may have on price movement. This is why when you said in my example that i was overstaking I believe that I am not. Because there is always the intention of trading out before kick-off occurs.

I agree with what you are saying but I believe that all about betting and not trading. This is just based on my beliefs of what trading is. I totally appreciate that different people have different interpretations of what trading is as opposed to betting.
For what it's worth I don't actually believe that trading exists when a football game is in-play, with the exception of half-time. It is all betting. That's just my belief though, i'm sure others will disagree.

Mathematically you should be trying to maximise the exponential growth of your bankroll. This is what Kelly effectively does as it assumes that there you will want to re-use your bank again and again and that you are not having a one-off bet.

Using this as a guide and the example above, if you were correct in your estimation of 2.30 as a fair price, then laying 3.63% of your bank at 2.20 would have a compound average effect of 0.079% on your bankroll.

If you then had the opportunity of backing at 2.24, a back of 1.45% of your original bankroll should increase that long-term growth to 0.092%.

I absorb both of your comments and can summarise this now.

There is only one correct response to the question "should you get out while you can when trading?" and that is NO, unless it is value to do so or if you are recycling.

How is that?

Hi Alex. I've not come across the term "recycling" but if your use of that term equates to my definition of trading and your definition of trading equates to my definition of betting then i would agree with this statement.

I do have a question for you though just out of my curiousity. If you place a bet as per the Man U example above and the price is then always worse than what you consider true value so that bet just stays there until the game is over, would you classify that as a "Bet", a "Trade" or both.

I do have a question for you though just out of my curiousity. If you place a bet as per the Man U example above and the price is then always worse than what you consider true value so that bet just stays there until the game is over, would you classify that as a "Bet", a "Trade" or both.



I guess that if that was the only bet then it would be an outright bet despite an intention to trade.  I am never in that situation as I will have already recycled and traded horizontally and if the price stuck at a price that was out of synch with my assessment of where it should be then I would ask myself if sixty million Frenchmen can be wrong.  But I won't trade out at poor value for the sake of greening unless it is to recycle.

Wonderful stuff Alex.
Might be a bit over their heads though ?

FAFH - ???????

You might well ask.
UA
Let me ask you the following.
Do you think that " trading" might in fact encourage people to overstake, based on the assumption that if things start to go pear shaped they can always get out at a small loss ?
As we've seen though with the latest outages, this may not always be the case.

No its not overstaking because you can get out with a small loss. It's called having another account with with another company froggy, so if something totally strange does happen like Betfair going down, you can trade out with another account. There is gambling life outside BF.

Or are you saying that you should never offer up more than say 2% of your bank on any single outcome just in case the site happens to crash. You may have to clarify a bit as this is currently going over my head!

No I'm saying that even if you're trading you should treat each bet as an independent wager.
If it doesn't make sense on a stand alone basis then you shouldn't do it.
IF you can trade out later at even better value, the all and good.
And that is probably the real difference between traders and gamblers.
If and when even better value comes along the gambler will increase his exposure, whilst the trader will take a profit and move onto other areas of interest.

" -- then all and good "

There's no simple answer to the question of whether or not you should green out value bets- it depends on the exact circumstances and it's not as easy to work out the implications of greening as many people seem to think. 

On the face of it it appears that there are many situations in which you will maximise your returns by greening out a position even though the closing bet may be only true value or less, you can think of this as being like taking only the value portion of the bet and giving the rest back. The way comm works often appears to make greening more profitable, many people will use an example like the following to argue this position:

let's say you can consistently back horses at 2.1 that have a true value of 2, if your comm rate is 5% then the expected return on 100 of these horses if they are straight punted at level stakes is 2.25 stakes (no argument there I hope).

Ardent "greeners" will often argue that if you greened all of your horses out at 2 (true value) then you'd be left with 4.75 stakes after comm, which is more that twice as much as outright punting them. on the face of it that looks impressive but actually it's missing so much of the point that we can call it flat wrong.

That's because this approach doesn't allow for all the horses that you backed at value but nevertheless never got to 2. If you want to make a fair comparison of greened or non greened strategies then you have to allow for that too. This is not as easy as it first appears, normally you can infer the likelihood of a selection going to X price to Y price from it's true price, for example you would infer that a horse with a true price of 4 has a 0.5 (or 50%) probability of matching at 2, but how do you calculate how many horses that match at 2.1 with a true price of 2 will fail to even match at 2? Worse yet, how many of those that do match at 2 will only do so after having drifted out and matched at 2.2 first? How long are you going to wait to green out before you give up and red out? How many of the horses that you could have greened out end up redded because they drifted before they came back in? what is the probability of your lay getting matched at 2 even if the price does go there? All of these things (and more besides) make it extremely difficult to make a simple comparison of most greened and non greened strategies - even when the opening bets are identical and the scenario appears simple. Simple answers to this question just don't cut it, you need to do some fairly intricate maths and be very careful not to miss anything before you can make a call on the best way to play any particular set of bets.

The only examples that I can think of where greening is unambiguously advantageous are when both bets can be struck and matched simultaneously, and even in these strategies there is the probability of "non matches" to deal with.


In my own betting I never deliberately green up but I do very often end up with green markets because the markets that I work in are so volatile that if you're looking for value all over the place then it's likely enough that you find good bets to cover every eventuality. It doesn't bother me one bit to see some red on the screen though.

"If and when even better value comes along the gambler will increase his exposure, whilst the trader will take a profit and move onto other areas of interest." - As mentioned before a trader's primary concern isn't about value it's about which way they think the price is going to move. A trader will move on when he thinks that the price has gone as far as it can and won't go any further, or the game/event is about to start, or because he needs to free up funds because he see's a better trading opportunity.

Traders and gamblers work in a very different but both effective way.

aye robot is that why you chose IR horse markets?

Also robot , in your example who's not to say that of the horses that do drift in don't actually win at a higher strike rate than 50 %, and so you might in fact be losing out on extra profits by greening out in those scenarios.
As you summarize, there are far too many what if scenarios to inject.
But suffice it to say that if you're opening at value and closing at non- value, intrinsically it must be bad in the long run.
Is that at least a given ?
Probably not even that ?

UA
All I know is that if prices open at 1.01 bet and 1000 lay, then bet price will lengthen and the lay price will shorten.
That's why I don't trade price trends pre-off.
I do not have the skills.
As far as trading IR, I fail to comprehend how anybody can predict the vagaries of any game as it unfolds.
People like robot who take advantage of IR horse prices moving too far due to emotions I sort of understand where they are coming from, but all the others no way.

Think of it like a see-saw. The bit in the middle is the value and the bit either side is the bit you want to get on

Actually don't think of it like that

Just when I was getting my head around what you said Rocket, you go and change your mind.
Too much.

I wasn;t happy with the bit about the bit in the middle

But suffice it to say that if you're opening at value and closing at non- value, intrinsically it must be bad in the long run.
Is that at least a given ?

It's not quite a given - there are exceptions. As ever the important thing is to work through all the implications of everything you do. My point is really just that it's a complex problem and no simple explanation covers it.
In my own results I've never found a case for any kind of greening up. Sometimes when I'm experimenting with new bots i get them to actively seek to green up their positions but to date I've always found that the "greening" bets have cost me money and I've cut them in every case.

Rocket: I like the IR GGs for a lot of reasons, it's highly volatile and it has a good liquidity balance, but most of all I like it because I think it's close to impossible for manual punters to comprehend the whole situation and respond fast enough with good decisions. By the time they've seen the whole price board my bots have already anylised it in detail from a dozen different angles and by the time their bets have submitted the situation in the race has already changed. It's bot heaven. To make it even better there are at least 18 races every day, they're rarely simultaneous and the whole thing's done and dusted in just a few of minutes- all of which makes for very efficient use of IT resources.

By way of contrast in football markets the action is relatively slow, the spreads relatively tight and the delay ridiculously long- all of which favours strategic picture players and smart market makers. Plus you have monitor dozens of the damn things all at the same time with (usually) nothing much happening, so that's a lot of refreshing without much betting. Also I just can't be bothered to sit through a whole bunch of football matches watching to see how the betting works - they're just so feckin long.....

To make it even better there are at least 18 races every day, they're rarely simultaneous and the whole thing's done and dusted in just a few of minutes- all of which makes for very efficient use of IT resources.


That's what I was talking about.


Have you put much thought into expanding into football ?

"By the time they've seen the whole price board my bots have already anylised it in detail from a dozen different angles and by the time their bets have submitted the situation in the race has already changed"
But if your bots are following the odds in running and makign decisions on them, then someone (surely manually) is changing the odds that you are following depending on how the race is going? Is that some big bookies with very very fast fingers at the course?

Or a load of punters all heading towards a common target....

AyeRobot: Let's say you can consistently back horses at 2.1 that have a true value of 2, if your comm rate is 5% then the expected return on 100 of these horses if they are straight punted at level stakes is 2.25 stakes (no argument there I hope).

Ardent "greeners" will often argue that if you greened all of your horses out at 2 (true value) then you'd be left with 4.75 stakes after comm, which is more that twice as much as outright punting them. on the face of it that looks impressive but actually it's missing so much of the point that we can call it flat wrong.

Absolutely AR, the "Greener" would have to show a success rate of 96% just to make 0.56 of one stake profit. A success rate of 95% would return an overall loss of 0.49 of a stake according to my reckoning, but I have had a couple of shandy's. Having said that, I do green my positions, but only to reduce volatility, nothing to do with profits.

AyeRobot: I like the IR GGs for a lot of reasons, it's highly volatile and it has a good liquidity balance, but most of all I like it because I think it's close to impossible for manual punters to comprehend the whole situation and respond fast enough with good decisions. By the time they've seen the whole price board my bots have already anylised it in detail from a dozen different angles and by the time their bets have submitted the situation in the race has already changed. It's bot heaven. To make it even better there are at least 18 races every day, they're rarely simultaneous and the whole thing's done and dusted in just a few of minutes- all of which makes for very efficient use of IT resources.

By way of contrast in football markets the action is relatively slow, the spreads relatively tight and the delay ridiculously long- all of which favours strategic picture players and smart market makers. Plus you have monitor dozens of the damn things all at the same time with (usually) nothing much happening, so that's a lot of refreshing without much betting. Also I just can't be bothered to sit through a whole bunch of football matches watching to see how the betting works - they're just so feckin long

Very good descriptions, I'm agreeing with you left right and centre tonight, hehe. You're reasons for disliking the football markets are my reasons for liking them - the markets move much slower which gives us manual punters time to assess, react and adjust comfortably, without the need for manic input.

So going back to the original question, my strengths lie in football markets and my weaknesses in horse markets, which I therefore quit a long time ago. Oh no, that's a different thread :)

You're being Your :)

AyeRobot: Let's say you can consistently back horses at 2.1 that have a true value of 2, if your comm rate is 5% then the expected return on 100 of these horses if they are straight punted at level stakes is 2.25 stakes (no argument there I hope).

Ardent "greeners" will often argue that if you greened all of your horses out at 2 (true value) then you'd be left with 4.75 stakes after comm, which is more that twice as much as outright punting them. on the face of it that looks impressive but actually it's missing so much of the point that we can call it flat wrong.

Trevh: Absolutely AR, the "Greener" would have to show a success rate of 96% just to make 0.56 of one stake profit. A success rate of 95% would return an overall loss of 0.49 of a stake according to my reckoning, but I have had a couple of shandy's. Having said that, I do green my positions, but only to reduce volatility, nothing to do with profits.

Trevh, perhaps when you've sobered up you can review your post and provide a more expansive justification for those, seemingly completely erroneous, figures.

With regard to Aye Robot's post that you quoted above, I disagree entirely with his description of the 'greener's' claim as 'flat wrong'. The claim is preceded by the word 'if'; an explicit acknowledgement that it may not be possible to trade out at the correct price. Given this, the 'greener's' claim appears to be entirely accurate.

I'll assist:

The 'greener' has backed 100 selections, each with a 50% probability of winning, at the price of 2.1 (a price with a positive expectation after commission). Regardless of how many of those selections he succeeds in laying off at the correct price of 2 within the same market, whilst the probability of the selection winning remains at 50%, his bets would have a positive expectation.

Whilst Aye Robot mentioned the possibility of the 'greener' laying off bets at a price higher than the original back price of 2.1, you made no reference to this in your post above. Without a specific reference of this nature (detailing the average lay off price and the number of selections to which this applies), in my opinion you cannot possibly justify the figures you stated above.

Regardless of how many of those selections he succeeds in laying off at the correct price of 2 within the same market, whilst the probability of the selection winning remains at 50%, his bets would have a positive expectation.


That's not quite right- his OPENING bets would have the same positive expectation but this may well be wiped out by his greening bets. Remember - by the nature of it he's much more likely to be able to lay against those bets that win than those that lose.

Not if he lays off at the correct price of 2...whilst the probability of the selection winning remains at 50%.

I understand what you're saying JT and it's right that if the true value of his lays is 2 then they should come out clean, but my point is that he must be more likely to lay off his winners that his losers - so if he simply lays off all the horses that match at 2 and leaves any that don't he will inevitably reduce his return.

Aye Robot,

I agree with your post above but in that situation many of his lay bets will be taken when the probability of the horse winning is regarded to be greater than 50%. That is why I stated succeeds in laying off at the correct price of 2 within the same market, whilst the probability of the selection winning remains at 50%. I accept that this would be very difficult to achieve in practice. 

On a similar note, I largely agree with your post highlighting the difficulties in attempting to use the 'ardent greeners' suggestion. Nevertheless, the theory is correct and I therefore do not accept that it is reasonable to describe it as 'flat wrong', simply on the basis that it may be difficult or even impossible to put into practice.

* Assuming that the probability of each of the horses that he backed at 2.1 winning remains at 50% [;)]

Sorry for not replying sooner froggy but had switched my comp off before getting your reply. Re your last post addressed to myself, this is the point that I am trying to say, in that traders and gamblers have very different skill-sets. One is concerned with value and one is concerned with price movement. This is the problem with Aye Robot's post is that it's from a gamblers perspective not a traders perspective, i shall try to give an example to illustrate my point.

When Bolton played Everton last Sunday the draw price at about 15 mins before kick-off was 3.6 back and 3.65 lay.
A gambler may well look at this and calculate that the value price is 3.4 (this was what the price was last season for example) so I will back this.
A trader may well look at this and say that the draw price has been going up, i think it will continue to do so right up until kick off and so will look to lay.
They are both doing opposite things so who is right? Answer is they both are. The draw price did go up to 3.75/3.8 just before kick-off at which point the trader would have traded out and made a profit. The gambler is right because he got good value so over the long term can expect to make a profit.

My point is that the original question refers to trading and not gambling, so to answer it from a gambling perspective is not necessarily correct. I do however totally acknowledge that this is all dependent on how you define the term "trading" and "gambling" and that other people (maybe even everyone) have different interpretations to myself.

As such I shall ask the question to FAFH (as I know he would like to answer it anyway) Aye Robot, Alex, the original poster and anyone else who cares to venture their opinion. How do you define the term "trading" and the term "gambling".

UA
I think a trade is just a gamble in that you are speculating on a trend, which is really no different from speculating on anything else.
The commission implications are slightly different, that's all really.

ok FAFH i know what you are saying maybe i should ask for definition of the term "Trading" as opposed to "Betting" or do you define betting and gambling as the same.

whats the point in getting on if all u want to do is get off

Why do I all of a sudden start thinking of suitable forum names ?

jt45: Trevh, perhaps when you've sobered up you can review your post and provide a more expansive justification for those, seemingly completely erroneous, figures.

With regard to Aye Robot's post that you quoted above, I disagree entirely with his description of the 'greener's' claim as 'flat wrong'. The claim is preceded by the word 'if'; an explicit acknowledgement that it may not be possible to trade out at the correct price. Given this, the 'greener's' claim appears to be entirely accurate.

Still drunk as a lord here jt, but my calculations were as follows...

100 back bets at 2.1 for 1 stake (£10).

Green at odds of 2.0 for 0.05 stakes profit (50p).

Success rate of 96% = 96x50p = £48 minus 5%comm = £45.60 minus 4 losses @ 1 stake (£40) = £5.60 profit or 0.56 of 1 stake.

Success rate of 95% = 95x50p = £47.50 minus comm = £45.12 minus 5 losses @ 1 stake (£50) = -£4.88 loss or 0.49 of 1 stake.


No doubt you will say that the 'greener' will have redded up the 4 losses that didn't reach 2.0, but that wasn't in the original comparison between letting them ride or greening them at 2.0, and if they were horses that stalled or fell at the off that option wouldn't be available anyway.

I don't understand why you're singling out the word "if", as obviously not all of the bets would drop below 2.0, "if" was used to make a sentence, you could substitute "if you greened" for "greening" if you like.

Trevh,

In the original comparison, Aye Robot didn’t appear to suggest that any unmatched lay bets were to go IR.

You have assumed that the unmatched lay bets go IR.  Presumably, on that basis, you have also concluded that the 4%/5% of selections which the greener didn't successfully lay-off must, therefore, be losers.*

Without the assumption that unmatched bets are to go IR, your figures are nonsense. That is what I originally concluded. Had you stated that unmatched bets were to go IR, I wouldn't have disputed your figures. I remain unconvinced that your analysis represents a fair comparison.

If the unmatched bets do not go in play, every bet he succeeds in laying off, at the correct price of 2, enhances his expected profit in comparison to allowing the original bet to run its course.

For example, again assuming 5% commission and no premium charge,  if he successfully laid 50 of his 100 original bets at the correct price of 2 pre-off, his expected profit would be (50*0.05*0.95) + (25*1.1*0.95) - 25 = 3.5 units. This compares very favourably to laying off no selections for an expected profit of 2.25 units.

With regard to the other matter:

if you greened all of your horses out at 2 (true value) then you'd be left with 4.75 stakes after comm, which is more than twice as much as outright punting them.

In my opinion, in the above statement the greener is implying that if you could trade out of all your selections at the correct price of 2 then it is advantageous to do so for exactly the reasons given. He is not implying that it will be possible to trade out at the correct price of 2. Therefore, the greener's statement is entirely accurate.


* I have ignored the possibility of one or more of the horses on which the lay bet was unmatched, prior to or during the course of the race, being unexpectedly awarded the win due to the disqualification of the first past the post. I presume that you now accept that this is possible, albeit very unlikely.

Frankly I think I worded my earlier post poorly - it was a bit ambiguous. My point was really just that a very simplistic view of the advantages of greening up is sufficiently flawed to be useless and that without a sufficiently sophisticated and specific analysis it's impossible to say. I think we all agree on that.

A better wording would have been;

Ardent "greeners" will often argue that you would increase your overall return by greening out all of your horses out at 2 (true value) which would leave you with 4.75 stakes after comm, which is more that twice as much as outright punting them. on the face of it that looks impressive but actually it's missing so many of the relevant factors that it's simply not a worthwhile argument. It's equivalent to saying that it doesn't matter what odds you bet at as long as you win.


That's what I meant but it's not what I wrote - sorry about that.

It's fairly simple maths when it comes down to it and I'm always a little surprised that these discussions generally end inconclusively.

The larger your perceived edge, the larger open position you should have relative to your bankroll. Therefore if you open a position due to positive value being available and then that edge decreases, either due to the market price moving towards your price or the other way round, then the size of your position should also be reduced. If the market price is equal to your price and you have an open position, there will always be a long-term gain in fully closing your position.

jabmast
Aren't you assuming that your strike rate actually equals your " perceived " edge ?
It could in fact be better or worse. Who is to know ?
Don't the " fairly simple maths" change then ?

As long as it is measurable (i.e. your actual edge against perceived edge) then it can and should be incorporated in. The only way to measure that is over time, but once you have an indication of how reliable your edges are then the maths can be adjusted.

And just when you think you've established how reliable your edge is, it starts to become unreliable.
In hindsight the maths is easy I agree, but otherwise ?
Btw to all and sundry who say that it makes sense to get out at a profit all the times you can, does it equally make sense to get out at the first sign of a loss ? What's the " fairly simple maths " on that ?

FAFH, it doesn't matter whether you are getting out for a win or a loss, it's about whether your edge has increased or decreased. The market may have gone for you or against you, but if your edge has increased you should be opening up and if your edge has decreased you should be closing out.

Admittedly if your edge has suddenly disappeared, then clearly your results will be affected, but they would be whichever way you were betting.

But how do you know such trends are not just momentary and can reverse very quickly ?

Btw I've only ever seen convincing maths supporting " greening out" on a retrospective basis.
That is persons go back and analyse the price movements after the fact and say that this and that time were clearly the optimum times to take profits or cut losses.
In real time situations all is not so cut and dried unfortunately.
Many persons tend in greening out to take profits are the first sight of them and conversely to cut and run at the first sight of any losses.
Resultantly they end up switching in and out inefficiently and to their ultimate overall net loss.
Can I mathematically prove that simply ?. No way I'm afraid to say. But can anyone disprove it simply ? No way I also say.

The only thing that is subjective in this is the perceived edge. If you have a high confidence in that (or a %age consistent confidence in it) then the maths I don't believe to be subjective.

Give me a P&L, a bankroll, a calculated price and an available market price and I can calculate what to do to maximise long-term growth of bank.

Godd for you jab.
I suppose we'll never know if you're right or wrong till we get to the end will we ?

Btw perhaps the point maybe really is not that if you should green or red out out but when. And please don't say that you only do it in proportion to the % movement for or against you.
I real life, most often prices are fluctuating around like a yo-yo, particularly if circumstances are very fluid and fast changing.
Constant changing of one's mind will result in the back/lay spread killing you.

The combination of do it in proportion to the % movement for or against you and most often prices are fluctuating around like a yo-yo I find results in quite the opposite of will result in the back/lay spread killing you. 

If the proportions are right (easily calculable and happy to show you how) then this sort of volatility is exactly what you are looking for to maximise the opportunity availed by having an "edge".

No need jab
I'm not a trader as must be obvious by now.

If the proportions are right (easily calculable and happy to show you how)

Go on then Jabmast - If a horse is trading at 2.1 when true value is 2,what is the probability of it matching at 2 before the off?

And after the off?

Give me a P&L, a bankroll, a calculated price and an available market price and I can calculate what to do to maximise long-term growth of bank.

Kelly already did that for you.

aye robot, that's not what I'm saying. As cpfc4me states, Kelly (or an extension of to include current position) allows you to do what I say. It's not about predicting where the market will go, but knowing what to do when it gets there.