With CoinMarketBook, you can evaluate cryptocurrencies by a different metric: "buy support".
Buy support is the sum of buy orders in USD equivalent from top 20 exchanges (planning to include more in future). It shows the actual amount that people want to buy.
One might say that it's easy to place a large buy order at very low price & increase buy support artificially. However, because of low price, the increase will be negligible.
So do you suggest that buy (and sell) orders on Bitfinex, for example, are genuine and there's no wash trading, no other tricks to make people believe market will go one way or another? LOL.
As for manipulative orders - you're right, they flash in the orderbook on and off. However, when I needed to exit, I actually filled a couple of them - so they were real.
Back in 2017 I traded a bit in cryptocurrencies. I noticed back then that a lot of buy and sell orders were being canceled. My assumption is that a lot of that was being done by bots that were either arbitraging, or trying to manipulate the market into putting orders at certain prices.
Based on that I am a bit skeptical towards trusting buy and sell orders as showing what amounts and prices people actually want to buy and sell at.
I would like to see someone gather and analyze data to show orders placed vs orders retracted and orders filled. But even the insight gained from that could be of little or limited use.
Yes, orders are being placed and cancelled all the time, but it mostly happens around bid / ask (and you're correct - this activity is generated by arbitrage bots).
The metric is useful for comparison - since the arbitrage bots work on all altcoins, their activity is included in the metric for all altcoins.
In crypto, it's always better to do your own research instead of following buy recommendations. That said, you can use the numbers in your research: for example, if two coins have a similar market cap, but one has much higher buy support, then it might be a better investment.
There aren't many fiat trading pairs for smaller coins. That means, if people are holding fiat, they most probably put it on low buy order for Bitcoin. I'd like to see exactly which markets are calculated in the results.
Currently, it's `'bitmex', 'binance', 'bithumb', 'bitfinex', 'okex', 'huobipro', 'bittrex', 'poloniex', 'kucoin', 'cryptopia'` (planned to include top 20, just fixed the message on home page)
- What does support/cap tell us? I see that it's a "gamble vs investment" metric but I don't understand the logic behind that statement.
- Is buy support at any level? As in, there's hug buy support for bitcoin, but is this counting people who have buy orders at the $6500 range all the way down to the $1 range?
- It shows you how much of market cap is backed by actual buy orders (~ whether it's vapor or not). The more buy support, the higher the probability that current price will hold (discounting for manipulative orders, of course).
One of the problems with this metric is it's easy to manipulate. Exchanges can easily lie about their book depth and/or people can put in lowball buy orders they have no intention of ever filling. To find out if those buy orders are real you would have to purchase them and all the orders above them, which would be very expensive and move the price a lot.
On the other hand, the instantaneous price of a coin is easy to verify. Everyone who buys or sells at the exchange can verify it, so exchanges can't easily lie about it.
Here's an interesting solution for ranking exchanges by their order book depth in a way that's more difficult to manipulate. Perhaps it could be adapted to rank coins. https://bitcoinity.org/markets/rank_explanation
Unfortunately as a metric this is far too easy to game and is heavily influenced by algorithmic order regimes. Maybe decent as signal in trading strategies but not useful for real fundamental analysis. Market makers create phantom interest all the time to try and scare the market through stop loss orders.
Also why not focus on the selling interest too? Any one sided metric is even less trustworthy.
Yes, you're right - it's possible to place fake orders and cancel them as soon as the actual trades start happening near them. However, from my experience, it only happens during the pump (when the whales are pushing the price higher).
Sell pressure is a bit different: while the lowest price is zero, the highest price is unlimited (technically, there is a limit enforced by exchange engine, but it's very high anyway). So actually, sell pressure should be calculated in base currency (in coin itself) instead of USD.
Market cap isn't perfect by any means, but if there was a better metric that could be used to get a quick valuation of something then the stock markets would already use it.
> Buy support is the sum of buy orders from top 20 exchanges.
> This is the actual amount that people want to buy.
Curious how you have arrived at $1.5T of buy support for BTC?
Given an example, a $1 bid on 100 BTC equates to $100 of buy support. In this example, would you count this as 100 * current price?
For reference, my custom tool reports BTCUSD pairs across major exchanges and shows approx 40kBTC of liquidity 10% either side of the current price. This is approximately $0.25B.
Whatever that validity of the metric itself, it's an interesting idea that if a market is different "enough", standard market metrics fall short of their goal.
On the website design, a fixed-width font would be helpful for the numbers. Currently, ETH at $22,898,033,476 looks one digit larger than BTC at $113,768,712,510.
As someone who is not very financially orientated, I find this difficult. Firstly because I don't really have a feeling for what that % means - presumably the ratio of market cap to book depth gives an indication of how sensitive the market cap is to large orders. So it combines liquidity and total volume - but I don't know why you would combine them into one statistic? Either I want a single number which means something clear, which this fails on. Or I want more information, in which case isn't it just better to tell the market cap AND book depth?
Also whilst market cap is easy to measure, book depth seems more difficult. How far away from the touch do you count? A million dollar offer which values bitcoin at 3cents represents almost no information. Firstly because by the time the price drops there's no way that offer will still be good and secondly because even if that offer were still good that change in price would represent a total shift in the market.
Also, this seems to compare apples to oranges, because some coins will contain market makers who are likely to leave orders in the book and cancel in response to signals, thereby creating phantom liquidity whilst others won't have enough volume to attract market makers and so all the depth is real.
Ratio (%) is calculated as "Buy support / Market cap". You can also see separate metrics for both "Buy support" and "Market cap" in the same row.
We count all buy orders (on all prices). It's true that low-hanging orders might be cancelled before they are executed. We've decided to count them anyway, because even low orders provide support during flash crashes.
I look at Daily Issuance (last column in https://onchainfx.com/v/SMT45r) as a measure of how well a (mineable) coin is doing, as this reflects the cost of energy spent on mining that coin every day.
This is one of the very few measures that is not easily faked...
42 comments
[ 3.6 ms ] story [ 103 ms ] threadBuy support is the sum of buy orders in USD equivalent from top 20 exchanges (planning to include more in future). It shows the actual amount that people want to buy.
One might say that it's easy to place a large buy order at very low price & increase buy support artificially. However, because of low price, the increase will be negligible.
What do you think? Tell us in comments.
As for manipulative orders - you're right, they flash in the orderbook on and off. However, when I needed to exit, I actually filled a couple of them - so they were real.
Based on that I am a bit skeptical towards trusting buy and sell orders as showing what amounts and prices people actually want to buy and sell at.
I would like to see someone gather and analyze data to show orders placed vs orders retracted and orders filled. But even the insight gained from that could be of little or limited use.
The metric is useful for comparison - since the arbitrage bots work on all altcoins, their activity is included in the metric for all altcoins.
Please note that some markets may allow leverage - we'll mark them with asterisk on next deploy.
Buy support shows demand for this coin.
Any plans to include the top decentralized exchanges? (https://etherscan.io/stat/dextracker)
You might want to only include US exchanges (poloniex, bittrex, coinbase pro + bitstamp.. maybe Kraken)
- Is buy support at any level? As in, there's hug buy support for bitcoin, but is this counting people who have buy orders at the $6500 range all the way down to the $1 range?
- Yes, we sum buy support at all levels. You can see the explanation here: https://news.ycombinator.com/item?id=18162456
On the other hand, the instantaneous price of a coin is easy to verify. Everyone who buys or sells at the exchange can verify it, so exchanges can't easily lie about it.
Here's an interesting solution for ranking exchanges by their order book depth in a way that's more difficult to manipulate. Perhaps it could be adapted to rank coins. https://bitcoinity.org/markets/rank_explanation
Do you have any advice on how to encourage sharing?
Also why not focus on the selling interest too? Any one sided metric is even less trustworthy.
You can see the explanation for including all buy orders here: https://news.ycombinator.com/item?id=18162456
Sell pressure is a bit different: while the lowest price is zero, the highest price is unlimited (technically, there is a limit enforced by exchange engine, but it's very high anyway). So actually, sell pressure should be calculated in base currency (in coin itself) instead of USD.
Curious how you have arrived at $1.5T of buy support for BTC?
Given an example, a $1 bid on 100 BTC equates to $100 of buy support. In this example, would you count this as 100 * current price?
For reference, my custom tool reports BTCUSD pairs across major exchanges and shows approx 40kBTC of liquidity 10% either side of the current price. This is approximately $0.25B.
So your tool counts the closest 10%? Would be interesting to expand it to 100% and compare the metric.
On the website design, a fixed-width font would be helpful for the numbers. Currently, ETH at $22,898,033,476 looks one digit larger than BTC at $113,768,712,510.
Good idea, we'll switch to fixed width.
Also whilst market cap is easy to measure, book depth seems more difficult. How far away from the touch do you count? A million dollar offer which values bitcoin at 3cents represents almost no information. Firstly because by the time the price drops there's no way that offer will still be good and secondly because even if that offer were still good that change in price would represent a total shift in the market.
Also, this seems to compare apples to oranges, because some coins will contain market makers who are likely to leave orders in the book and cancel in response to signals, thereby creating phantom liquidity whilst others won't have enough volume to attract market makers and so all the depth is real.
We count all buy orders (on all prices). It's true that low-hanging orders might be cancelled before they are executed. We've decided to count them anyway, because even low orders provide support during flash crashes.
Flash crash horror movie: https://youtu.be/R7WfSTePuA0?t=171
This was possible because there wasn't much buy support. And that's ETH - imagine what can happen with lowcap altcoins.
P.S. The fact that some altcoins don't even have market makers is a little... unsettling, don't you think?
This is one of the very few measures that is not easily faked...