Handinaps

This article relates to the horse tipping competition organised on twitter by Handinaps.

I have run a few basic simulations to give an idea of distribution of scores for different player types (sharps/guessers) and different strategies (low risk/high risk).

Fav Sharp: This is a low risk strategy for a sharp punter that is adept at finding value on the shorter priced odds.  The parameters I have used here is that he has a 10% edge and he can find two horses about 5.0 decimal odds in each of the 40 races.  He tips both.

Fav Guesser: This again is low risk, the favourites are short enough prices that you would expect the final position to converge to close to the expected value over 40 races (80 picks).  This type of player however is not sharp, so his EV will average out at some negative number to account for the margin the bookie applies to favourites.  Again the parameters I use here is 5.0 decimal odds and giving the relatively uninformed punter a -15% edge.

Mid Range Sharp: As the Fav Sharp but fancies himself at picking midrange horses.  This strategy is more high risk - there is a broader range of outcomes (both good and bad).  Parameters I've used are 21.0 picks, and sticking with the 10% edge. 

Mid Range Guesser: This guesser is shooting to be lucky by going for a more high risk approach.  There is usually a bigger margin applied to bigger priced horses, so I penalise this approach by decreasing his edge.  The parameters here are 21.0 picks and -50% edge.  This is the equivalent of betting 21.0 odds when the true price is 31.5.  This looks feels like a really harsh judgement on this Guessers ability, however the midrange horses in the Lincoln last year ISP compared to BFSP (the 20/1 is 28.0 on exch) and research I have done before does suggest that this is the sort of margin built in at these mid range odds.

Longshot Sharp: This is the most high risk strategy for sharp players. Parameters I used here are 51.0 picks and still a 10% edge.  I have used the same edge for all the sharp strategies.  The bigger margin applied on the longshots implies these odds are harder to beat (you have to beat a big margin just to simply break even).  The bigger margin applied is precisely because the books are less sure of these prices, so there is also a bigger chance for the sharp to spot a rick.  So I have assumed that these can be balanced and I do not penalise the Sharps for betting the bigger priced horses as I have penalised the Guessers.

Longshot Guesser: The parameters used here are 51.0 picks and -75% edge ie a true value of 90.0 for the picks.  I think this is consistent from the previous links provided comparing ISP to BFSP.  

By being good the Sharp players have made 10% ROI.  On 40 races, 2 picks per race at 2 units that equates to 16 units.  By being lucky one of the sharps on the low risk strategy has made an additional 104 units, whereas two players on the high risk strategy have made an additional 548 units (over 1000 sims).

So its seems you have to be both smart and lucky, but most importantly create the opportunity to be lucky.

There are only 411 entries in the competition.  This is much less than 1000 sims that were run to generate the charts.  So I dont really envisage a scenario where one person has luckily picked eight 50/1 winners in forty races as has happened in a couple of the sims for Sharp player types.  My assumption too is that nobody will fall into the 6 defined categories precisely, but will be a combination of all 6.

I am also very skeptical that anyone can have a 10% edge on SP.  The list of entrants shows that many are not professional punters, many have jobs (traders, media, journalists mostly).  I am not saying this to diminish their tipping skills, just to point to the fact that gambling is probably not their main source of income.  Others on the list are professional gamblers, but football punters or poker players.  Finally, if when you do identify professional horse punters on the list - what good is that when you are betting at SP.  Perhaps their skill is accurately modelling each horse's probability, you still have no way of picking whether or not this is value against a price that you dont know.  If you are sharp you would expect the market to move to your price.  Perhaps their skill is picking good early prices, again no good in this competition you'll be logged against SP.  Perhaps their skill is spotting and trading market movements on betfair, no good.  Perhaps their skill is not being a judge at all but their skill is in getting on and have a very well organised team of runners.  This competition betting at SP is testing a very sepcific skill.  To me, it is as if you need to be both expert in handicapping and expert in reading the industry/markets.  To be able to accurately evaluate each horse and also to be able to predict what the Industry SP will be. And hence be able to identify some cases where for some reason the market does not move to your sharp number.  Don't get me wrong, all the above mentioned experiences may help.  However, in my opinion then there will be very very few players with a quantifiable edge of anything near 10%.

I don't mean to be insulting to anybody here, becuase I include myself in this analysis too.  The point is that this is going to be really hard (imo).  If you can do it, and then the previous paragraph doesn't apply to you.  Fair play.

With all that said, I think the Mid Range Guessers will vastly outnumber the sharp player types and dominate. The number of players on the Sharp strategies will be in the low single digits imo.  I think the most likely strategies will begin as Mid Range Guessers and perhaps morph into Longshot Guessers as some will launch a Hail Mary play to get back into it if luck runs against them.  According to sims 6 playeres in 1000 of the Mid Range Guessers hit 176 units profit or greater, this could translate to about 2 or 3 people out of the 411 entrants hitting similar heights.

I will probably take a hybrid approach to my picks and mix things about.  Aim to win 4 units per race.  When I fail, then similar to martingale I will try to win 8 units the next race and so on.  All going to plan 160 units (4 units x 40 races) should be right near the top of the final table.  Martingale is the very definition of a strategy that you have managed your distribution of outcomes in a way that favours this type of competition.  It doesn't matter that that one time you get in so deep that you cant bet your way out - you haven't lost more than your entry fee no matter how big you lose.  

How this strategy can (and probably will) derail me is for example after race 6, I am -24 units then I will need to win 24 units (to wipe out my debt) + 28 units (to put me back on track for 4 units per race by 7 races).  So I could quickly find myself struggling, desperately looking and hoping for 2 units on a good 33/1 offering from the books.  That will quickly descend to my no way out scenario, which I accept.  I am trading EV for Variance. 

I want a strategy that I either win alot or lose alot, rather than one that will just average out to my EV, because even if I am 10% sharp 16 units is not going to win it.

The article focuses on the strategy, but after this point the picks will be genuine.  Leverage any each way angle, market knowledge, form and handicapping skills I have or might like to improve in order to pick the most efficient bets to maintain those 4 units of profit per race.  Please don't read this as an article on trying to game the system.  

I write this article because I think it is a big problem with tipping/betting competitions.  Even massive tipping services suffer from this kind of survivorship bias where being lucky seems to be more important than being good.

I think the format proposed is very good.  The fav-longshot bias penalises players that want to just try and get lucky in this scenario.  Also, there are fair restraints on the stakes and the races that you can bet.  The cup aspect of the competition also discourages high risk shots.  I still think it can be improved upon, and am interested in general in how tipping/betting competitions can be designed so that you can strip away variance and the best bettors rise to the top.  But I have no answers.

Lets wrap this up with a quote from an Unabated article on the DraftKings National Sport Betting Championship (which had absolutely bonkers rules unlike Handinaps, but can still be a much needed warning).

"In this contest, strategy matters more than handicapping skill. Don’t be afraid of -EV wagers if those wagers will present an opportunity to catapult to the lead late in the tournament. The winner may be crowned the 2021 Sports Betting National Champion, but in the end they were really just the King of Variance one weekend in November."