Wednesday, 1 May 2013

Looking For A Silk Purse

Webbo asked on the last post:

What price did you get out at Cassini? Or did you let this one run? I remember you saying once you have no intention of losing your whole stake when you trade so why wouldn't you have 'redded up' at 1.01?
Unfortunately it is not clear to which event or bet his comment relates to, but when trading, it is true that I have no intention of losing my whole stake. Unfortunately, whether it is my intention or not, that unfortunate outcome does result from time to time. Sometimes the market moves against you, and you do not get an opportunity to trade out at anything close to value. In trading the NBA for example, one risk of getting involved late on in close games is that you will see huge swings. Great if you called it right, but unfortunately us mere mortals sometimes call it wrong, and we are left holding a large red. For less liquid sports such as baseball, the problem is often that there are no takers at the true price, but you are better served in the long run to hold a position than you are to trade out at an unfavourable price - assuming of course that you are staking sensibly, which is easier said than done when a great opportunity presents itself. Webbo says “why wouldn’t you have ‘redded up’ at 1.01?’ In my experience, 1.01 is not usually a value back. It can be, but more often it is the case that the 1.01 is taken and the price then moves out. Years ago, I read a reported statistic from Betfair that only 70% of 1.01s actually won. I have no idea if that was true then, or is true now, but what I do see in the sports I trade is that the 1.01 gets taken too early.

Some of you may have woken up yesterday morning to find that one baseball game was still in-play. The Oakland Athletics v Los Angeles Angels of Anaheim game went to 19 innings and didn’t end until 9:42am BST. Money had traded at 1.01 on the Angels – not mine unfortunately - and they went on to lose.

I mentioned the Brooklyn Nets – Chicago Bulls game at the weekend in this post. Here, not only 1.01 traded, but also 1000 traded, and the ‘sure-thing’ lost.

Even when a 1.01 wins, more often than not, a little fear creeps in to the market, and you can back at 1.02 later on. I caught the end of the Golden State Warriors game on Sunday night / Monday morning, and the Warriors had a comfortable lead in the fourth quarter, up by 16 / 18 points or so. I layed five figures at 1.01, not because I expected the Warriors to lose (although that would have been a nice bonus financially) but because I expected Denver to close the gap at some point, which would result in a few rattled nerves. Greed and fear drive the markets. Sure enough, I was able to back some back at 1.03, some at 1.02 and make a little low-risk money when the lead dropped to 13 or 14. All it takes is a turnover, an ‘and one’ or a three-pointer to shake someone’s confidence.

The second part of Webbo’s comment went like this:
BTW, I'm trying to test an edge I may have over thousands of bets on the football. It's puzzling me that using a Kelly staking system this is profitable but to level stakes the bets make a slight loss just short of break even.
I wanted to know what you thought of this as I remember you and someone (Al maybe?) debating whether a system can make a loss at level stakes but a profit using another staking method. I previously thought you were right but now I'm not so sure.
Not sure about me being right?!! Oh ye of little faith.

A system can certainly make a loss at level stakes but be profitable using another staking method - in the short-term. In the long-term, no. For an example, look at Roulette. The true odds on black are 2.0556, whereas you will get paid at 2.0. Can you employ Martingale to overcome this negative expectancy? For a while, most likely you will, but sooner than you might think, probability catches up with you and you hit the house limit, run out of money, or you lose your nerve. Labouchere might seem a gentler way of playing this bet, but again, sooner than you might think, probability catches up with you, and you have a losing run and the stake size becomes too high. There's a reason there are no professional roulette players.

Unlike in Roulette, the edge on football bets will vary from event to event. It is quite possible that the edge on the smaller Kelly stake selections is negative. Perhaps Webbo would tell us what the level stakes P&L is when the lowest Kelly bets are expunged from the data set. A few higher value bets with larger stakes would compensate for the lower edge bets.

And now the moment you have all been waiting for - the Friendly Tipster League updates from the weekend. The XX Draws (Extended) had three draws, all 0-0, from 10 selections, three one goal games, and a 90th minute goal to take a draw away, so overall not bad. The Monday night game between Aston Villa and Sunderland was the final selection of the weekend, and not the best suggestion the spreadsheet has ever had.

Just two Classic selections, with the 1-1 draw between Real Vallodolid and Sevilla ensuring a profit on the draw bets. No joy in the Bundesliga though, where an 87th minute goal prevented one draw from two selections.

The top six are all from the XX family, with the Extended HT 0-0 moving into a Champions League spot:
while the next six, and the remaining profitable, entries are:
Premier Betting moved up form 13th, while Drawmaster came up empty and drop down from 6th.

In the red by less than 10 points are these nine:
The Bundeslayga selections are having a bad run with just two winners from 12, and the big gainer was the Extended U1.5 selections which has mentioned above had six winners from 10 selections. Neil's poor run continues too, with just one profitable weekend since mid-March. At the end of January, Neil was in 6th spot, and up 9.63 points, just 10 points off the lead.

The Hall of Shame candidates are this sorry lot:
Football Elite's nightmare season continues with just 4 winners in the last 25 selections, and my own XX Draws (Classic) remain in the relegation zone. Credit to Football Formbook though for keeping the selections coming on his blog. Plenty of others gave up long ago.

And hello May. Until this morning, April was something of a nightmare month, and the only one in the red in total.
April Daily Averages (Net)
After seven years of roller-coaster results, I finally realised where I was going wrong and after a decent, if not spectacular April this year, every month is now green, although April remains very much an outlier.

1 comment:

Webbo said...

Thanks for your comments Cassini. There perhaps not being an edge in the smaller bets makes sense to me and I'm going to go over the data and see if this is the case.