Am I Really Winning? — Betting Variance In Action

One of the biggest problems in gambling is that people convince themselves that their winning streaks were more than just pure chance.

But it’s not about what you’ve made before, or how many winners you’ve picked out. You need to be consistently hitting on value, full stop.

Here’s some basic tips to avoid being deceived by variance (betting swings).

Never Assume (That You’re Winning)

Just because you’ve been winning doesn’t mean you’re onto a winning trend. Variance (swings) in results are very common. You have to keep a level head and be careful not to let good/bad spells misguide you.

An Example of Variance…

Recently I’ve been undertaking a Mug Betting Experiment, where I bet on every single Premier League fixture until the season is up. I bet where I feel there’s an advantage and monitor my progress on a weekly basis.

Up until the previous 2 betting weeks the Mug Bets were doing great. But as I expected, normality has started to set in. Inevitably, several of the exceptionally profitable weeks have been offset by a run of bad form.

From +£180 down to +£100…

These kinds of swings are all part & parcel of betting. I’ve experienced it before, so I wasn’t remotely tempted to raise my stakes following the run of good form in weeks 1 through 8. This proved to be a wise decision. The moral is: it’s dangerous to assume that you’re going to continue winning without sufficient proof.

As for the Mug Bets… well, I’ve never claimed that my views on Premier League football would conjure up any ‘edge’ whatsoever. All I’m really interested in is whether I can generate a profit by the end of the season, with all peaks & troughs incorporated. Therefore I’ll remain rigid with my equal £5 stake.

…And so far so good.

Collect A Large Set Of Data

This is an important point and it leads on directly from the ‘Never Assume’ section above.

Refer back to the downturn in the graph and note that 80 bets in the experiment had been placed prior to this. So is 80 bets a little or a lot to go by? Can we learn anything from a sample of this size?

For many gamblers 80 bets is considered a lot. For example, if you placed 3 bets every weekend, it would take over 26 weeks (approximately half a year) for you to reach 80 bets. Furthermore, if you bet with large stake sizes — such as £50 — then that’s £4,000 worth of bets.

But that’s the problem. It’s not what you’ve staked, or how long it took you — it’s the frequency of bets that matters. Unfortunately 80 bets simply isn’t enough to draw conclusions from. We can’t be confident how the Mug Bets will perform until more data is collected, or alternatively, I back-test my ‘strategy’.

In the case of the Mug Betting Experiment I cannot back-test my selections because I have no way of knowing what I would have selected on all historic fixtures. Therefore the % 'edge' can only be established in the future.

Have Realistic Expectations

It’s important to have an idea of what your expectations should be when you’re betting.

I’m very much aware that even after the downturn of the Mug Bets after week 8, the current ROI is still a very high +20.37%. I’m not expecting to maintain this level of performance, because I have never seen a betting strategy that makes this long-term!

I know for a fact that 5-10% is a realistic expectation to have on a betting strategy, and I would be very pleased if I achieved this on the Mug Bets.

Using the Betfair Exchange helps, because I’m getting the best deal on odds.

ROI is ‘Discovered’ Over Time

Whilst I expect the current ROI of the Mug Bets to decrease over time, I don’t expect it happen at once. It could grind down over weeks or months, as the results gradually even out.

The same principle goes for Tipsters. Their selections could earn in the short-term. But they need to be ‘proofed’ over a large sample of bets before you begin to determine that they’re “profitable”. The most transparent sites for proofing are Tipstrr and Betting Gods.

Over time the ROI % will converge to where it really belongs. It’s like a casino with a fixed house edge on Roulette. The house doesn’t always win on every spin, but over time their % advantage reveals itself in their profits. It’s inescapable.

Variance in sports betting can be deceitful — so keep my advice in mind. I also recommend reading my post on Strike Rates, as it ties in nicely.

hey mate – nice website 🙂 regarding number of bets. What do you think is a good number to say ‘yes, my system has proved XX% ROI and it will continue to do so..’. Currently I am testing a system, just about 300 bets in with 10% roi but is that enough.. i guess it depends on the average odds or the average occurance of losses within those bets..?

Author

Reply to chris

Hi Chris,

It’s a tough one to call. 10% is a realistic ROI.

But whilst 300 bets sounds a lot to some people, it can soon turn around. That’s especially true if you’re using Betfair (or didn’t have an edge to begin with). Sometimes other bots/traders work out what you’re doing, and soon enough it takes all the winnings back. For proof of that, take a look at this: https://punter2pro.com/sample-size-betting-results-analysis/

If you continue to modify your strategy, adjust stakes, tweak the criteria, then that might not happen.

If the average odds are high so far, then the variance (swings) can also be pretty big. So that’s a factor with a “medium” sample size like yours. It might not have evened out quite yet: https://punter2pro.com/sure-confident-winning-betting-gambling-swings

Strike rates don’t you tell a lot without checking the odds. So I’d also take a look at this post: https://punter2pro.com/betting-strike-rate-verify-results/

In most cases, 1,000 bets is a healthy sample. Hope this helps.