As usual Shaun is on the button.ShaunWhite wrote: ↑Sun Jun 28, 2020 5:32 amProblem is James that charts are like fractals, you don't know what zoom level you're looking at. That dip might be almost invisble after a year, and you might not know what happened before it started. How does it look if you put a trendline on that? origin zero. I'm guessing it's still +ve.jameegray1 wrote: ↑Sun Jun 28, 2020 2:14 amHere is my latest "Think-I've-Cracked-It' automation.
500 markets in it was all roses. 1,000 markets in and it's a complete dog!
I do most of my betting based on graphs and statistics. The shortcoming I see in the usual backtesting graph is that it assumes you started on a particular day. What if that was a 'lucky day'? I find it more informative to plot it in reverse so instead of running a total from old to new I run it from new to old because it tells me what the value today would have been for every starting day in the past and identifies if your backtest started on a lucky day.
For example, let's assume you backed every Home team so far in the Premier league this season at Max Odds according to football-data.co.uk (with thanks to them). Your accumulated return would be …
… a profit so far of c.20 points (note this is an 'accidental' profit and not a recommended system!). But, +20 points only if you started on the first day of the season, what about if you started it on a different day. If you are working on a new system, you could be starting it on any day in the season so this is where I like to see what I call the 'Backward' graph where the running total is run backwards to show the total had you started on any date …
… Now you can clearly see that the 'system' would have been profitable whenever you started, other than a few weeks each side of the lockdown period, at around 15 points.
Obviously you would usually backtest over a longer period, this is just to show the alternative way to view the data. Hope it helps.