Critical analysis

[ShaunWhite]

What information do people find useful when doing critical analysis?

Video seems the easiest and contains a lot of info, ladders, graphs, tv stream etc. That seems good for individual situtions or types of situation, but I wondered if there was any other data/trade capture that would be useful to tell the story over a longer timeframe. What are the essentials?

This question probably stems from an anxiety about not collecting enough information, and when I do set something up, I soon find it doesn't really help. The months keep rolling by and adding to the other months that I really should have data for.

[rinconpaul]

I'm presuming you're not a handicapper as that will involve sooo much work, which data services provide anyway? Pick a category or categories, but one's you will honestly have enough time to devote, update and be interested in? it might be UK greys, AUS gallops, English Soccer etc.

Then set up a bot to record your data each day. Your parameters are everything that Betfair offers (price, vol, field size, rank, gap % etc) For Greys as an example, it might be from 3 mins to advertised start to actual start at 5 second intervals? Soccer's much different? The work comes in updating results and cleaning data the next day. I devote at least 2 hrs a day to that and no time to betting, as that's all automated. There is a big difference to what you can get from Betfair's files (not worth a pinch) to what you collect yourself. You need to be committed long term and put in the hard yards. After 6 months you will have a treasure trove of data to start analysing. The time will pass quickly and you will be rewarded. You might think that you will be competing against thousands doing the same thing? Not so, very few are that committed. You will have an edge! it's all about finding parameter combinations that have an actual probability consistently outside implied probability.

Then you need a regime to properly statistically analyse, but that's for another day.

[Derek27]

I've never bothered with statistical analysis of historic data.

I honestly don't know if I'm saving time and making more money investing the time on trading, or whether I'm missing a major opportunity to improve my trading.

[rinconpaul]

Curious then Derek? If you've never done any statistical analysis and you don't do form, how do you know what's successful and what's not, so that you can continue to repeat a successful technique? Do you just guess at a direction and scratch if a wrong call? I don't trade myself, only straight back or lays. I would've thought there was a bit of science behind successful trading, but could be wrong? Do you reference any technical indicator like MACD or MA crossovers?

[Derek27]

Curious then Derek? If you've never done any statistical analysis and you don't do form, how do you know what's successful and what's not, so that you can continue to repeat a successful technique? Do you just guess at a direction and scratch if a wrong call? I don't trade myself, only straight back or lays. I would've thought there was a bit of science behind successful trading, but could be wrong? Do you reference any technical indicator like MACD or MA crossovers?

I used to consider myself a horse racing expert and made money betting. When I first opened an account with Betfair for betting purposes, I quickly realised the advantage of cashing out a back bet when the price has got ridiculously short, and laying over-backed favourites. I gradually shifted to trading and I'm now a cold trader - I don't closely follow horse racing. The main reason for that was the sheer amount of time it takes to keep on top of form, compared to the flexibility of trading.

Trading is a bit like chess in that there is no technique that will always win. Every game of chess, and every market, is different. The most crucial factors, in my opinion, are the prices, direction of prices, volatility of price movement, and overrounds. It's possible to profit by just using your judgement to estimate where a horse is getting too short or too big, and which way the price is more likely to go.

[ShaunWhite]

Thanks riconpaul... noted.

In the OP I wasn't actually thinking in terms of data mining, the critical analysis I'm thinking of is being able to look back at your trades and have all the information necessary to assess the effectiveness of the plans you have in place for each type of setup. It's probably only of limited use in our markets although it's fairly standard practice in the city.

[ShaunWhite]

I've never bothered with statistical analysis of historic data.

I honestly don't know if I'm saving time and making more money investing the time on trading, or whether I'm missing a major opportunity to improve my trading.

Once saw an interview with a very successful trader who said that like an army sniper, pulling the trigger is only a very small part of his job.

Sports markets are admittedly much more involving than financials but even so I think I only spend about a quarter or third of my day looking through the gun sight, and actually managing open positions is probably just a small fraction of that. The extreme case would be the botting specialists who arguably don't spend any time trading at all.

It sounds like you're more of a machine gunner. Horses for courses though init.

[Derek27]

I wasn't thinking in terms of data mining, the critical analysis I'm thinking of is being able to look back at your trades and have all the information necessary to assess the effectiveness of the plans you have in place for each type of setup. It's probably only of limited use in our markets although it's fairly standard practice in the city.

I'm probably different to most or many traders in that I only have one setup, or one approach to trading. That is, back a horse if you have reason to think it's more likely that the price will go down, taking into consideration the extent it's likely to go down and the extent it can go against you, or lay a horse for opposite reasons.

So this isn't very relevant to your question but whenever I look back on my losing trades, like today when I overstaked and took a 12-tick loss, I can conclude without critical analysis that at that moment in time, I was a complete dickhead!

The obvious solution is to pay more attention to what I'm doing and I'll get back to winning ways.

[Derek27]

I've never bothered with statistical analysis of historic data.

I honestly don't know if I'm saving time and making more money investing the time on trading, or whether I'm missing a major opportunity to improve my trading.

Once saw an interview with a very successful trader who said that like an army sniper, pulling the trigger is only a very small part of his job.

Sports markets are admittedly much more involving than financials but even so I think I only spend about a quarter or third of my day looking through the gun sight, and actually managing open positions is probably just a small fraction of that. The extreme case would be the botting specialists who arguably don't spend any time trading at all.

It sounds like you're more of a machine gunner. Horses for courses though init.

On that analogy I would say I'm very much a machine gunner. I don't drink as much now but I used to trade US racing, to small stakes, when I was completely pissed. I would sometimes lay a horse at 2.2 and then back it at 2.6 shortly afterwards. When a market is that volatile and people are backing and laying such a wide spread, it doesn't take much thinking (or sobriety) to make a few quid.

You can apply the same approach to UK racing, albeit with smaller margins. But if I used a bot to record data of all my trades and how the market moved thereafter I think it would be very useful information.

[rinconpaul]

Derek's not winning me over with a technique to follow really ...but being a stats person, I thought I might be able to use data analysis to determine market direction? I took a sample of a certain rank in a certain field size.

I plotted the open prices across the X axis and count of steamers and drifters on the Y axis. See 'Frequency of Open Price' chart. There were more steamers than drifters in the lower 3rd of the trading range of prices. Then it becomes a mixed bag of volatility. Then I added a graph of Lay profit, just laying a price range (middle chart) and you can see that all the profit for Laying is in the bottom 3rd of trading range. Similarly added a graph of Back profit (bottom graph). The middle (average) of price range, there's a turning point between profit and loss and almost all of the top third of the trading range Backing makes money.

Fundamentally there's nothing new here! Prices below the average are Lay propositions and above Back propositions (straight bets only!) To get back on track trading wise, using statistical analysis of past data, I could program a bot that if a price opens in the bottom third of a particular known trading range, that it will Steam more often than not, so Back it first. As for any other price range : don't trade, unless you're looking for volatility (danger Will Robinson)!

Derek you say, "if you have reason to think it's more likely that the price will go down"....but what is that reason and how did you come by it, if you haven't used stats of past performance?

[Derek27]

My aim is to make money trading rather than to win people over with a technique to follow, and it wouldn't be possible anyway, because a technique isn't easily transferable - you can't write it precisely on paper.

Derek you say, "if you have reason to think it's more likely that the price will go down"....but what is that reason and how did you come by it, if you haven't used stats of past performance?

There's a number of complexed reasons why you may think a horse is going to move in one direction, such as money coming in for it, money for other horses, overly biased overrounds, etc. Surely what's happened in the last five minutes is more relevant than any statistics?

[rinconpaul]

"Surely what's happened in the last five minutes is more relevant than any statistics?"
Yep, you're right, my steamer/drifter parameter is determined by price action data based on the last 5 minutes action at 5 second intervals with a gap back-lay > 90%, so on the same page.