Hi,
I’m struggling to test a mult deal bot (50deals ASAP) on a longer time period (like 3+ years).
There are tons of data and I would like to understand what kind of equipment is good to test such a strategy. Is it CPU Speed or is it RAM or is it GPU?
Thanks
If I’m not mistaken it depends on the cpu and ram. Worse case you can run it server-side.
okay, thanks for that. It takes quite a long time server side. I sent one this morning and still wait for the results… is it normal or is something wrong with it?
Hey Markus maybe you can take periods of 6 months of different years to check the overall performance of the bot, what you need to see is if the whole project is resident to different starts and conditions. If numbers are more or less similars could be the Bot can survive… Also you can test random coins for the same purpose.
Hope this help too
I like to see how it performs also in a longer time period… I normaly test my bots in several market conditions like you mentioned but it’s a workaround
While a backtest over longer periods is statistically better, I prefer to analyze certain difficult and varied periods where there is significant manipulation, like what we are experiencing now. In such cases, I examine the entries where losses or significant gains appear to assess their quality for each token. This approach makes it easier to discard a poor strategy or identify where the failure lies without having to conduct lengthy tests.
On the other hand, considering that most of the time we select good tokens with strong historical performance, if we conduct a long backtest, the strategy would have to be very poor not to yield good results if we had bought BTC during the Nakamoto era. 
Personally I have a Macbook Pro with an M3 Pro chip and it works well. But I find that the most is to recover the data. Even with fiber, it takes a long time. For your information, on tradingview, when you test more than 1000 combinations, it takes a long time too. The solution we found is to launch several virtual machines that we leave running in the background.
It looks like server-side backtests have been pending since last night, so there is no speed problem. Something in the configuration is possibly halting the backtest, which Maksym will fix.
The problem is in the settings. Low DCA step, low TTP, dynamic price—all of them lead to an enormous number of deals in the backtest (10k in Aug, if extrapolated to 3y, it will be around 350-400k). The backtest struggles with the number of deals. It starts fast, and then with more deals, it becomes increasingly slower.
We have optimized some deal methods. Locally, I’m able to run backtest 2.5x times faster. Also, execution speed seems no longer to depend on deal amount, at least in a one-month test. But I cannot guarantee that your browser will be able to handle the load.
Yes I thouht about this…the amount of deals is heavy to handle 
can I do your method as well or is it something only you and the team are able to do?
We already updated the backtester with the new method. You can try again, maybe in a smaller period first.