BETTING NHL TEAMS ON A LOSING STREAK FACING A TEAM ON A WINNING STREAK

Today I hope to give you some insight about the following question:

Considering the public’s tendency to overreact to recent results, should we bet on teams which are currently on a losing streak facing a team riding a winning streak (in the NHL)?

First of all, you should note that all the results presented in this article come from a dataset containing information about all NHL games from the last 9 seasons (2007/2008 to 2015/2016).

Let’s kick off this study with a very simple analysis: we are betting 1$ on each team coming off at least one loss playing against a team coming off at least one win. Here are the results:

• Record = 2899-3098 (win percentage = 48.3%)
• Profit = -178.13$ (-3.0%)

To put these numbers in perspective, how much money would we have won/lost from using the opposite strategy (i.e. betting on teams coming off at least one win facing a team coming off at least one loss)? The answer is: -249.54$ (-4.2%). As you can see, the initial suggested strategy seems to be doing much better than doing the opposite. We are off to a good start.

From now on, I will use odds with “reduced juice”. In my dataset, when two teams were evenly matched the juice was 20 cents (for example -110 versus -110, or -135 versus +115). That is not representative of the reality as you can obtain much better odds simply by playing at casinos like Pinnacle or 5Dimes. Also, if you have an account with multiple bookies you can get even better odds by shopping for the best line. Consequently, I have adjusted odds in my dataset (I won’t go into details about the specifics here).

With reduced juice, the profit numbers above become -62.88$ (-1.0%) when betting the team on a losing streak versus -147.24$ (-2.5%) when betting the team on a winning streak. In other words, you are losing 2.5 times more when you go against the suggested strategy.

All right, let’s now refine our analysis to see if we can find a winning strategy. But first I’d like to introduce a little bit of notation to simply the writing for the remainder of the article:

• x = number of games from the losing streak
• y = number of games from the winning streak
• Teams L = teams coming off at least one loss
• Teams W = teams coming off at least one win

Let’s study the performance of the suggested strategy both as a function of “x” and “y”. Ideally, we’d like to see our winnings increase as these two variables go up. In plain English, are we making more money when betting a team on a big losing streak facing a team on a big winning streak?

For each couple (x, y) I am showing you three things:

• The number of bets;
• The profit when betting Teams L (suggested strategy);
• The profit when betting Teams W (opposite strategy).

Let’s see the details (I have highlighted in blue the strategy yielding the highest profit):

	WINNING STREAK
	1			2			3
LOSING STREAK	#G	Teams L	Teams W	#G	Teams L	Teams W	#G	Teams L	Teams W
1	1315	-32,46 $	-20,47 $	826	-35,71 $	2,57 $	411	-5,74 $	-12,85 $
2	846	-26,07 $	3,24 $	344	4,68 $	-7,15 $	179	-15,16 $	11,64 $
3	441	-10,94 $	-11,93 $	157	19,82 $	-27,64 $	81	-16,96 $	12,64 $
4	223	5,92 $	-20,66 $	87	1,05 $	-1,61 $	33	-4,12 $	3,48 $
5	92	-6,76 $	8,02 $	44	-0,87 $	-2,76 $	24	-6,28 $	4,19 $
6	49	0,04 $	-1,83 $	25	0,99 $	-2,85 $	10	0,97 $	-0,43 $
7+	52	10,34 $	-16,52 $	22	2,76 $	-2,28 $	8	-2,19 $	0,97 $

 

	WINNING STREAK
	4			5			6
LOSING STREAK	#G	Teams L	Teams W	#G	Teams L	Teams W	#G	Teams L	Teams W
1	205	14,42 $	-21,48 $	92	-12,71 $	14,81 $	59	14,83 $	-14,03 $
2	82	18,31 $	-15,71 $	46	17,46 $	-17,20 $	20	6,46 $	-6,72 $
3	42	10,42 $	-11,10 $	19	4,89 $	-2,51 $	6	-1,25 $	-0,23 $
4	18	-10,57 $	6,53 $	15	-3,36 $	2,02 $	10	2,22 $	-3,78 $
5	11	5,43 $	-2,67 $
6	4	2,59 $	-2,37 $
7+	4	-1,90 $	1,04 $

 

 

	WINNING STREAK
	7+
LOSING STREAK	#G	Teams L	Teams W
1	61	-12,44 $	4,02 $
2	18	1,31 $	0,64 $
3	9	0,48 $	0,31 $
4	7	-2,78 $	3,42 $
5
6
7+

 

Let’s start by focusing exclusively on the performance of the suggested strategy as a function of “x” (i.e. depending on the number of consecutive losses by the team we are betting on):

LOSING STREAK	Profit Teams L
1	-69,81 $
2	6,99 $
3	6,46 $
4	-11,64 $
5	-8,48 $
6	4,59 $
7+	9,01 $
TOTAL	-62,88 $

 

As you can see, most of the damage is done when betting teams on a 1-game losing streak. If we remove those cases and only bet when Team L is coming off at least 2 losses, we are earning +6.93$ (+0.2%). Such a return on investment is very small, but at least we are not losing anymore. We’re talking about a VERY SIMPLE strategy here: bet teams coming off at least 2 losses facing a team coming off at least one win. That’s it!

Let’s turn our attention to the performance of the suggested strategy as a function of “y” (i.e. depending on the number of consecutive wins by the team we are betting against):

WINNING STREAK	Profit Teams L
1	-59,93 $
2	-7,28 $
3	-49,48 $
4	38,70 $
5	6,28 $
6	22,26 $
7+	-13,43 $
TOTAL	-62,88 $

 

This is where it gets very interesting. I see a clear pattern where we are losing money for y = 1, 2, 3 but we start observing positive results for y >= 4 (except for the last case). So here is another very simple strategy that leads to even better results: bet Teams L coming off at least one loss facing Teams W coming off at least 4 wins. This strategy would have earned +53.81$ (+7.4%)!!

 

THE ROAD/HOME SPLIT

I have read a report claiming that betting all visiting teams over the last 10 years led to a ROI (Return On Investment) of -0.7%, as opposed to -3.1% when betting all home teams. I figured I would check if our strategy leads to better returns when looking at road versus home teams. The numbers are pretty convincing. Let’s start with the situation where the road team is Team W and the home team is Team L:

• Betting on all Teams L who were at home: profit = -50.90$ (-1.7%)
• Betting on all Teams W who were on the road: profit = -45.89$ (-1.5%)

As you can see, betting either team is a wash: you are losing as much money on both sides.

Let’s now see the case where Team L is on the road while Team W is at home:

• Betting on all Teams L who were on the road: profit = -11.96$ (-0.4%)
• Betting on all Teams W who were at home: profit = -100.61$ (-3.3%)

WOW!!! You are losing 8.5 times more money on Teams W as opposed to Teams L. Specifically, when you have a road team on a losing streak facing a home team on a winning streak, you should definitely avoid betting the home team! The public’s overreaction to recent results is therefore more pronounced when the team who is hot is playing in front of their fans.

As we did earlier, let’s have a look at the detailed results as a function of “x” and “y”:

	WINNING STREAK
	1			2			3
LOSING STREAK	#G	Teams L	Teams W	#G	Teams L	Teams W	#G	Teams L	Teams W
1	658	-18,86 $	-16,54 $	423	-32,14 $	12,44 $	195	-0,15 $	-8,17 $
2	442	-1,18 $	-8,02 $	165	-2,70 $	-1,68 $	92	-2,59 $	1,39 $
3	243	-2,38 $	-16,91 $	67	-1,66 $	-3,41 $	37	-7,30 $	2,96 $
4	94	-7,13 $	2,25 $	46	7,55 $	-6,48 $	14	-0,83 $	-1,10 $
5	47	3,43 $	-3,14 $	23	0,60 $	-2,08 $	10	-3,19 $	1,27 $
6	30	3,81 $	-4,83 $	16	-1,46 $	0,67 $	8	-0,35 $	1,57 $
7+	24	-3,59 $	-0,67 $	10	2,15 $	-1,73 $	4	-1,60 $	1,32 $

 

	WINNING STREAK
	4			5			6
LOSING STREAK	#G	Teams L	Teams W	#G	Teams L	Teams W	#G	Teams L	Teams W
1	109	12,12 $	-13,36 $	46	-7,77 $	6,69 $	29	14,50 $	-8,28 $
2	45	18,03 $	-14,34 $	25	14,84 $	-10,40 $	13	2,00 $	-1,77 $
3	21	11,40 $	-9,77 $	11	4,85 $	-2,25 $	5	-3,38 $	1,79 $
4	14	-9,32 $	5,47 $	10	-3,77 $	1,44 $
5	8	3,62 $	-1,22 $
6
7+

 

	WINNING STREAK
	7+
LOSING STREAK	#G	Teams L	Teams W
1	32	-6,09 $	1,57 $
2	6	3,86 $	-2,78 $
3	7	2,72 $	-2,51 $
4
5
6
7+

 

The potential winning strategies are exactly the same as the ones described earlier (no road/home split), but this time with more convincing ROIs. Recall how earlier we found that betting Teams L when the losing streak is 4 games or more led to a 53.81$ profit (+7.4%). If we restrict ourselves to such teams playing on the road only, the profit is 57.61$ with a ROI of +15.1%. Basically, the profit is the same but we are betting pretty much half the number of games we did before, which doubles the ROI.

 

THE ODDS SPLIT

Are we performing better for a certain set of odds? That question came to my mind after realizing we may be winning less (or even losing money) on teams where the odds were pretty high. Not because the system does not work, but simply because the vigorish (the casino’s commission) is bigger. Let’s see the results, still based on the strategy which consists of betting a road team on a losing streak (1 game or more) facing a home team coming off at least 4 consecutive wins:

 

Road odds		# Games	Profit
0	1,67	14	7,11 $
1,67	1,77	12	1,76 $
1,77	1,84	17	-2,49 $
1,84	1,92	15	0,11 $
1,92	1,97	12	-0,31 $
1,97	2,06	24	-1,76 $
2,06	2,16	26	-2,67 $
2,16	2,255	31	2,37 $
2,255	2,36	23	9,66 $
2,36	2,46	28	10,66 $
2,46	2,59	27	6,10 $
2,59	2,71	33	6,48 $
2,71	2,79	25	-5,65 $
2,79	2,89	18	2,00 $
2,89	2,99	19	-1,30 $
2,99	3,11	21	3,43 $
3,11	3,26	16	9,55 $
3,26	3,56	10	7,05 $
3,56+		10	5,50 $
TOTAL		381	57,60 $

 

The very first row of the above table corresponds to the case where the odds on the road team were 1.67 or lower (American odds: -150 or more). Our record was 13-1 on such games! I am fully aware that we are talking about a small sample size, but that’s still pretty impressive!

It looks like my initial fear of losing money on teams where the odds were high was wrong. We are earning a profit in each of the last 4 odds categories.

Here is the conclusion I draw from the table above: the suggested strategy seems to work for any odds, except maybe in the range 1.77 to 2.16 (American odds: -130 to +116). The losses in that range are fairly small, so it may just be an anomaly due to small sample sizes. My personal decision would be to bet a smaller amount in this range, but you are of course free to act as you’d like.

 

FINAL WORD

If I had to summarize this study in a few sentences, here is how it would go:

Bet all NHL road teams coming off at least one loss facing a home team coming off at least 4 consecutive wins. The estimated ROI on such bets is around 15%. Bet a smaller amount if the odds on the team you are betting on range between 1.77 and 2.16 (-130 to +116).

I would like to end this analysis by putting the results in perspective. You may wonder “how much can I expect to earn during one full season by following this strategy?” The truth is you won’t get rich because there aren’t that many games fitting the criteria for betting. We made 55-60 units over 9 seasons, which amounts to around 6.5 units per year. If you are a 100$ bettor, you can thus expect to make a 650$ profit during one NHL season. Nothing to get overly excited about, but you need to find several such angles in order to become a truly successful sports “investor”.

Thanks for reading this report and don’t hesitate to contact me for comments/questions!

Professor MJ (www.professormj.com)