Building a great Bitcoin exchange, part II: should all orders be limit orders?

[aes1]

Exchanges commonly have two options for buying and selling: limit orders (executed only at the specified price or better) and market orders (executed immediately or not at all).

So, a stupid reductionist idea: why do we need the two different types? Why not just use limit orders?

Pros:
- it would make the user interface simpler - the user wouldn't have to think of options
- it would make the trade engine simpler (and therefore easier to optimize, verify, and test)

Cons:
- but that's not what we are used to!
-

The user interface would, by default, calculate the price which is needed for the limit order to be executed completely at a high probability (i.e. the market doesn't suddenly move into the wrong direction), and have that set as a default. And that's a "market order" for you.

[Come-from-Beyond]

Exchanges commonly have two options for buying and selling: limit orders (executed only at the specified price or better) and market orders (executed immediately or not at all).

So, a stupid reductionist idea: why do we need the two different types? Why not just use limit orders?

Pros:
- it would make the user interface simpler - the user wouldn't have to think of options
- it would make the trade engine simpler (and therefore easier to optimize, verify, and test)

Cons:
- but that's not what we are used to!
-

The user interface would, by default, calculate the price which is needed for the limit order to be executed completely at a high probability (i.e. the market doesn't suddenly move into the wrong direction), and have that set as a default. And that's a "market order" for you.

We need stop orders. Most of traders can't afford to buy/develop a trading bot and can't sit in front of monitor 24/7.

[Furyan]

We need stop orders. Most of traders can't afford to buy/develop a trading bot and can't sit in front of monitor 24/7.

This. I care far more about this than limit vs market orders. They're tools, and both are useful and necessary.

The lack of stop orders on any exchange is a travesty.

[aes1]

We need stop orders. Most of traders can't afford to buy/develop a trading bot and can't sit in front of monitor 24/7.

This. I care far more about this than limit vs market orders. They're tools, and both are useful and necessary.

The lack of stop orders on any exchange is a travesty.

Did you already take a look at Kraken?. It's in private beta right now, but you can request an account if you're interested.

They have plenty of options for closing the trade, including stop-loss, take-profit, stop-limit and about 7 others. Looks quite promising if you're the pro-trader type.

(I'm not affiliated with Kraken, just participating in the beta.)

[codro]

Exchanges commonly have two options for buying and selling: limit orders (executed only at the specified price or better) and market orders (executed immediately or not at all).

So, a stupid reductionist idea: why do we need the two different types? Why not just use limit orders?

Pros:
- it would make the user interface simpler - the user wouldn't have to think of options
- it would make the trade engine simpler (and therefore easier to optimize, verify, and test)

Cons:
- but that's not what we are used to!
-

The user interface would, by default, calculate the price which is needed for the limit order to be executed completely at a high probability (i.e. the market doesn't suddenly move into the wrong direction), and have that set as a default. And that's a "market order" for you.

A market order is either a sell at $0 or a buy at $inf as far as the trading engine is concerned. You won't be selling at 0 or buying at $inf because it matches bids so you get the best deal. However, there's no extra complexity in a market order.

[aes1]

A market order is either a sell at $0 or a buy at $inf as far as the trading engine is concerned. You won't be selling at 0 or buying at $inf because it matches bids so you get the best deal. However, there's no extra complexity in a market order.

Yes indeed. Thanks for pointing that out. I was trying to make the situation overly complex by thinking about doing a fill-or-kill instead of an actual market order. That's probably not necessary at all.