Friday 10 February 2017

Basic Models, Priorities

In one of Steve M's recent posts, he writes:

Taking a leaf out of Cassini’s blog to generate as much content out of one post as possible, I’ll be replying to a very well written comment on my last affiliate post. Instead of in the comments (where no one really reads them), I think it will be better to make it a whole new post.
Good to see that at least one of my 'best practices' has found some support, and it's one that makes sense, if I do say so myself. 

I try to comment on all published comments, because not only does it give me a topic for a new post, but it also gives the comment some oxygen - as Steve says, comments typically tend to get overlooked.

It's rare that I'll publish a comment and not respond, at least in passing, to it in a subsequent post, because if the commenter has taken the time to post their thoughts in a polite way, I feel they deserve a response.

Unfortunately not all comments are polite, well-written or even timely, and so never make it to the blog. One recent comment was asking about the XX Draws service that was retired after the 2014-15 season! Keep up people.  

Sadly, I'm not sure a couple of the 'best practices' there - namely conditional bets and money-back guarantees, have made it into the mainstream tipster world yet, and I doubt that they ever will.

Betting Tools' Brian wrote in response to my post about his post about my posts (you get the picture) on court-siding:

Appreciate the kind words, thanks Cassini. I'd be interested to hear why you think the US sports that you used to trade so well are more difficult for you now?
I would think that the introduction of the cross-matcher (which certainly seems to prevent selections trading as low as they used to inplay), as well as the premium charge and lower liquidity levels, have been the biggest factors but I would think that these sports could have 'courtsiders' as well?
Apologies if you've touched on this before and I do remember you saying you have less time for trading in general.
The in-play trading environment in general has, as expected, become more difficult in recent years for the home-trader. The Premium Charges are certainly a big reason, but as the exchange concept has matured, the expertise of the competition has improved greatly. Court-siders are certainly one weapon, but behind the scenes, models in all sports are far more sophisticated now than they were ten or more years ago. Markets become more efficient and sports trading is no exception.

Opportunities for the part-time / home-trader are now few and far between and, in my opinion, effectively non-existent in sports such as tennis where court-sider data and efficient pricing models exist in large numbers. 

Stephen High, who should know, stated in 2015 that some 75 court-siders were present at a Wimbledon Final. If you seriously think that you are going to have any kind of an edge against that level of competition (and remember, these 75 court-siders and their back-offices are in it to make a profit), frankly you're delusional. 

Add in that you need an edge large enough to cover the commission cost, and the time (and thus money) that you would need to spend on this venture, and it should be obvious that it's a non-starter. 

No two of us are at the same stage in life or have the same goals, so it's a personal decision how much time to spend betting or trading. As Brian says, I do spend much less time trading now. My real job takes priority and while I'm not the CEO of a multi-billion dollar global enterprise, I do currently have responsibility for the careers of some 43 people right now, a responsibility which is more deserving of my time than trading a sports event, and the reason the Draw Service was retired. 

With a 50% Premium Charge and court-siders / more sophisticated in-play models now out there, the time required to find a rare opportunity can no longer be justified. Hence the move to a more cost-effective system / portfolio strategy these days. 

It's also fair to say that another factor is that I'm not getting any younger and have other priorities. I plan to travel again next month, as well as in June, August, November and probably next February. Paid holidays is another advantage of a 'proper' job.

JokerJoe, perhaps trying to convince himself rather than me, writes on this same topic:

I'm not sure I agree with you. Courtsiders are trading information flow and only need a basic model to profit. They need to know the direction price will move in and approximately where the market will settle after the news. They want to capture the bulk of the move and close out the trade immediately afterwards. They do not need to predict to the tick where the price should be afterwards as sweeping those last few ticks aren't worth the effort. As such a basic model calibrated against market prices is sufficient. Essentially the courtsiders are just reflecting the market's opinion, if the market knew all the information.
If they have a more sophisticated model then they may end up holding positions, so they aren't "pure" courtsiders and a portion of their funds is being used speculatively. That means their model is competing against others. One can view that as their courtsiding money sweeping to the basic model (market) price, and the speculative portion as being a trade at that price.
Either way, no matter where the market settles, it isn't necessarily at the fair price. Markets aren't perfectly efficient, I presume you'd agree otherwise you wouldn't be involved in them at all. I don't see why there shouldn't ever be moments when your model/view might not significantly disagree with the market price.
It's been previously covered, but these pricing models are extremely efficient. If the model is only 'basic' as JokerJoe suggests, they wouldn't be in business, and as new competitors enter the market, models need to be ever more efficient. If all they knew was 'approximately' where the market will settle after a point is decided, they would lose out to a model that knows 'exactly' where the market would settle, either by not getting any action or by getting action but with a negative expectancy. With the vast amount of data that some syndicates have accumulated, the 'true' price is known. 

Why would the price ever NOT settle at this fair / exact price? Back at greater than that price or lay at less than that price, and they make a profit. Get it wrong and they make a loss. 

If losses start to accumulate, a model will soon adjust (or go extinct), but there are at least tens of full-time models out there for tennis. I'm not convinced that a part-time or home-trader's model is realistically going to have an edge that will cover commission and be better than those other, well-financed models.

Essentially, any time you get matched in-play in an event where court-siders are active, you are taking the opposite side of someone who is ahead of you both in time and (almost certainly) in ability to price the market accurately. The 70 or so syndicates have already passed on that price, or are offering you that price. Are they all wrong?   

As for holding positions, why would they not? No one should ever trade out of a position for the sake of trading out, but only if there is value in doing this.

It's the single biggest trading error to close out a position as soon as a profit can be locked in. If you back at 2.0 and can lay at 1.9, it makes no sense to do so if the 'true' price is less than 1.9 and your staking is sensible. You are giving value away at that 1.9. Positions don't need to be held until expiry, but until there is value in closing out the position. 

JokerJoe suggests that "court-siders are jut reflecting the market's opinion". I'd say that the market IS the opinion of court-siders, and that in an efficient market like tennis, it's almost always correct. 
Moving on, and for those watching the Apple story, unfortunately for our protagonist, the price continues to soar. Shorting a company such as Apple is a dangerous game, and as I asked in my last post, before entering any trade, be it financial or sports betting related, asking the question "what do I know about this that others don't?" is always time well spent.

Finally, I mentioned last month that:
Using just two filters in one league, the system would have generated 417 bets for a profit of 174.25 points, an ROI of 41.8%. My P-value calculator gives a value of 0.000368, (or 1 in 2,716) for these results.

Too good to be true? With half the season left, I'm adding this to my portfolio and will provide updates here from time to time.
From time to time is here, and since that update, another 9 bets have added 0.32 points to the pot. Unfortunately this system isn't for sale, but to the person making the offer, it's appreciated, but we can only sell it when it starts to lose!

1 comment:

Trader 247 said...

It would be a waste to ignore what works for those with more experience. So I have also adopted the 'best practice' of responding to comments in posts. Thank you.