Draw System Drawdowns: How should we calculate maximum drawdowns?

I had an interesting question by email the other day and I want to share my answer here, too.

I found a post of yours on Reddit saying the Betaminic draw system has suffered a 115 point drawdown possibly due to a 0-0 correct score bot that had been added.  Is that something that’s always existed in the Betaminic draw system or is it something that’s recently been added and can be removed again to bring that maximum drawdown back down to 60?

That’s a good question. 
 
Short answer, 
I think -220 is a more realistic maximum drawdown for the Betaminic Draw system, meaning it needs a 300 point bank, but I think it will have fewer big drops than the BSG draw system. Removing the 0-0 bot will not return the drawdown to -60. But it would reduce the possible drawdowns, since that bot is betting an extra point each time and it has a much lower predicted win rate of 12%. Following fewer trigger draw strategies might reduce the volatility, but I think that would be a false economy, because the lower volatility is just coming from a lower bet volume. The same drawdown would eventually be reached. It would just take longer And I think that drawdown will be similar across all betting strategies with a 33% win rate.
 
Long answer,
I think there are 4 ways to judge drawdowns.
1. Looking at actual historical results.
2. Looking at all available historical data.
3. Looking at the Estimated Losing Sequence for that win rate.
4. Considering the law of numbers in regards to bet volume.
 
1. Looking at actual historical results.
I have been using my Betaminic Draw System for 4 months. At the beginning, there were just 5 Betaminic draw strategies providing trigger tips for my bot set. At that time there was little volatility.  After 1 month, I added another strategy to make it 6. There was still little volatility. After another month I added 3 more to make it 9 strategies. The maximum drawdown was still -60. In the last month I added 1 more to make it 10 strategies. Then in March there has been a losing run of -180 points. Only about -20 points of that was the new 0-0 correct score bot which had done very well in testing and I still think is a good addition.
So just going by the actual results of my 6 bot set in the past 4 months I had a -180 point drawdown from a 27% win rate from the last 200 bets. This is well below the usual 35% win rate of the bots and the aggregated average 33% win rate of the 12 years of historical results of the 10 strategies. I am very sure we will see a swing back in results to push that average backup to 33%. So the next 200 bets will probably have a win rate above 33%. But my -60 and -180 point drawdown figures came from the actual results of the 6 bot set.
 
2. Looking at all available historical data.
If I look at the past data from 12 years of the draw strategies I am following, using the Betaminic aggregator tool, it gives a combined maximum drawdown of -168 points over the past 12 years. But these figures include double/triple betting on overlapping tips. At the moment I am not double/triple betting on overlapping tips. So this figure cannot be used directly to calculate the drawdown, but it is an indication. Also, this drawdown is only for the core B2 trend. The other in-play bots and Correct Score bots will be different. So for the BetaminicDraw system this information is not that useful.
3. Looking at the Estimated Losing Sequence for that win rate.
I believe this ELS is a better way to judge drawdowns if a system has a consistent win rate.
The core trend according to the Betaminic aggregator has a win rate of 33%. 

Putting that number into the ELS calculator shows us the expected longest losing runs. And the maximum drawdown is likely to be a combination of 2-3 losing runs in close succession. So for the 800 tips the BetaminicDraw system has had in 4 months, we would expect an ELS of around 17 for the A1 and B2 bots. Also, I place 3 extra bets in-play that have win rates of around 24% which give an ELS of around 25.  

And the correct score bot has a win rate of around 12% which gives an ELS of 36 after 100 bets.

If I put them together we get a combined ELS of 145. (109 without the correct score bot.)
 
So the -120 drawdown looks very standard and expected in this context.
Furthermore, from experience with other strategies, I would expect 2-3 times that losing run to be the eventual maximum drawdown. So 220 or even 330 might be the eventual maximum drawdown of the system looking at those figures. And this leads on to the next point.
 
4. Considering the law of numbers in regards to bet volume.
Looking at the Estimated Losing Sequence (ELS) figures again, we can see that the ELS gets higher the more bets we make. This is the natural law of numbers. The more bets we place, the more chance of an unlucky or negative variance run coming. Also the larger those drawdowns and variances will be. So the longer we follow a betting system, the larger its drawdowns will get, and over the long, long term it will break its previous largest drawdown records. 
This means we can never truly say “-180 is the maximum drawdown” of a system. We can only say that in the case of a 33% and 24% win rate trends that “-180 is the actual maximum drawdown from the last 800 bets. and -109 is the expected longest losing run of the 5 bots after 1000 bets. And -135 is the expected losing run after 5,000 bets.” But such losing runs could happen close to each other, so we should plan on a drawdown of the ELS x2 for the number of bets we plan to follow the system for.
My BSG draw system has had over 9,000 bets in its 16 month history and has reached drawdowns of -220. This is well under its ELS x2 figure of 292 (146 x2). So while my BSG draw system appears to have a bigger volatility than the BetaminicDraw system, one significant factor is that my BSG draw system has had +9,000 trigger bets compared to the BetaminicDraw system which has had only 800 trigger bets. So in very simple terms the BSG draw system has a higher drawdown because it has placed more bets. We could even see the BSG draw system as an indicator of the possible drawdowns the BetaminicDraw system could face over the next 9,000 trigger bets, which for the BetaminicDraw system would take over 2 years to reach. So I would say that a 300 point bank would be the right choice for both systems.
 
So what do we do with this information?
We use this for bank and stake planning.
Since we are planning to follow these betting systems for the long term, and the longer we bet, the more chance of a big losing run coming, this makes me want to use a staking method of fixed level stakes plus a fixed bank over a fixed time period
Yes, that’s right, boring level stakes. Despite researching and writing “The Staking Plans Book”, I have gone back to level stakes for my big data, high volume betting systems. Progressive staking plans like percentage staking will just end up betting more on losers and less on winners. And absolutely stay away from any recovery staking plans. They just don’t work long term.
So looking at my own calculations I have done just here. I think having a 300 point bank for the BetaminicDraw system and also for the BSG draw system are correct if we are betting fixed level stakes. With both systems making over 50-70 points per month on average, then after 6 months they should have made enough to cover their banks being bankrupted in the case of an unusually large drawdown coming, which WILL come if we follow it for years and years on end, which I plan to. So by using level stakes, then we protect ourselves from the maximum drawdown run (ELS x2 or x3) that may come once every thousand bets, depending on our bet numbers, since the system will have made enough profit to restart itself.
So I recommend betting 10 GBP level stakes with a 3,000 bank for BetaminicDraw and a separate 3,000 bank for the BSG draws.
And when I run these banks, I risk 100% of them. I go until the bank is gone (due to extreme volatility) or until the trend win rate no longer makes positive ROI. 
In my mind, I am planning to run them for 1 year, and then review things. So I can always plan for how long I am going to risk how much for what expected profit.
 
When can we increase stakes with fixed level stakes?
We should only increase stakes during a short term losing run. When the long term win rate is 33% and the short term win rate is around 20% or lower, we KNOW there will be an above average winning run coming soon. Then we can increase stakes knowing that we are “buying the dip” and results will improve. Once we increase stakes, some people might choose to reduce stakes back to their original levels once the short term win returns to the trend line of 33%. This is okay, too. 
But each time you increase stakes, you must set aside a new bank for that stake increase. And we must accept that this bank could all be lost.
 
Are Betaminic strategies less volatile?
I don’t have hard data on this, but my Betfair market triggered bots tend to have more volatility than my Betaminic database triggered bots.  My Betfair triggered bots reach their stop limits earlier than Betaminic triggered bots do. I think this is because Betfair is the equivalent of the aggregated open market with no actual “opening odds”, but Betaminic database odds come from Pinnacle opening odds which are based on a consistent market model and only based on league games which will be much more predictable and consistent than all the various cups, friendlies, youth and women’s competitions on Betfair. So if you wanted to choose one of the draw system based on lower volatility, then I would recommend the Betaminic one first.
 
What is the lowest drawdown / lowest volatility system I know?
If you just want a working betting system with the lowest possible drawdowns, then take a look at these 4 strategies. https://www.tippingsports.com/en/profile/9159  The maximum drawdown for the past 11 years is just -25.

1 thought on “Draw System Drawdowns: How should we calculate maximum drawdowns?”

  1. Pingback: Betting System Results: +£1,468 in the last month after costs (Monthly Newsletter) - bettingsystemsguide.com

Comments are closed.