spot_price definition in the input csv files

[topherbuckley]

According to the docs the spot_price is defined as:

(optional): value of 1 unit of the given cryptocurrency at the time the transaction occurred; If the value is unavailable, to direct DaLI to read it from Internet historical data, write in __unknown and use the -s command line switch;

I would guess this is meant to be the value in the paired coin/currency, but it is a bit find ambiguous so wanted to check. For example lets say I have a BUY line on a ETHBTC exchange, where I'm buying ETH and paying in BTC, I'd assume spot_price should be how much 1 ETH is worth in BTC, or is it perhaps how much 1 BTC is worth in ETH? With the column called fiat_out_no_fee, I'd assume spot_price in this case wouldn't have anything to do with fiat, but looking at the definition for the fees:

crypto_fee: (optional) crypto value of the transaction fees. This field is mutually exclusive with fiat_fee. If fiat_fee is assigned, crypto_fee is set to 0. If crypto_fee is assigned, fiat_fee is set to crypto_fee * spot_price. The reason for this behavior is that if the fee is paid in fiat, then no crypto is used for the fee, but if the fee is paid in crypto, then its converted fiat value is needed to compute taxes. Note that RP2 models a non-zero crypto_fee with a separate fee-typed out-transaction.

I'd say spot_price is supposed to be the fiat equivalent (i.e. in the first example, how much 1 ETH is worth in USD or JPY). If this is the case, then which fiat is supposed to be used or does it not matter so long as you always use one? I've got transactions from both USD and JPY accounts, so I'm not quite sure how this is supposed to be handled. Should I be looking up historical exchange rates myself and converting them all to one currency or can dali-rp2 handle this somehow?

Also worth noting is that spot_price also appears in the intra_csv_file. If I'm moving ETH from one account to another, what then does spot_price represent? Here again I'd guess it is some fiat equivalent.

[macanudo527]

The spot_price should be the price of the digital asset in fiat, which defaults to USD. You can declare a different fiat by passing in the fiat_code of it. If it is not easily obtainable, I recommend letting the CCXT pair converter handle looking up the price for you.

The CCXT pair converter will 'route' the price to fiat. It will try to prioritize the exchange you made the purchase/sale at, but if that is not available will use the default exchange (currently Kraken).

For example, if you make a trade for ETH in BTC. It will look up the price for ETH in BTC at that moment if the market exists on that exchange. Then find the price of BTC in the fiat you are using.

ETH -> BTC -> USD

The final fiat is determined by what version of Dali-rp2 you are using. For Dali_us it is USD and Dali_jp it is JPY. so if you are filing taxes in Japan and America, you will need to run the program twice with the same data but different entry points (dali_us and dali_jp).

I hope that makes sense.

[topherbuckley]

Adding untested PR here. I mention in the comments of the PR that I'm having issues with the build/test workflow, so if you'd prefer, we can continue to discuss on the PR and I can make updates as necessary.

[eprbell]

The best way to fix this would be to add field index qualifiers to the manual plugin (similar to what RP2 does with the in_header, out_header and intra_header sections of its configuration file. This would allow the user to select their custom format. Absent that, yes, we'll have to add fiat_ticker at the end to preserve backward compatibility.

The UNKNOWN values in intra-transactions are used only if the transaction is partial.

Sure, let's continue the discussion in the PR: it'll be easier.

[topherbuckley]

I had another comment on this unrelated to the PR. Is there any way of making use of a spot price given by an exchange for crypto pairs? For example, I get the "traded price" from bitflyer for a ETH/BTC buy transaction which is given as the value of 1 ETH in BTC used at the time of trade. If the spot_price in the input csv is only intended for fiat price, I guess my recommended course of action would be to use __unknown for all crypto2crypto transactions, and only use the spot_price for crypto2fiat transactions. Is that correct?

[macanudo527]

Not really, the spot_price is always fiat since that is what is important for taxes. A crypto to crypto trade is actually a conversion. RP2 will model this as a sale of one of the assets into fiat, then the purchase of the other asset will use that amount of fiat as the basis cost. This is how pretty much all tax agencies want you to record the transaction.

So, let's take the ETH/BTC example.

You trade 0.1 BTC for 1 ETH.
Dali (and your tax agency) will calculate this as:
sale of 0.1 BTC for 3000 USD (if BTC is at $30,000)
purchase of 1 ETH for 3000 USD (that becomes your basis cost if you sell it later).

So, the "spot price" of ETH in BTC is actually not really needed in terms of taxes. A sale of ETH for BTC is actually two transactions in Dali and for taxes.

I hope that makes sense.

[topherbuckley]

Yep, that makes sense. I was aware of that, but just thought making use of the data might improve accuracy in some way but regardless, we'd have to look up the fiat price so I suppose not. Thanks for confirming!