Forums

General Betting

There is currently 1 person viewing this thread.
cpfc4me
08 Sep 13 16:34
Joined:
Date Joined: 30 Mar 04
| Topic/replies: 639 | Blogger: cpfc4me's blog
Do any traders who keeps detailed records have the ratio for their average loss to average win size? I suspect that most traders cut winners too short and let their losers run which would result in the average loss size being larger than the average win size. On the two sports I trade most, my win size is 87.4% of the loss size, (NBA) and 72.9% (NFL) which suggests this is true at least for me, but having posted on my blog about it, the number of replies 24 hours later is precisely zero so I thought I would try for some comparisons here.
.
http://green-all-over.blogspot.co.uk/2013/09/average-losswin-ratios.html

Post your reply

Text Format: Table: Smilies:
Forum does not support HTML
Insert Photo
Cancel
Page 1 of 2  •  Previous 1 | 2 | Next
sort by:
Show
per page
Replies: 61
By:
JLivermore
When: 08 Sep 13 18:03
just wrote a verbose opinion on this and then clicked wrong button, so I'll give concise one.

the reverse logic of
"most traders cut winners too short and let their losers run which would result in the average loss size being larger than the average win size. "
does not hold
(average loss size being larger than the average win size DOES NOT IMPLY cut winners too short and let their losers run)

Average Loss:Win Size Ratios captures skewness which just depends on the type of bets the trader tends to make (for example whether he tends to profit or lose from goals)
By:
cpfc4me
When: 08 Sep 13 18:39
Football isn't a true trading sport in that the price movements are pre-determined, re-setting after suspensions - i.e. goals or significant events. That's why my numbers are from sports where the prices move more gently and genuine trading takes place.
By:
JLivermore
When: 08 Sep 13 19:26
I'm not sure that changes my opinion.
All bets (apart from @ 2.0) are skewed, so you'd expect profits to be skewed depending on the kind of strategy the trader pursues
By:
cpfc4me
When: 08 Sep 13 20:42
I'm really looking at a 'proper' trader, who is in and out of a market several times in a game. My theory is that (for example) a novice trader who makes the makes the mistake of cutting his winners short and letting his losers run will see several small sized wins offset by the rarer, but larger sized, loss. His average win size will be a lot smaller than his average win size.

It makes no difference if you typically trade at odds-on or odds against, since a trader is simply trading the movement. 10 ticks is the same whether you back at 1.5 and lay at 1.4 or lay at 3.1 and back at 3.1. The skewed statement refers more to a punter (rather than a trader) who may prefer lots of winners at low prices or the occasional winner at a long price.
By:
JLivermore
When: 08 Sep 13 21:07
I understand what you're trying to get at, I was just saying I'm not sure it's a good target to aim for.

I still think the prices will explain and win:loss ratio (mainly because probabilities are bounded between 0 and 1, causing skew in returns), but if you trade sufficiently small price movements then I'll agree it wont be the dominant factor.
By:
askari1
When: 08 Sep 13 23:15
I can't speak for football, but supposing I am purely trading the horses my loss size is more than 3X bigger than my win size, but I'll win about 78% of the time.

In practice I will expose myself to big reds on judgment-based plays and will edge backs, so I don't have a good measurement of this ratio.

My sense is that there is a psychological benefit to someone who on average takes positive expectation bets of winning most of the time. Typically when I evenly red out a losing trade I'll admit I don't understand the market and move onto the next.
By:
Darlo Bantam
When: 09 Sep 13 02:41
Average win:loss size ratios can only be compared properly when set alongside average win:loss number ratios.
By:
cpfc4me
When: 09 Sep 13 06:05
Darlo - we are talking about profits and losses from trading events, and the one ratio (not ratios) is for the average loss size compared to the average profit size per market, not per trade. The number of winning or losing trades is irrelevant for this discussion and is not needed. I'm not asking if people are profitable or not or by how much - only how close to 1:1 their average loss size : average win size is.
By:
Darlo Bantam
When: 09 Sep 13 16:40
The number of winning or losing trades is irrelevant for this discussion and is not needed.

I can't agree there. If you are trying to make lots and lots of basically 1.01 trades, then you need a high win:loss ratio. But if you are trying to make the likes of a "double or loss" trade then you need a 50% ratio to break even. So the converse can also apply.
By:
JLivermore
When: 09 Sep 13 17:20
I have a question for you OP:

When you're going about 'trading the movement'  do you aim to:
1 - have an edge on a equal move in price (ie a 60% chance of making 10 ticks versus a 40% chance of losing 10ticks)?
OR
2 - have an edge on the price movement itself (ie a 50% chance of making 10 ticks versus a 50% chance of losing 5 ticks)?


I ask because the way I see it is:
If you are restricted to (2) then I can see why you are interested in win:loss size ratio (since number of wins=number of losses).
If you are trading with (1) in mind then win:loss size ratio is going to deviate from 100% just because of your strategy (and looking at it ignorant of number of wins/losses would be missing out valuable info).

I agree with Darlo:
"Average win:loss size ratios can only be compared properly when set alongside average win:loss number ratios."
By:
cpfc4me
When: 10 Sep 13 01:20
I enter a market when it is value to do so and exit (or switch sides) when the value is to do so. In theory.

In practice, I believe I am cutting my winners a little short or letting my losers run a little too long, as evidenced by my average loss size exceeding my average win size.

My theory is that a 'perfect' trader will, over a long period of time, have a win size to loss size ratio of close to 1:1 - but with more wins than losses.

Of course the average win size * number of wins will exceed the average loss size * number of losses for any successful trader, but how much people make is not of interest here, only the ratio of loss sizes to win sizes, which very few people seem willing to share, it has to be said.
By:
JLivermore
When: 10 Sep 13 07:52
"I enter a market when it is value to do so and exit (or switch sides) when the value is to do so. In theory."
So some of your trades will likely have skewed returns (ie type 2 bets in my example).

"In practice, I believe I am cutting my winners a little short or letting my losers run a little too long"
Probably - this is a well known behavioural bias seen when trading/investing (from finance, although generally over >1 week periods)

"My theory is that a 'perfect' trader will, over a long period of time, have a win size to loss size ratio of close to 1:1 - but with more wins than losses."
still don't agree. 

"Of course the average win size * number of wins will exceed the average loss size * number of losses for any successful trader, but how much people make is not of interest here, only the ratio of loss sizes to win sizes,"
If I was you, and I suspected I was making errors and that I should have no skew in profits I would simply look at SKEW (you can calculate in excel) of profits and plot it over time to see if I it was becoming closer to 0.

"which very few people seem willing to share, it has to be said."
I'd share, but I'm not a trader (by popular definition).  I know my average win / avg loss is >1 though (although I have no idea what it is).
By:
mardock
When: 10 Sep 13 11:17
During my 3 years full time trading experiences I have following stats:

1. After over 3000 markets i have 72% positive and 28% negative markets
2. My average profit = 82% of my average loss.
By:
Bayes.
When: 10 Sep 13 12:15
In tennis, my main trading sport, I win in about 60-65% of markets and my average win to average loss ratio is very close to 1. Four or five years ago this ratio would have been nearer 0.8 but I am much better nowadays at staying 'at market'.
By:
TheInvestor2
When: 10 Sep 13 23:28
Looking at it per football match is probably most interesting for me.

For the first quarter of 2013, my average winning match gives me about 88% of the average losing match, and I won on 75% of matches (counting manual trading only). That's was a pretty good quarter though.
By:
Thin and Crispy
When: 11 Sep 13 13:59
Horse racing pre off win/loss 3 to 1 ratio...can't be arsed to work out the figures it would only depress me.
By:
Bayes.
When: 11 Sep 13 14:42
I think it's actually a very interesting topic. I am going to have a closer look at my figures...
By:
Darlo Bantam
When: 11 Sep 13 16:16
It is an interesting topic. However, I've done the reverse of the OP hence my question or debate that you need to know the ratio of wins:losses. I've tried to calculate that ratio to try work out my wosrt case red out point. It doesn't always follow but it should allow me to consistently make back any losses, especially if I keep losses to that worst case scenario point.
By:
sens
When: 12 Sep 13 08:48
If you like statistics, you can use Mybetlog.com to track Betfair trading stats/progress ;)
By:
jptrader
When: 12 Sep 13 20:41
For a large number of mainly inplay football markets my average win size is about 88% of the average loss size, remarkably close to both OP and Investor.
By:
askari1
When: 13 Sep 13 02:38
Thin and Crispy, I trade 1) my view of the form, 2) my perceptions of how punters view trainers, 3) my perceptions of market evolution and 4) my reading of data I have for how horses / yards have been backed in the past. My losses are far too big. I may trade the favourite on a market view but otherwise don't touch anything in the top four of the market.

I agree w/ Darlo that you need to consider win/loss ratio for the question to be meaningful. I doubt, though, that a bettor can suppose that his given win:loss number ratio, given win:loss size and history of holding onto e.g. 20% of his average trading win can determine his strategy for playing markets. One inference might be that he should bet more, because he is on average a winner; with a positive expectation per bet he should increase his stake. But the prices won't always be there and individual markets will have their own specific characteristics that determine how they should be played.
By:
U.A.
When: 14 Sep 13 14:08
"I suspect that most traders cut winners too short and let their losers run" I've seen this comment before and I just don't like it, it makes no sense. I think it must come from people who bet who don't seem to like traders. It basically seems to say if you trade and you make a profit then you will make a bigger profit if you just stopped trading out and let everything run and so bet instead of trade.

It's also making a generic statement that whenever you trade out you are giving away value. Hmm so where was that value on the Liverpool v Man U game? Well i had a position at 17 minutes so that was ok but look I traded out on 34 minutes, ah there's the value on the other side of course, if only i'd seen that and not traded out. Crazy.

The statement "I suspect that most traders cut winners too short and let their losers run" is equivalent to saying that position takers stake too much on losses and not enough on wins so if they could just master their staking system then they could really become successful.

If you really want to know if you are trading out too early or not, just record all the potential profits if you had let all your positions run and compare with your actual results. I looked at this on pre-match trading for football as I was cutting a lot of positions short on my opinion on what I thought prices would do and I looked at how I compared with if I let the prices evolve and traded out right on kick off and I found that my profit would have pretty much the same.
By:
cpfc4me
When: 15 Sep 13 07:39
U.A. - You are making some false assumptions. There is no suggestion that anyone should stop trading and become a punter, nor do I imply that every time a trader exits a position he is giving away value. As I mentioned earlier in the thread, (third post), football is not a trading sport in the sense that (for example) basketball, cricket, and NFL are, so your sarcasm on that sport is wasted.

Novice traders cut winners short and let losers run too long. It's the single biggest error they make. The premise of my opening post is that not many of us become expert enough to totally erase this tendency. It does not say anything about staking size or strike rate.

Your suggestion to review how all positions would have turned out had they run is presumably (my turn to presume) another attempt at humour. When you trade in-play, and are in and out of a market maybe 40 times in a game, this is not only nigh on impossible, but also ultimately pointless. A trader takes a short-term position and has a quite different mindset, and probably stake size, to someone planning on a punt.

The data that others have contributed does tend to confirm my view that the better traders will have average win sizes close to average loss sizes, and the best will have high strike rates (60%+) Given that a novice trader will have a low win size:loss size ratio and a lower strike rate, it is perhaps a useful indicator to ourselves to see how we are improving.

The much respected Bayes said that prior to 2010 (trading tennis), his win:loss ratio was 0.961 (Strike Rate 62.6%) whereas now it is 0.998 and 62.4% respectively. 0.961 is a lot higher than most of us and 0.998 is excellent. A couple of others, Investor2 and jptrader both reported a ratio of 0.88 which may be more typical, with a strike rate of 75% for Investor2.
.
http://green-all-over.blogspot.co.uk/2013/09/ratios-and-risk.html

Perhaps U.A. has some numbers to share?
By:
U.A.
When: 15 Sep 13 13:50
Hello there cpfc4me how's it going.

You say that football is not a trading sport and my sarcasm is wasted. Perhaps in this instance then maybe Investor2 and jptrader's figures are wasted and you shouldn't be referring to them.

My viewpoint is that if you place a bet and then a counter bet when a game is in play, then provided some kind of action has occurred on the field/court in between those 2 bets then you haven't really traded at all you've simply made a bet. Your bet being that the match/event will evolve in a particular way, and if it does then you win and if it doesn't you lose (in a spread betting kind of way). True trading in my opinion is based on price changes pre-match, or in periods when no action is occurring in the event. Of course many people will disagree with me and that's fair enough.

What determines whether you are a successful in-play trader (I use the term in your context) is whether, overall, the bets (determined from your entry and exit point) you are placing are value. Novice in-play traders will enter and/or exit the market at the wrong time.

I still don't like the term, sorry if that offends you, I just find it a bit meaningless. Like you said you are in and out of a market umpteen times, so there is no way of measuring it. I prefer to measure my performance by profit i am making in relation to how much is in my account and how much time/effort I am putting in. Actually to be honest I just look at whether it is worth me doing it or not/and whether i have the inclination to do it.

Let me know what numbers you would you like me to share, and I will decide whether I would like to disclose such information (if I have this info). By the way i'm afraid all my work is football orientated.

Cheers
By:
Darlo Bantam
When: 15 Sep 13 14:34
When you trade in-play, and are in and out of a market maybe 40 times in a game, this is not only nigh on impossible, but also ultimately pointless.

I don't know if I've misunderstood this, but in tennis and ODI/T20 cricket, I can regularly go in and out of a market 100+ times. And it's not pointless.
By:
cpfc4me
When: 15 Sep 13 18:08
Darlo - I meant that the exercise of determining how your initial entries into markets would have fared as punts is pointless. Obviously trading itself is not pointless, and I picked 40 as a number at random because that is a rough average for myself in an NBA or NFL game.
By:
Darlo Bantam
When: 15 Sep 13 18:21

Sep 15, 2013 -- 6:08PM, cpfc4me wrote:


Darlo - I meant that the exercise of determining how your initial entries into markets would have fared as punts is pointless. Obviously trading itself is not pointless, and I picked 40 as a number at random because that is a rough average for myself in an NBA or NFL game.


I wondered if you meant something like that.

By:
Mr Magoo
When: 15 Sep 13 19:51
I keep track of something quite close to win size : lose size ratio - my average winning/losing matched bet liability. It can be a good measure of whether your bets are being snapped up or if you are betting too big for the market.

If your winning matched stakes (or liabilities, for lays) are much smaller than your losing stakes, it shows that people are reliably grabbing your money when you're wrong but leaving some of it on the table when you're right. Not a good sign - could be caused by:

- Faster bettors, grabbing my bets when they react to something quicker than me
- My bet odds were too greedy, asking for worse odds might get the winning bets matched better
- Stakes too big for the market. Normal trading wouldn't fill the bet but only when my bet is 'value' to others does it get fully taken (e.g. fast picture people, or nasty inside-knowledge horse racing people, perhaps?)

Figuring out exactly what's going wrong when your losing bets are bigger than the winning bets is tricky, but spotting that it is happening in the first place is a good thing. (n.b. you could still be in profit with smaller winning bets, but it's a hint that you could make more profit by adjusting your strategy)
By:
TheInvestor2
When: 15 Sep 13 22:35
Yes Mr Magoo, I find that very useful as well.

In 2007, I won on 97% of matches and my avg win was 20% of my avg loss, so I've certainly moved toward a more equal ratio.
By:
TheInvestor2
When: 15 Sep 13 22:37
And here is another stat:
Year    % Win Days
2007    95.83%
2008    87.99%
2009    85.16%
2010    79.10%
2011    74.38%
2012    72.20%
By:
cpfc4me
When: 15 Sep 13 23:50
U.A. Thanks for your contribution to this discussion. You are correct that I am guilty of double standards by dismissing football as a true trading event, by including numbers from that sport, but I am short of data!

You wrote "My viewpoint is that if you place a bet and then a counter bet when a game is in play, then provided some kind of action has occurred on the field/court in between those 2 bets then you haven't really traded at all you've simply made a bet." "Some kind of action" will always have taken place, even if it is only the clock ticking. The odds on 0:0 will trend steadily down to 1.01 without any goals or significant events taking place. In a sport like basketball, NOT scoring on a possession is almost as significant an event as scoring - sometimes more so! To trade, you need to first place a bet bet. It's only when you hedge your position, locking in a profit or a loss or a wash, that your bet becomes a trade.

I prefer to measure my performance by profit i am making in relation to how much is in my account and how much time/effort I am putting in. Actually to be honest I just look at whether it is worth me doing it or not/and whether i have the inclination to do it. My profits are not related to the balance in my account. as I have mentioned many times on my blog, I have only ever made one deposit, and have no intention of ever making a second, so I tend to keep too much relative to my stake size, and only withdraw when I have a good reason to do so (usually Mrs Cassini. asking for new windows, wood floors or a holiday). I guess if you are someone who starts each day with a set amount in the account and tops up or withdraws each day, then your profit / loss relative to bank size means something, but for most of us it's probably not a useful measure. We all keep track of our profits, or should, but profit size doesn't tell you how good a trader you are.

The factor of time you mention is important, and often overlooked. Trading for four hours and making £100 is (depending on opportunity cost) arguably not as good as making £25 in ten minutes. The problem I have is that I never know which ten minutes will present the best opportunities!

As for numbers, how about sharing your average win size and your average loss size per market. The closer to 1:1 you are, the better a trader I suggest you are.
By:
JLivermore
When: 16 Sep 13 09:33
"The closer to 1:1 you are, the better a trader I suggest you are."
Why 1:1?  Why isn't it the higher the better? 

I'm at 1.13 over 3k matches.  My trading involves placing a bet, then hedging right before the final whistle at 1.01 or 1000.
By:
JLivermore
When: 16 Sep 13 09:39
"cutting my winners a little short or letting my losers run a little too long"
Nowhere does this statement imply that winners and losers should be of similar size (which is presumably where the 1:1 ratio comes from).
By:
henok
When: 16 Sep 13 11:37

Sep 15, 2013 -- 10:37PM, TheInvestor2 wrote:


And here is another stat:Year    % Win Days2007    95.83%2008    87.99%2009    85.16%2010    79.10%2011    74.38%2012    72.20%


was that due to market wising up, or you taking more risk? i know it will be a combination of both, but was wondering if you can identify the more significant of the two?

By:
U.A.
When: 16 Sep 13 13:52
Hello there cpfc4me.

The reason why I said "some kind of action" is that there are periods when there aren't. What I really meant is things like half time, quarter time or after any good play in American football when there are 5 minutes of whooping and high fives.

Sorry to be argumentative but your profits are related to your balance in your account. In absolute terms I feel strongly confident that I could make more profit the more money I had in my account. It's easier to make £100 in a week when your account has £20,000 in it as opposed to £50. Likewise the lower your account the easier it is to increase it by higher percentages. This is the context that I was referring to, i'm not quite sure how you are interpreting it.

As for my figures (profit in relation to my loss).
Jan-Dec 2007 - .9649
Jan-Dec 2008 - .634048
Jan-Jun 2009 - 0.53201
2010-2011 season - 0.46173
2011-2012 season - .050671
Overall percentage .57604

These are based on pre-match football trading. I think the reason why there is no 2009-10 season is that I was live trading more heavily and wasn't able to split the figures between pre-match and Live trading. I had mastered this by 2010-11 season. Don't have any figures for in play. This may be regarded as rubbish by many but if I had left my bets and traded out at kick off then my overall profit would still be the same. I think the reason why these percentages are low are that I am able to trade out of a few games with a minute profit rather than no profit-loss when the prices don't really move so this skews the figures. Fortunately my goal is profit rather than opinion/approval and I have made a profit for every season so far so I will continue until this changes.

Anyway this got me thinking in the difference between my trading (pre-match) and yours. With mine i'd mainly be in and out of the market 1, 2 maybe 3 times whereas you would be more like 10, 20, 30. But I assume that you only look at the overall profit for the match rather than each individual trade, as you mentioned before how can you look at each one, it's just way too much effort. So with your trading it could be entirely possible that your one profit or loss for the game could be made up of a whole load of tiny profits and a few big losses. E.g. you could have a profit of £20 for the game but that could have come about from 30 £2 wins and 2 £20 losses. I'm sensing this 1:1 ratio is more important the larger the number of trades you average per game.

Anyway as I have divulged some information here is a quick question for yourself and anyone else who fancies answering it. When you place your bets (with a view to a counterbet) what rational do you use. Do you have statistics which will indicate that the prices may be a bit out, is it your intuition of whether you think a price is good or bad, or is it your feeling on which way the game is going to progress (momentum). Or is it something completely different.
By:
KOBEFROMDOWNTOWN
When: 17 Sep 13 03:19
Having a feeling or following what looks like momentum will get you in trouble long term in my opinion.  Momentum is overrated, I have lost count on how many times a tennis player has won a set 6-0 or 6-1 to level a match and then comes out and is broken to love in the very first game in a 3rd set.
By:
pulio
When: 18 Sep 13 01:18
^ new set, new start.
By:
cpfc4me
When: 18 Sep 13 06:36
U.A.: Do you have statistics which will indicate that the prices may be a bit out, is it your intuition of whether you think a price is good or bad, or is it your feeling on which way the game is going to progress (momentum). Or is it something completely different. If you watch and trade enough 'genuine' trading games, you instinctively know when a price is wrong. As I did in December 2008, I again recommend Blink, by Malcolm Gladwell.
.
http://green-all-over.blogspot.co.uk/2008/12/blink.html

IF you want a sport with statistics, there is none better than baseball where the game has a long history, and the probability of winning a game from every game state is known with some accuracy. The Book - Playing The Percentages In Baseball" by Tom Tango, Mitchel Lightman and Andrew Dolphin.

Regarding momentum, it is something that is key to some sports, especially (in my opinion) basketball. If you can know, or at least know more than guess, when the momentum will change, that is a huge edge. There are certain events in a game that I have written about on my blog many times which are momentum changers.

The lack of momentum in football is probably the big reason I don't consider it a true trading sport.

Tennis is, but to Lakers fan KBEFROMDOWNTOWN I would suggest that psychology (a key component when trading a single player sport) plays a big part here. Anyone who has ever played tennis knows that it is hard not to mentally concede a set (especially an early one) if you are two breaks down, and conversely it is hard to maintain focus after winning a set easily. I would be surprised if this was not already factored in to the price at the end of the set. When someone writes "I have lost count of how many times..." I presume they long ago opposed their initial thoughts and are making money from the trend they have observed. I sometimes wonder.

Insanity: Doing the same thing over and over again and expecting different results. - Albert Einstein
By:
KOBEFROMDOWNTOWN
When: 18 Sep 13 10:25
cpfc, knowing when the momentum is going to change would be the equivalent of finding a pot of gold and a huge edge.  You can never know when a momentum shift is going to happen and it is total guesswork and fluke if you happen to nail it on a few occasions.

I agree that single player sport is completely different to team sport in terms of psychology,some players will conceded a set and put more effort in a deciding set but the player who has just cruised through the set is the one who mentally drops off in the decider.

I don't think it matters if it is a team sport or single player sport when you are talking about momentum, you can never predict it happening otherwise this game would be so easy.  American sports are so unpredictable, you can have a basketball team go on a 20-2 run but you can't predict when this will occur.  You could have a team in the red zone that has driven 90 yards, only takes a pick 6 and that momentum you have predicted has been flipped back in the other direction in a blink. 

Each to their own, just my opinion.
Page 1 of 2  •  Previous 1 | 2 | Next
sort by:
Show
per page

Post your reply

Text Format: Table: Smilies:
Forum does not support HTML
Insert Photo
Cancel
‹ back to topics
www.betfair.com