Meugh. Cheng Gong was held by Enmeshed. Never mind, still a good day.
I got lucky with Be Proud, when just before the off I thought a part bet hadn't been taken, put on a bit more, which it turned out I didn't need to, but it won. There is a deity and he's the punter's friend. 🙂
12.30 Navan. Straight dutching the top 4 in the market. Top Moon / Alpha Mix / Drumacoo / Mengli Khan. 52% of book. Stake £23.66 to win £22.
1.30 Navan. 2 way dutch on La Tektor and An Fraoch Mor. 54% of book. Stake £26 to win £22.
Good selections Brian,keep them coming 👍
2.35 Navan. 3 way straight Dutch on Minella till dawn / Glorious Galway / Wonderoftheworld. 58% of book. Stake £30 to win £22.
Link to today's outcomes... https://www.dropbox.com/s/t3wc0fu9qh2fkit/bethistory%20%287%29.xls?dl=0
I thought it would help me if I did an “end of week review” of how this has gone and the lessons learned.
I had 19 bets this week: 12 wins; 6 losses; 1 break even. Ignoring the break even, that is a 66% strike rate, which is consistent with what I have had previously. Profit on the week is £109.13. (Profit since I started on 15th November is £239.80).
I am going to ditch the backing at the lay price half an hour before the race as it has caused me problems on a couple of occasions (that’s why I didn’t make the normal profit on a couple of races today). Keep it simple is a good motto. So, in future, I’m returning to 10 minutes before the off to place the bets.
I need a staking system. I’m thinking along the lines of 4% of the bank as the “aim to win”. I began this with a £500 bank to a £20 “aim to win”. As things went along I have slowly increased it to £21 and now £22, but I think a 4% approach would be more efficient.
Comment and suggestions most welcome.
Attached is the first draft of a dutch/saver recording sheet drawn up before todays results and Brian's previous post.
Column A is hidden and is just a unique reference to assist with columns M & N
Column H is hidden and contains the market chance (1/price) for use in columns M (%book), O (Dutch expected winners) & Q (saver expected winners)
Data should be entered in columns B to L direct from Betfair. I have assumed the figures in L are the loss on the event or the gross profit on the event less 5% commission.
The other columns auto-calculate to produce a breakdown of winners and expected winners by 'saver/dutch' and the % book of each event. Everything is summarised in rows 1 to 5.
Here are my observations,
1. The saver bets are few but clearly identify horses with much less chance and edge than the dutched bets. The savers' expected 1.45 winners and delivered 1 (A/E = 0.71)
2. Had you backed just the dutched bets to return 1 point you would have had an A/E of 1.24 and 23% POT
3. Had you backed both dutched bets AND savers to return 1 point you would have had an A/E = 1.16 and a POT after 5% commission of 15.4%
4. Backing the dutched bets to return the same amount (more or less) AND the savers to breakeven (more or less) has cost you nothing. You have an A/E of 1.16 but because the selections with the edge are staked higher (as the Kelly principle suggests) you have a POT of 24.4%
Hedging your bets increases strike rate and reduces losing runs. Splitting the 'hedge' and backing more where your edge is greater and less where it is lower enhances the overall return so long as the Kelly percentage of bank is not exceeded.
Let's have none of this, 'I never bet more than one selection in a race' nonsense and get smarter about assessing the horses' chances, identifying those with the greater edge and stake accordingly.
How should the Kelly criterion be applied to hedged bets (multiple selections in the same race)?
Many thanks for this Andrew - awesome stuff!
I've got a copy of the book you recommended coming next week.
Sorry Andrew, I have a lot of respect for you but I'll stick to my one horse a race policy, I just don't get the same enjoyment watching a race where I have one or more selections in it.
I may well be wrong but does it make any difference to the overall return if I back on average 1 horse in a race at £60 or 3 horses at £20.
Also, if I have found a horse in a race that has an edge then logically the other horses in that race are less likely to have an edge.
I'll stick to my one horse a race policy, I just don't get the same enjoyment watching a race where I have more than one selection in it.
That's perfectly fine Michael, betting is a very individual thing and it's whatever works for the individual and they're comfortable with. I just threw that comment in as a little 'pot boiler' to stimulate thought/discussion, it wasn't aimed at you specifically. :0)
It took me over two decades to understand dutching/hedging, bank management and Kelly type staking and that it worked for me. It may work for others if they are are aware of it.
does it make any difference to the overall return if I back on average 1 horse in a race at £60 or 3 horses at £20.
Yes and no. Yes, your strike rate goes up. Your POT goes down but as your turnover increases, your money works harder and the net result is more profit. No, because you wouldn't stake them equally unless they had the same edge and were the same odds. You would bet a proportion of bank according to odds and edge.
if I have found a horse in a race that has an edge then logically the other horses in that race are less likely to have an edge.
Yes, but it's not always easy to identify precisely where the edge lies. An individual selection cannot be proven to be value (have an edge). This can only be true of a subset of horses where their A/E is consistently greater than 1 over a period of time.
Your extra place system finds a subset of horses that has an edge at bookmaker place odds but no edge at win odds. It's one horse but two bets, one at value one not. It's the net position that counts. I would argue several selections with an overall net edge/value is no different. Indeed, if you can't identify and split out a non-value set, it is the only option. At least with your extra place system you can identify a non-value subset (the win bets) and have the option of laying them to reduce their losses.
Thanks Andrew. I agree it's difficult to calculate the edge, you can calculate it on past results but things change in racing and I've found to my cost that past results are not always an indicator of future results.
What I like about my each way systems is that there is a logic to them that is not based on past results, it is based on bookmakers odds and nothing else.
I'm not convinced that laying the win bets will leave me better off, the ROI is -3.8% so I assume that if I laid them I would be making bigger losses due to the Betfair commission.
I'm not convinced that laying the win bets will leave me better off, the ROI is -3.8% so I assume that if I laid them I would be making bigger losses due to the Betfair commission
You'd have to run an exercise across all win bets to get a true picture Michael but Betfair commission is now down to 2%.
With your last loser (Lust For Glory), you had £25 on at 6 (7 decimal) so £175 lay liability on Betfair at whatever price you could get. At BSP of 7.5 you would have won 175/7.5 = £24.31, less 2% = £23.81. A net loss on the win side of of £1.19 rather than £25.
Interesting discussion on betting styles chaps. I'm definitely on the Michael Wilding / aversion to long losing runs end of the spectrum, hence my interest in dutching. I found it psychologically too hard to handle long sequences of losers, so I have been looking to develop "high strike rate" based approaches which mitigate this.
Couldn't find anything of interest dutching wise in today's racing, just a couple of potential 80:20's on the favourites in the 5:10 and 5:40 at Wolverhampton.
2 way straight dutch: Perfect Moment and Todd in the 1.10 at Fakenham.
2 way straight dutch Mah mate Bob and Pogue. Catterick 1.25.